The 'Security Digest' Archives (TM)

Archive: About | Browse | Search | Contributions | Feedback
Site: Help | Index | Search | Contact | Notices | Changes

ARCHIVE: TCP-IP Distribution List - Archives (1994)
DOCUMENT: TCP-IP Distribution List for December 1994 (674 messages, 501022 bytes)
SOURCE: http://securitydigest.org/exec/display?f=tcp-ip/archive/1994/12.txt&t=text/plain
NOTICE: securitydigest.org recognises the rights of all third-party works.

START OF DOCUMENT

-----------[000000][next][prev][last][first]----------------------------------------------------
Date:      1 Dec 1994 00:06:49 GMT
From:      berger1@ehsn12.cen.uiuc.edu (Andreas  Berger)
To:        comp.infosystems.www.users,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.infosystems.www.providers,comp.os.ms-windows.networking.tcp-ip,comp.os.ms-windows.networking.misc
Subject:   Re: WWW Server/Client without INTERNET??

In article <3birvu$sd4@news.iastate.edu>, maad@iastate.edu (Madhu Rajamani) writes:

|> I need to run a World Wide Web server (on a PC or PC based network)
|> within an organization which does NOT have internet access.  Also, I'd like
|> to be able to dial in to the local network (or standalone PC) and access 
|> HTML documents using MOSAIC or NETSCAPE.
You have basically two options:
1) not to have a web server, using the normal filesystem to hold
   you HTML documents via the file:// URL. Mosaic then uses a normal
   DOS call to open the document. The exact procedure and syntax is
   described in the documentation of Mosaic.
2) to have a WEB server, in case you want to stay with PCs only most
   likely the Windows based HTTPD, WHTTPD. This can be found at NCSA or
   at least a refernce to it. You can then use Mosaic the normal way.
   Or use a PC UNIX (e.g. linux for no cost!).

That you are not connected to the internet is no problem, just make up
your own IP numbers for case 2. In case 1 you do not even need a TCP/IP
stack or for that matter even a network. In case two you would have to
setup a TCP/IP stack on server and clients (in case of the windows server
could be on a machine which is being used but I would not recommend it) 

|> Madhu Rajamani
Andreas

-----------[000001][next][prev][last][first]----------------------------------------------------
Date:      1 Dec 1994 00:17:50 GMT
From:      Jerry Scharf <scharf@corpmis.sjc.hw.sony.com>
To:        comp.protocols.tcp-ip
Subject:   Re: Subnetting to be avoided ?

> 
> Steinar Haug (Steinar.Haug@runit.sintef.no) wrote:
> 
> HP has informed me that supernetting will NOT be supported in 10.0, but 
> will be offered as a patch sometime at the end of 1995 / early 1996.
> 
> If you plan on using supernetting, it's important to be aware that not 
> all TCP/IP implementations support it.
> 
> Seth Bromberger
> 
I question how important it is that end stations do supernetting. In my
experience, user compiled gated would be the only thing that would
understand a supernetting protocol. What does HP say to supporting OSPF?
The way I set things up is we run either RIP or IRDP to have the end
stations discover their routers. All the fancy things like supernetting,
aggregated sunetting and the like is handled among the routers.

As long as Cisco, Wellfleet and that class of box supports the functionality,
it can be very effectively deployed in networks. My bias may show, but
anyone using an HP machine as a router deserves what they get.

Jerry


-----------[000002][next][prev][last][first]----------------------------------------------------
Date:      1 Dec 1994 01:26:39 GMT
From:      tli@cisco.com (Tony Li)
To:        comp.protocols.tcp-ip
Subject:   Re: IP multicasting

In article <3bhrp7$7ui@ubszh.fh.zh.ubs.com> bsz@dla831.ubszh.net.ch writes:
    I'm looking for the relevant information regd. IP multicasting, i.e.
    could anyone please quote me the list of relevant RFCs?  

RFC 1112 is probably a good start.

    Kindly also point me into the direction of an ftp-mail server where I
    can fetch these from.

ds.internic.net has them, but there are probably closer...

    Lastly: which routers currently support the IP multicast protocol?
    There is a big debate here, claiming that for instance CISCO does
    currently not support IP m-cast in their products. Is that correct? 

cisco shipped multicast support in 10.2, almost two months ago.

Tony

-----------[000003][next][prev][last][first]----------------------------------------------------
Date:      1 Dec 1994 02:01:50 GMT
From:      raj@cup.hp.com (Rick Jones)
To:        comp.protocols.tcp-ip
Subject:   Re: TESTING TCP/IP throughput over FDDI...

Teddy Wang (teddy@superbook.albany.edu) wrote:
: I'm doing some benchmarking of throughput (as close as actual maximum as 
: possible) of our FDDI ring...  Since I don't want to contaminate my results
: with bottlenecks in CPU/DISK performance, I was wondering if anyone out in
: netland has source code which will do these test???

I would suggest netperf (OK, I am slightly biased :) Netperf includes
the common bandwidth tests - unidirectional data stream. It also has
latency tests - so people can discover for themselves that high
bandwidth does not always mean low latency.

The latest netperf can be found by following links from
http://www.cup.hp.com/netperf/NetperfPage.html with a WWW browser, or
via anonymous FTP from ftp.cup.hp.com. It can be found on other
servers as well - ftp.sgi.com being one.

If you have a forms-capable WWW browser, then you can also peruse the
netperf results database at the time you download netperf. Better
still, you can *submit* netperf numbers to the database so that others
can benefit from the work you have done and perhaps build upon it.

rick jones
shameless promoter of netperf and the netperf database :)

-----------[000004][next][prev][last][first]----------------------------------------------------
Date:      Thu, 01 Dec 94 02:41:22 GMT
From:      schaft@best.com (Schaft)
To:        comp.protocols.tcp-ip,comp.infosystems.www.providers
Subject:   Re: Experiences with Internet providers in Bay Area?

	Check out  "BEST.COM" and "PORTAL.COM".  I use both (don't work for
either of them).

	PORTAL.COM is a nationwide service with 24 hour access for 2.95/hr
from almost anywhere.  It is located in Cupertino, and has PoPs in 415, 510
area codes.  Has a lot to offer - for $19.95/mo you get a Unix shell account (where
you can run TIA) and access to a BBS-like interface with file areas, chat areas, etc.

	SLIP/PPP accounts are more, and a  "vanity plate" costs a bunch. Max speed
 for dialins is 14.4.  I believe that max connect speed is 19.2, but may be higher for
leased lines. They have a T-1interface to the Internet.

	Telnet to "portal.com" or the URL is "http://www.portal.com"

	BEST.COM is a new kid on the street - 421 Castro in Mountain View, to be
precise... :)  They are looking to be a "niche" provider - local service for local folks.
They are currently local to 408 and 415.  Access from 510 is coming soon.
They work the hours most folks are on line....like until around midnight.  Trouble
calls are answered by a real person, and solutions are in real time...you type, they
watch and advise.  Much better than talking to an answering machine, IMHO. (They
do have an answering machine, and an IRC channel also... #best)

	SLIP/PPP, static, is $30/mo, no time limits,  with 35M of disk space for home
directory (10M) and FTP/WWW stuff (25M).

	They have two T-1s, and all modems are v.34 28.8s.  "Vanity plate" domain
address is a one-time charge of $30.  They have a "store front" approach - bring in
your box and they will help you load Windows/MAC software, plug it in there, and
 test it out...not a bad deal if the Internet (and its required software) is a bit daunting.
They have a couple of machines set up so you can see the various apps, like NetScape
and WAOL operate at *real* speed... :)

	Both of the above offer leased line capability - and Best also offers Class C
addresses.

	I like and use both services.  They offer me different capabilities, and
of course the prices are different.  I would be hard pressed to pick one over the other,
but I definitely use "best.com" more..... I like the speed, service, and attitude...

	Best can be reached at the addresses in my .SIG, "info@best.com", or
URL "http://www.best.com"


DISCLAIMER: Since I'm just a customer to these folks, I may have inadvertently
left out stuff they can do for companies.  All I've said is what I've seen as a private
customer.  Call them for the real scoop.....should be close tho.... :)

...........................................................................................................................

safari@nbn.com (Bob Walsh) wrote:
>In article <3b7bbk$4ji@aldwych.gig.nl>, hendrik@gig.nl (Hendrik te Winkel)
>wrote:
>
>> Hello
>>
>> I need some information on Internet Providers in the SF / Bay Area.
>> I already know netcom, little garden and SantaCruz but I would like
>> to know if any of you (esp. companies) have good/bad expreience
>> with one or the other provider.
>> I want to hook up our local SF office to the Internet through a
>> leased-line and can use any info you can give me.
>> I looked through some WWW pages and found an enormous list
>> with providers but the majority were SLIP/personal IP providers
>> and I look for a 'proffesional one' with good support, fallback, etc.
>>
>> Thanks for any help on this, or pointers to other info.
>> Bye, Hendrik
>>
>> --
>> Hendrik te Winkel       hendrik@gig.nl
>> ElectroGIG Europe
>> Amstel 222              tel: +31 20 5217 341
>> 1017 AJ Amsterdam       fax: +31 20 622 68 01
>
>If you are in Marin County, Check out North Bay Network - they've been
>very good to work with. they're at info@nbn.com.
>
>--
>Safari Software, Inc. - Software that makes a difference.


                             --  Schaft  --
                             schaft@best.com
                URL: http://www.best.com/~schaft/schaft.html

                     "I've felt better but it cost more"
        _______________________________________________________

       Shell & SLIP/PPP at 28.8 with 35M of space for 408/415
                     $30/mo Flat Rate - No Time Limits
            URL: http://www.best.com  e-mail: info@best.com

-----------[000005][next][prev][last][first]----------------------------------------------------
Date:      1 Dec 1994 03:09:52 GMT
From:      mschlack@interserv.com
To:        comp.protocols.tcp-ip
Subject:   ATM backbones

I'm looking for is an ATM user who has installed an ATM switch as the
backbone. Basically a company with a plain old Ethernet LAN that goes
into a router which then plugs into the ATM switch. I'm not interested in
ATM to the desktop. Just data moving back and forth across. And I
'I'd like to know what the network looked like before installing the
switches. Contact Kim Twist on MCI mail or leave a message for me here. Thank you.




-----------[000006][next][prev][last][first]----------------------------------------------------
Date:      1 Dec 1994 04:13:21 GMT
From:      maad@iastate.edu (Madhu Rajamani)
To:        comp.protocols.tcp-ip
Subject:   Q: WWW Server/Browser without Internet on Novell/WFWG??



I need to run a World Wide Web server (on a PC or PC based network) within an 
organization which does NOT have internet access.  Also, I'd like to be able 
to dial in to the local network (or standalone PC) and access HTML documents 
using MOSAIC or NETSCAPE.

Has anyone ever tried something like this before?  Or if you have any ideas or 
sugestions as to how I can do it, please mail to me or post here.

Any help will be sincerely appreciated.

Thank You

Madhu Rajamani

--
maad@iastate.edu
404 Marston Hall
Iowa State Univ.
(515) 294-1675

*** HITCH HIKING on the INFOBAHN ***
-- 


-----------[000007][next][prev][last][first]----------------------------------------------------
Date:      Thu, 1 Dec 1994 04:46:35 GMT
From:      chiejin@nfx.com (Chie-Jin Cheng)
To:        comp.protocols.tcp-ip
Subject:   What is CSLIP?

Hi there,

What is CSLIP?  What does it differ from SLIP?

Thanks,

Chie-Jin

------------------------------------------------------------------------------
Chie-Jin Cheng			nFX Corporation
				A new type of Special Effects........
				nth dimension Facial Effects........
(408) 748-9200 (voice)		4800 Great America Parkway
(408) 748-9272 (fax)		Santa Clara, CA 95054

-----------[000008][next][prev][last][first]----------------------------------------------------
Date:      Thu, 1 Dec 1994 05:12:35 GMT
From:      cslater@netcom.com (Charlie Slater)
To:        comp.protocols.tcp-ip
Subject:   Re: (Q) SLIP & Subnetting

In article <bobley-2111941901210001@kslip1.apl.jhu.edu>,
Brett Bobley <bobley@mailstorm.dot.gov> wrote:
>Yes, the Netblazer supports proxy arp.  However, for proxy arp I would
>have to enter the MAC address of all the machines which might be dialing
>in via SLIP.  This kind of defeats the convenience of having a SLIP ip
>pool.


The NetBlazer "auto-arp" feature will automatically proxy-arp for a
host that dials in to the NetBlazer.  

I agree with those who have suggested proxy-arp as the way to go.  You
can put a workstation on the Ethernet at the office for a while, then
take it home and dial in via SLIP.  The responds to arp requests for
your workstations IP address while you are dialed in.

Your routing tables stay simple and you don't have to change them when
you dial in (either with RIP or by hand).

Wal-Mart has a LAN with seven (or more) netblazers on it and 150 (or
more) dial-in lines in one hunt group.  Because auto-proxy arp directs
IP packets to the right netblazer, it does not matter what netblazer
the remote site ends up connecting to.

To my knowledge, this auto-proxy arp feature was first invented at LBL
and is part of the current CSLIP distribution.

-charlie

-- 
Charlie Slater  (cslater@imatek.com)    408.244.8864  (Voice or FAX)
1119 Timberpine Ct.
Sunnyvale, CA 94086

-----------[000009][next][prev][last][first]----------------------------------------------------
Date:      1 Dec 1994 05:58:03 GMT
From:      kbeaucha@gpu.srv.ualberta.ca (Kevin Beauchamp)
To:        comp.dcom.lans.ethernet,comp.dcom.net-management,comp.protocols.tcp-ip,comp.security.misc
Subject:   Q re: Unassigned addresses and security

Hello: 

I have a few questions about unassigned addresses and their use in a
departmental LAN environment.

The gist of most of my questions will be with respect to the use of
unassigned addresses as a security measure, and methods to provide
access to so-numbered nodes from other nodes on a common segment,
and on a private segment.

My understanding of the unassigned addresses, from reading through the 
RFC1597, is that there exist blocks of IP addresses reserved for use in
so-called private networks.

They are allocated to allow selected nodes or networks to exist within a 
department attached to the Internet which are not accessable ("visible" may
be a better term?) from the Internet.

Let's say that in one instance we have a network that looks like this:

                          Campus Backbone
       =======================================================
                              |
                              |
                           ------
                          |      |
                          |      | Departmental
                          |      | Router
                           ------
                              |
                              |       Departmental Segment
         --------------------------------------------------------
            |                     |                         |
            |                     |                         |
          -----                 -----                     -----
         |     |               |     |                   |     |
         |     |               |     |                   |     |
          -----                 -----                     -----
          Public                Public                    Private
          Node                  Node                      Node

The private node is an instrumentation unit, with no local user interface
and no need to ever send or receive packets from beyond the departmental
router. Because of this, I am leaning towards the unassigned addresses
instead of firewall technology.  This seems like a much easier to configure 
method to prevent outside tampering with the data-gathering private nodes.

As I understand the RFC, the responsibility for rejecting internet packets 
TO _or_ FROM the private node (using unassigned addresses) resides with the 
departmental router.

Is this configured by the administrator, or is it built into the router as a 
default filter?

It is also important that the public nodes on this segment have easy,
tranparent access to the private nodes, as this is where the user
interface for the instrumentation data will reside.

I assume that since they are on the same segment, and all nodes will "see"
the packets, the private nodes will respond to packets sent from the public 
nodes without requiring editting any cofiguration files on the public nodes.

Will use of unassigned addresses allow me to accomplish my goals of easy 
access for users on the same net, but anonymity from users outside the
department?

Now, suppose that the network looks like this:

                          Campus Backbone
       =======================================================
                              |
                              |
                           ------
                          |      |
                          |      | Departmental
                          |      | Router
                           ------
                              |
                              |       Departmental Segment
         --------------------------------------------------------
            |                     |                         |
            |                     |                         |
          -----                 -----                     -----
         |     |               |     |           Private |     |
         |     |               |     |           Router  |     |
          -----                 -----                     -----
          non-user              user                        |    
          Node                  Node                        | 
                                        Private Segment     |
                      ---------------------------------------------
                        |                         |
                        |                         |
                      -----                     -----
                     |     |                   |     |
                     |     |                   |     |
                      -----                     -----
                     Private                   Private
                     Node                      Node

First, what do we gain in security from:

	1) Locally attached nodes that don't need access to the private
	   Nodes:

	2) unwanted Internet access?

Here, I assume that User nodes would have an additional entry in their 
local router table for the private router (Non-user nodes would omit this 
entry) and that the private router would have a static routing table that 
includes the addresses of all of the nodes on the private segment.

Is the address of the private router also an unassiged addresses?

Is this adequate security to prevent interference with the private 
nodes and the tasks running on them?

Is there a danger of the private node IP addresses being propogated
beyond the private segment, or to non-user nodes? How?

How do I provide name services which allow user nodes to reference
private nodes by a name or alias? (I have seen articles which refer
to "public" and "private" name servers. Will they be required here?)

How difficult is it to set up a router to block traffic on the basis 
of datagram type?  Lets say that the data-gathering private nodes used 
IPX between themselves, but TCP/IP to communicate to the user nodes.  
If I want to prevent the IPX packets from traversing the private router, 
how hard is this?

And lastly again, will the use of unassigned addresses meet my goals
for access and security?

As an aside, can anyone recommend a reference on address allocation
guidelines?

As you can see, I lack experience in configuring systems of this type.
Any help will be very much appreciated.

Direct e-mail replies prefered if possible.
Thank you in advance
kmb


-----------[000010][next][prev][last][first]----------------------------------------------------
Date:      Thu, 01 Dec 94 15:38:09 PST
From:      Johnam@dsusmail.dsus.datastorm.com
To:        comp.protocols.tcp-ip
Subject:   How to contact NIC for registering domain name

What is the Phone number, address, E-mail address etc. for NIC for 
registering a domain name.
jam


-----------[000011][next][prev][last][first]----------------------------------------------------
Date:      1 Dec 1994 07:52:34 GMT
From:      teklogic_osd@apollo.is.co.za (Teklogic - Altech Systems)
To:        comp.object,comp.protocols.tcp-ip,comp.unix.solaris,comp.lang.c++,comp.client-server
Subject:   Re: ACE version 2.15.5 now available

Greg Gibson (gibsong@agcs.com) wrote:
: In article <3aul4m$bh6@tango.cs.wustl.edu> schmidt@tango.cs.wustl.edu (Douglas C. Schmidt) writes:
: >COPYRIGHT INFORMATION FOR ACE
: >
: >...  It would be great to see this distributed
: >evolve into a comprehensive, robust, and well-documented C++ class
: >library that would be freely available to everyone.  Naturally, I am
: >not responsible for any problems caused by using these C++ wrappers.
 
: It is not clear to me why you would 'naturally' not be responsible for
: any problems caused by using this library.  What if a problem within this
: library caused all of the Iridium satellites to become useless?  I'm not
: saying you have full responsibility by any means, but to claim a desire
: of a comprehensive and robust library without responsibility for any
: lack of those characteristics seems like a conflict.

Don't look a gift horse in the mouth!

If you want someone to take responsibility fork out the $$$$. The above 
is simply a standard disclaimer that anyone in their right mind would put 
with free software.

cheers,
Wayne


-----------[000012][next][prev][last][first]----------------------------------------------------
Date:      1 Dec 1994 10:29:44 GMT
From:      ian@spider.co.uk (Ian Heavens)
To:        comp.protocols.tcp-ip
Subject:   BSD support for Variable length subnet masks?

I'm trying to find out whether (and how, if so) BSD TCP/IP supports
variable length subnet masks.  The code doesn't seem to support it
 - at least, the routines that convert network addresses to numbers
 don't - yet a netstat -r on our local BSD machines shows different
subnet masks.

ian

---
	 Ian Heavens			ian@spider.co.uk                 
 	 Spider Software
 	 Spider Park, Stanwell Street		
	 Edinburgh, EH6 5NG, Scotland	+44 31 555 5166 (Ext 4735)
--

-----------[000013][next][prev][last][first]----------------------------------------------------
Date:      1 Dec 1994 13:35:20 GMT
From:      pascal@decgraph.evt.DEC.COM (Pascal d'ORNANO)
To:        comp.protocols.tcp-ip
Subject:   Re: WANTED: Automated FTP client with scripts



Have a look to .netrc although this may not be sufficient to achieve 
your goal.

Regards

-----------[000014][next][prev][last][first]----------------------------------------------------
Date:      1 Dec 1994 14:27:36 GMT
From:      Gordon McAndrew <gordon@aec.env.gov.ab.ca>
To:        comp.infosystems.www.users,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.infosystems.www.providers,comp.os.ms-windows.networking.tcp-ip,comp.os.ms-windows.networking.misc
Subject:   Re: WWW Server/Client without INTERNET??

maad@iastate.edu (Madhu Rajamani) wrote:
> 
> 
> I need to run a World Wide Web server (on a PC or PC based network)
> within an organization which does NOT have internet access.  Also, I'd like
> to be able to dial in to the local network (or standalone PC) and access 
> HTML documents using MOSAIC or NETSCAPE.
 
> Madhu Rajamani
> 

There is no requirement for an Internet connection to run a WWW server.
But you probably have to be using the TCP/IP protocol on the LAN.  If you
have a SLIP server of some kind you will be able to dial in and use whatever
services are on the LAN.  You will probably have to use the IP addresses
directly or set up a DNS on the LAN.

We use local FTP & WWW servers to give out info to the staff.

Gordon McAndrew
Vegreville, Alberta, Canada

gordon@aec.env.gov.ab.ca

> 
> *** HITCH HIKING on the INFOBAHN ***
> -- 
>       o o o o o o o . . .   _________________________ _____=======_||__________
>     o      _____           ||Madhu Rajamani         | |rajamani@cs.iastate.edu|
>     o                       |404,Marston,ISU,50011  | |off:(515)294-1675      |
>   .][__n_n_|DD[  ====_____  |103 Stanton Ave. #37   | |Res:(515)296-2646      |
>  >(________|__|_[_________]_|_______________________|_|_______________________|
>  _/oo OOOOO oo`  ooo   ooo  'o^o^o             o^o^o` 'o^o^o             o^o^o`
> -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-
> The hardest thing to learn in life is which bridge to cross and which to burn.
> 


-----------[000015][next][prev][last][first]----------------------------------------------------
Date:      Thu, 1 Dec 1994 14:55:45 GMT
From:      orel@lpuds.oea.ihep.su (Oleg Orel)
To:        comp.protocols.tcp-ip
Subject:   Re: WANTED: Automated FTP client with scripts

Carlos Miguel Paraz (cparaz@admu.edu.ph) wrote:
: Hello,
 
: I would like to have an FTP client which can take a list of sites and
: filenames, and get them, logging in as anonymous and entering my 
: email address as password.  This would ideally give me a log of 
: the operations it made, and even schedule retries if the sites are full.
 
: Any suggestions will be appreciated.
: --
: Miguel A. Paraz -=-=-=-=-=||-= "How far can this Net go?" =-=-=-=-=-=-=-=-=-=
: cparaz@balut.admu.edu.ph =||-=  Ateneo de Manila University, Philippines   -=

This client is uftp: ftp.oea.ihep.su:/libs/libftp/libftp-2.0.zip (for unixs)



--
ˆD

-----------[000016][next][prev][last][first]----------------------------------------------------
Date:      Fri,  2 Dec 94 01:00:00 -0600
From:      alfred.harding@windmill.com (Alfred Harding)
To:        comp.protocols.tcp-ip
Subject:   Remote TCP/IP to Unixware

I have a Unixware server that connects local workstation over TCP/IP.

We are adding a remote office that has three workstations running
Windows for Workgroups over Ethernet.

I would like to allow these remote workstation access to the Unixware
server.  I know that I could install a modem in each workstation and
allow them SLIP access into the Unixware server.

Is there a way where I could allow the remote workstation to access the
Unixware server over their Ethernet adapters into an Ethernet/Modem with
dialed up access to the Unixware server?

We will be growing the number of workstations at the remote site to 10
stations so I do not want to purchase 10 modems and 10 phone lines.  If
I could use one phone line with a 28.8 modem between the remote stations
and the Unixware server this would be what I would prefer.

Each remote workstation would be running telnet into the Unixware
server.

Any ideas or assistance would be appreciated!

Alfred S. Harding

-----------[000017][next][prev][last][first]----------------------------------------------------
Date:      2 Dec 94 00:00:34 -0500
From:      s2700142@nickel.laurentian.ca (Cyan Bloodbane)
To:        comp.protocols.tcp-ip.ibmpc,comp.protocols.tcp-ip
Subject:   TCP/IP pkgs for Dos.. ftp sites?

Hi.

I'm rather clueless when it comes to networking and such, so I'll come down
to the point.

I'm looking for ftp sites that may contain shareware tcp/ip pkgs for ms-dos.
(if there IS such a thing) .. reason being, I'm more or less looking
to telnet, ftp from dos, instead of using Winsock or Trumpet.. Windows in
general, as it's WAY too slow to do me any good. Other than that, it
takes up more drive space than I'd like, and I figure a nice little
dos version would be a hell of alot smaller.

If there are such sites, anyone care to mention them to me? ;)

If you can, please email me a reply as I'm on a lame Vax (ugh, NOW you know
why ;) and reading threads proves to be.. well, difficult, to say the least. ;)

Thanks,
-Gilles Melanson


-----------[000018][next][prev][last][first]----------------------------------------------------
Date:      1 Dec 1994 16:21:07 GMT
From:      birknerm@email.exide.com
To:        comp.protocols.tcp-ip
Subject:   Duplicate IP Software?

Does anyone know of any good software that will detect duplicate IP addresses?


-----------[000019][next][prev][last][first]----------------------------------------------------
Date:      1 Dec 1994 16:36:56 GMT
From:      landmark@cs.tu-berlin.de (Torsten Kerschat)
To:        comp.protocols.tcp-ip,comp.protocols.misc
Subject:   Re: ***Best Reliable Multi-cast Protocol???***

hlester@relay.nswc.navy.mil (Hiram W. Lester Jr.) writes:
>Like so many before me, I am interested in a reliable multi-cast
>protocol.  I have briefly perused the rfc's and found numerous
>attempts at a solution to this problem, but does anyone know what
>the latest is?  Is there a comparison, say, of the best protocols
>that address this problem?  
An overview about all (or almost all) Multicast Transport
Protocols is available on the following WWW-page:
http://hill.lut.ac.uk/DS-Archive/MTP.html
bye
torsten
-- 
Torsten Kerschat - Interdepartmental Research Center for Process Control
		   Technical University of Berlin (PRZ - Room HE 104)
Internet: torsten@prz.tu-berlin.d400.de / landmark@cs.tu-berlin.d400.de
Phone   : ++49 / (0)30 / 314 - 26822    Fax: ++49 / (0)30 / 314 - 21114

-----------[000020][next][prev][last][first]----------------------------------------------------
Date:      1 Dec 1994 16:37:11 GMT
From:      gdiehl@gandalf.endicott.ibm.com (Gary Diehl)
To:        comp.protocols.tcp-ip
Subject:   Host handling of LF bits in 802.5 source routed frames

Does anyone know what is the most common method for handling the 3 "Largest Frame" bits that may come in on IEEE 802.5 source routed frames ?   Specifically, RFC 1042 says that the "receiver should compare the LF received with the MTU.  ...   If the LF is less than the MTU, the frame is rejected"

But then there are 3 options:

1) Reject the frame
2) For all remote hosts, reduce the effective MTU to the LF
3) Keep a separate MTU for each remote host

And, I guess a 4th option is just ignore the LF bits and accept the response.

Anyone know what is industry standard ?

Thanks,  Gary Diehl 

-- 
\begindata{text,538725120}
\textdsversion{12}
\template{plaintext}



\bold{Gary A. Diehl} 

AS/400 Divison

IBM Corporation

Endicott, New York   

Phone: 607-752-5505 or Tie Line: 852-5505.   FAX: 607-752-5421 

\italic{INTERNET: gdiehl@rchland.vnet.ibm.com           VNET: DIEHL@GDLVM6\
}\enddata{text,538725120}

-----------[000021][next][prev][last][first]----------------------------------------------------
Date:      Thu, 01 Dec 94 20:49:04 EDT
From:      Bill Bell <pp001529@interramp.com>
To:        vmsnet.networks.tcp-ip.ucx,comp.protocols.tcp-ip
Subject:   Dual Ethernets




What is the accepted way to switchover to another ethernet
card when the one you are connected with goes down?

We are using Alpha OSF/1 (UNIX) and are writing a
BSD socket TCP interface to transfer message across the
network.  We need to be able to "sense" when the connection
goes down and try the other connection.

1. How do you "sense" when one connection goes down?  Is there a way
   to know this without using a watchdog function that writes
   a byte every 30 seconds?  This has to work for when the ethernet goes 
   down and the connected host machine is unplugged.  Will select()
   handle this?

2. We figure on using a configuration file with the 2 hostnames
   to keep trying to connect() to.  Is there a better way?

3. I usderstand that accept() using INADDR_ANY handles incoming
   connections on either IP number. Is there anything I need to
   specifically configure?

4. When bind()ing to a port using INADDR_ANY, I get "port in use"
   users for 10 seconds or so if I hit ctrl-c and restart my
   app.  Is there anyway to reset this so I don't have to wait
   10 seconds ??  What is causing it?

Bill Bell


-----------[000022][next][prev][last][first]----------------------------------------------------
Date:      Thu, 01 Dec 94 21:08:50 EDT
From:      Bill Bell <pp001529@interramp.com>
To:        comp.unix.programmer,comp.protocols.tcp-ip
Subject:   Re: Address already in use message


In article <19941126.121601.95@comptech.demon.co.uk>, > >set the socket option 
"SO_REUSEADDR" (or something spelt very similarly)
> >on the socket, this should make that "problem" disappear.
> >
> 
> Wont that just make the symtom disappear? It wont get rid of the
> problems as such, ok so it may sort itself out after a few (10?) mins or
> so...
> 
> I did have something similar once, and think I cured it by doing a
> shutdown(s, 2) to dump pending data before closing the socket.
> 
> BTW, are you using SO_LINGER?
> 

I have read Rago and Stevens books and I cannot find a good explanation of the 
socket options.  WHat does SO_LINGER and SO_REUSEADDR do?




-----------[000023][next][prev][last][first]----------------------------------------------------
Date:      01 Dec 1994 18:11:26 GMT
From:      joost@cadlab.de (Michael Joosten)
To:        comp.protocols.tcp-ip
Subject:   Re: TCP connect() on non-blocking socket

In article <2354@sun3.IPSWITCH.COM> ddl@harvard.edu (Dan Lanciani) writes:


   | socket, but will the threeway-handshake continue?

   Indeed.  The connection process is usually already asynchornous and
   the blocking status of the socket merely controls whether your process
   will wait.  In other words, the connect call does approximately:

   if(already connected)
           return EISCONN
   if(already connecting)
           return EALREADY
   if(attempt to start connect fails)
           return reason
   if(socket non-blocking)
           return EINPROGRESS
   while(socket is connecting)
           sleep
   return status

A source-code example might be found in Mosaic's/CERN's libwww2/HTTCP.c. It's
at least in the Mosaic-2.4 distribution.

-- 
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| Michael Joosten   |       Tel.  : (+49) (+) 5251-284 168                 |
| CADLAB            |       Fax   : (+49) (+) 5251-284 140                 |
| Bahnhofstr. 32    |       E-Mail: joost@cadlab.de                        |
| 33102 Paderborn   |        	    ...!uunet!unido!cadlab!joost           | 
| FRG               | Mass mail to: joost@uni-paderborn.de                 |
|--------------------------------------------------------------------------|
| CADLAB is a cooperation between Uni-GH Paderborn & SNI AG                |
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

-----------[000024][next][prev][last][first]----------------------------------------------------
Date:      1 Dec 1994 20:13:48 GMT
From:      wsmargiassi@acm.org (William Smargiassi)
To:        comp.infosystems.www.users,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.infosystems.www.providers,comp.os.ms-windows.networking.tcp-
Subject:   Re: WWW Server/Client without INTERNET??

In article <3bkmgo$b0u@nntp.gov.ab.ca>, Gordon McAndrew <gordon@aec.env.gov.ab.ca> says:
>There is no requirement for an Internet connection to run a WWW server.
>But you probably have to be using the TCP/IP protocol on the LAN.  If you

Definitely. 

>have a SLIP server of some kind you will be able to dial in and use whatever
>services are on the LAN.  You will probably have to use the IP addresses
>directly or set up a DNS on the LAN.

What's wrong with hosts files? (If the network is small enough)

bill

-----------[000025][next][prev][last][first]----------------------------------------------------
Date:      Thu, 1 Dec 1994 20:31:39 GMT
From:      micoke@nucleus.com (Michael O'Keefe)
To:        vmsnet.infosystems.gopher,comp.protocols.tcp-ip.ibmpc,comp.protocols.tcp-ip.domains,comp.protocols.tcp-ip,comp.os.ms-windows.apps.comm,comp.infosystems,alt.gopher
Subject:   Re: Resolve Host Problem (IP Adrses)

GOD IM LOST: 
X-Newsreader: TIN [version 1.2 PL2]

: if anyone has a complete list of what it takes and simple 
:steps to follow to use mosaic send me mail at MicOKe@nucleus.com



-----------[000026][next][prev][last][first]----------------------------------------------------
Date:      Thu, 1 Dec 1994 22:33:14 GMT
From:      knaru@rxsvr2.cup.hp.com (Kim Naru (Contr))
To:        comp.infosystems.www.users,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.infosystems.www.providers,comp.os.ms-windows.networking.tcp-ip,comp.os.ms-windows.networking.misc
Subject:   Re: WWW Server/Client without INTERNET??

Madhu Rajamani (maad@iastate.edu) wrote:



: I need to run a World Wide Web server (on a PC or PC based network)
: within an organization which does NOT have internet access.  Also, I'd like
: to be able to dial in to the local network (or standalone PC) and access 
: HTML documents using MOSAIC or NETSCAPE.
 
: Has anyone ever tried something like this before?  Or if you have any ideas
: or sugestions as to how I can do it, please mail me or post here.
.
. deleted.
.

To get the WWW you need a network. The network does not need to be
connected to the Internet. I have a web server runnning on a lan
at home.

To dial in, you need to suscribe to a service(i.e American On-line,Netcom).


--kim





-----------[000027][next][prev][last][first]----------------------------------------------------
Date:      Thu, 1 Dec 1994 22:47:57 +0000
From:      Andrew@cucumber.demon.co.uk (Andrew Gabriel)
To:        comp.protocols.tcp-ip
Cc:        david_diamond@skymir.usc.edu
Subject:   Re: Sending SMTP mail...

In article <3bgcr4$2au@usc.edu>
           david_diamond@skymir.usc.edu "David Diamond" writes:

> I'm writing a database system that will send SMTP mail.  SO far, I've
> got it working properly except for a few things:
> 
> Small messages (less than 15 or so lines) work fine, but larger ones
> either just "abort" without any error or I get "message header line too
> long".  I'm sending all of the mail to a campus unix mail host which is
> where I'm getting that error from.

Are you remembering to leave a blank line between the header
and the body of the message? If not, I suspect the server chokes
on what it reagrds as an overly-long or garbage header, which is
actually part of your message text. It could also be a problem with
your line delimiters (causing the same effect) - should be CR LF.

> 
> The second question I have is regarding multiple addresses.  Is it
> neccesary to issue a "RCPT TO:" command for each address?  I was hoping
> to be able to send a entire list at once.  I read about the "EXPN"
> command, but I don't quite understand it.  I'm assuming it refers to a
> list that the mail host knows about and not a variable list such as I
> want to use.  

I'm a little less sure here (someone will correct me if I'm wrong), but
I believe separate RCPT TO:s are used so that when using several, they
can be individually accepted or rejected.

EXPN, as you say, returns the members of a mailing list. A lot of SMTP
servers don't provide it, because you might not like the contents of all
your mailing lists to be readable from anywhere. I'm not aware of any
SMTP clients which try to use it either, although there must be one
somewhere?

> 
> How do I refer to an address as a cc: or bcc:?

I will use an analogy with a real letter typed on paper.
The To: Cc: and Bcc: are the fields that you would typically get at the
top of the letter so all the recipients know who the other recipients
are. (Bcc:s wouldn't be included, but that's irrelevant).
When you post your letters, you put them in envelopes. The RCPT TO: is
the name you put on the outside of the envelope. You need to do this
whether the recipient is a To:, Cc:, or Bcc (and you don't distinguish).
If two of the recipients shared the same room, you could put both their
names on a single envelope, i.e. use more than one RCPT TO: - again it's
irrelevant whether they are a To: or a Cc: (a Bcc: gets a little
complicated in this simple analogy, but in principle the same is true).

> 
> David Diamond                                           ____
> Technology Services Coordinator                         \  /
> USC News Service - University of Southern California     \/
> (213) 740-2215                                            
> 

If your database is running under Unix, you should look at calling
/bin/mail or sendmail to do all this for you.
Similar considerations probably apply on other operating systems too,
i.e. look for a higher level API to mail than TCP.

-- 
Andrew Gabriel                        Home: Andrew@cucumber.demon.co.uk
                                      Work: Andrew.Gabriel@gpt.co.uk

-----------[000028][next][prev][last][first]----------------------------------------------------
Date:      2 Dec 1994 01:06:09 GMT
From:      kaat@xs1.xs4all.nl (kaat)
To:        comp.protocols.tcp-ip
Subject:   Re: bytes received from ping

Noah_Davids@vos.stratus.com wrote:

: I've noticed that the local copy of ping reports X+8 bytes received where
: X is the number of bytes that I requested it send. As far as I can tell the
: extra 8 bytes represent the ICMP header. I've looked at several systems
: and Unix Network Programming by R Stevens and they all report the same +8
: bytes.
 
: Is there a reason I am missing for including the ICMP header length in the
: bytes received count? If it includes the ICMP header why not the IP header?

I'll really like to know this too!
Currently I'm doing a project for school conserning TCP/IP.
One of the major questions I had was indeed why the extra 8 bytes are
there?

Is there anyone who can help us here?

thanks in advance,

Rob.

software engineering student, The Netherlands.     e-mail: kaat@xs4all.nl


-----------[000029][next][prev][last][first]----------------------------------------------------
Date:      Fri, 2 Dec 1994 01:32:37 +0000
From:      Chris@lunchbox.demon.co.uk (Chris Butterworth)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.unix.solaris,demon.ip.developers,demon.ip.support,demon.ip.support.pc,demon.ip.support.atari,demon.ip.support.unix
Subject:   Re: TCP Protocol problem Solaris<->KA9Q

In article <199411301302.NAA09562@ntl.com>
           tonymo@ntl.com "Tony Mountifield" writes:

  [much snipped]

> Here is an example of the failed connection from the Solaris client
> to the KA9Q server.
>     
 [SYN packet snipped]
>     
>     Tue Nov 29 21:38:30 1994 - sl0 sent:
>     PPP: len  82        protocol: IP
>     IP: len 78 193.132.77.63->193.132.77.1 ihl 20 ttl 63 prot TCP
>     TCP: 25->45214 Seq x112d8000 Ack x9456f792 ACK PSH SYN Wnd 11680 MSS 1460
>  Data 34
>     0000  ff 03 00 21 45 00 00 4e 00 16 00 00 3f 06 5e 4b  ...!E..N....?.^K
>     0010  c1 84 4d 3f c1 84 4d 01 00 19 b0 9e 11 2d 80 00  A.M?A.M...0..-..
>     0020  94 56 f7 92 60 1a 2d a0 e5 d5 00 00 02 04 05 b4  .Vw.`.- eU.....4
>     0030  32 32 30 20 73 6f 66 74 69 6e 73 2e 74 63 70 2e  220 softins.tcp.
>     0040  63 6f 2e 75 6b 20 53 4d 54 50 20 72 65 61 64 79  co.uk SMTP ready
>     0050  0d 0a                                            ..

   The above packet carries a SYN as well as some data. This is said to 
   bother some TCPs. In the example trace of a working connection I note
   that data and SYN were not present in the same packet.

   In the KA9Q version that I use (ibm PC) the option to turn this off
   is TCP SYNDATA OFF - this may exist in the version you are using.

   Hope this helps!


-- 
 Chris Butterworth                |  *   *  | "Everybody does it in the Zone"
 Mail: Chris@lunchbox.demon.co.uk |    |    | Hours: 9:30pm -> 1:00 am
       Chris@lunchbox.dircon.co.uk|   \_/   | telnet lunchbox.demon.co.uk 7777

-----------[000030][next][prev][last][first]----------------------------------------------------
Date:      1 Dec 1994 12:25:08 +0800
From:      cparaz@admu.edu.ph (Carlos Miguel Paraz)
To:        comp.protocols.tcp-ip
Subject:   WANTED: Automated FTP client with scripts

Hello,

I would like to have an FTP client which can take a list of sites and
filenames, and get them, logging in as anonymous and entering my 
email address as password.  This would ideally give me a log of 
the operations it made, and even schedule retries if the sites are full.

Any suggestions will be appreciated.
--
Miguel A. Paraz -=-=-=-=-=||-= "How far can this Net go?" =-=-=-=-=-=-=-=-=-=
cparaz@balut.admu.edu.ph =||-=  Ateneo de Manila University, Philippines   -=

-----------[000031][next][prev][last][first]----------------------------------------------------
Date:      2 Dec 1994 02:18:36 GMT
From:      snoack@calvin.bellahs.com (Scott Noack RD)
To:        comp.protocols.tcp-ip
Subject:   ESTABLISHED connection with no ends?

--Protocol gurus,

	Sometimes under Solaris 2.3 TCP connections remain ESTABLISHED (as
	noted by netstat) even though both ends have either shutdown or 
	died.  These orphane connections remain in this state forever (
	several days)!!!!!  Why aren't these connections cleaned up by the
	OS?  Is there a system call or funtion I can use to clean then
	up?

--Scott (snoack@bellahs.com)


-----------[000032][next][prev][last][first]----------------------------------------------------
Date:      Fri, 2 Dec 1994 08:51:44
From:      ace@iesl.forth.gr (Andreas C. Enotiadis)
To:        vmsnet.infosystems.gopher,comp.protocols.tcp-ip.ibmpc,comp.protocols.tcp-ip.domains,comp.protocols.tcp-ip,comp.os.ms-windows.apps.comm,comp.infosystems,alt.gopher
Subject:   Re: Resolve Host Problem (IP Adrses)

In article <3b3rdd$bci@cybersys.mercy.org> rputnam@cybersys.mercy.org writes:


>I'm using Chameleon 4.01 for Windows and am having a problem with domain 
 
>names.  I can't enter domain names in any of the programs (gopher, 
 
>telnet, FTP, etc...) and get it to work.  For some reason it can't find 
 
>the IP addresses when I put in the domain name.  This has been very 
 
>troublesome for me since I have a book of domain names -- but no IP 
 
>addresses!  Can anyone help me with this problem?  Here are the error 
 
>messages I get in each of the utilities for Chameleon:
 
>Telnet: "Unable to resolve host"

Seems to me you don't have a nameserver to tell you the IP address.
Typically you would have a machine you can connect to that will resolve your 
DNS queries.

Andreas

--------------------------------------------------------------------------
Andreas C. Enotiadis
Internetwork Ltd
"My views DEFINITELY represent those of InterNetwork Ltd - It's mine dammit!"
ace@ics.forth.gr
ace@iesl.forth.gr
ace@praxis.forth.gr
Snail-Mail :7 Fokidos Str, 11526 Athens, Greece. Tel : 7486222-3, Fax 7486223
--------------------------------------------------------------------------

-----------[000033][next][prev][last][first]----------------------------------------------------
Date:      Fri, 2 Dec 1994 20:30:38 -0800
From:      Robin Callender <robinc@halcyon.com>
To:        comp.protocols.tcp-ip
Subject:   Re: DHCP Info ???

On Wed, 23 Nov 1994, Frances K. Selkirk wrote:

> In article <3ap0gr$e37@cnj.digex.net> ajguido@cnj.digex.net (A.J. Guido) writes:
> 
> > I'm looking for any info on DHCP ( Dynamic Host Configuration 
> > Protocol). I have RFC1531, so I sort of know how the protocol 
> > works, but what I'm really after is real-world experience.  Has 
> > anyone successfully implemented DHCP? If so, with what products?
> 
> We have a DHCP server in our new Services OnNet product for Windows -
> I think it may still be the only commercial DHCP server available. In
> contrast, several (most?) of the current TCP/IP packages for DOS
> and/or Windows include DHCP clients. 
> 
> Enjoy,
> 
> --
> Frances K. Selkirk                                        fks@ftp.com
> FTP Software, Inc.    Technical Information Services   (800) 382-4FTP
> ---------------------------------------------------------------------
> FTP server = ftp.ftp.com        BBS =508-659-6240  |  support@ftp.com
> WWW server = http://www.ftp.com                    |  info@ftp.com
> 
> 
> 
I'm pretty sure (vis. know) that Win/NTAS has a DHCP server with it.  
Also I believe that HP has a DHCP server for their unix system.


---------------------------------------------------------------------------
 Robin Callender
 robinc@halcyon.com


-----------[000034][next][prev][last][first]----------------------------------------------------
Date:      2 Dec 94 12:38:12 WET
From:      norley@ccvax.ucd.ie
To:        comp.protocols.tcp-ip
Subject:   ddddd







-----------[000035][next][prev][last][first]----------------------------------------------------
Date:      2 Dec 1994 12:40:20 GMT
From:      mm1402@iunet.it (ANDREA SPINELLI)
To:        comp.protocols.tcp-ip,comp.sys.hp.apps,comp.sys.hp.misc
Subject:   POP3 server for HPUX

Hi everyone.
I am looking for a public-domain POP3 server for a HPUX system.

There is also the possibility that /etc/smtpd will do the job,
but I am pretty ignorant; will some kind soul answer to me?
I'll post summary, if I receive answer(s).

Thanx in advance

---------------------------------------------------------------------------
Andrea Spinelli			aspinelli@ismes.it
ISMES SpA			phone +39-35-307777
Via Pastrengo 20		fax   +39-35-307710
Seriate - Italy
---------------------------------------------------------------------------

-----------[000036][next][prev][last][first]----------------------------------------------------
Date:      Fri, 2 Dec 1994 13:07:25 GMT
From:      johannes@titan.westfalen.de (Johannes Stille)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.unix.solaris
Subject:   Re: TCP Protocol problem Solaris<->KA9Q

In article <199411301302.NAA09562@ntl.com>,
	Tony Mountifield <tonymo@ntl.com> wrote:
[...]
>
>The problem is that clients on the Solaris seem unable to make TCP connections
>to servers on the KA9Q, and complain of a protocol error. Other, BSD-based
>clients can connect to the KA9Q without problem, but I am told a WinSock
>client also couldn't connect to the KA9Q server.
 [...]
>    
>    Tue Nov 29 21:38:30 1994 - sl0 sent:
>    PPP: len  82	protocol: IP
>    IP: len 78 193.132.77.63->193.132.77.1 ihl 20 ttl 63 prot TCP
>    TCP: 25->45214 Seq x112d8000 Ack x9456f792 ACK PSH SYN Wnd 11680 MSS 1460 Data 34
>    0000  ff 03 00 21 45 00 00 4e 00 16 00 00 3f 06 5e 4b  ...!E..N....?.^K
>    0010  c1 84 4d 3f c1 84 4d 01 00 19 b0 9e 11 2d 80 00  A.M?A.M...0..-..
>    0020  94 56 f7 92 60 1a 2d a0 e5 d5 00 00 02 04 05 b4  .Vw.`.- eU.....4
>    0030  32 32 30 20 73 6f 66 74 69 6e 73 2e 74 63 70 2e  220 softins.tcp.
>    0040  63 6f 2e 75 6b 20 53 4d 54 50 20 72 65 61 64 79  co.uk SMTP ready
>    0050  0d 0a                                            ..
>    
 [...]
>
>That is the complete conversation. It appears that the client, having made
>the connection, doesn't like something, and closes it again.
>
>Any suggestions would be most welcome.

KA9Q sends data together with the SYN in the same packet.
AFAIK, this is quite uncommon and might not be handled correctly by
some implementations. So I'd assume this is what Solaris doesn't like.

	Johannes

-----------[000037][next][prev][last][first]----------------------------------------------------
Date:      2 Dec 1994 12:18:18 +0100
From:      sergior@tid.tid.es (Sergio Rodriguez Lombardero)
To:        comp.protocols.tcp-ip
Subject:   how TCP send unblocks without writing bytes at all ?

Hello.

I am starting with nonblocking sockets, and I'm  wondering if it is
possible that a send operation on a TCP socket unblocks without 
writing bytes at all, when there is no space in the buffer to hold 
all the bytes requested.

Thanks.


-----------[000038][next][prev][last][first]----------------------------------------------------
Date:      Fri, 02 Dec 94 22:14:43 PST
From:      zabin@unlisys.in-berlin.de (Hal Zabin)
To:        comp.protocols.tcp-ip
Subject:   Q: Optimal Trumpet Winsock config

Can anyone point me to some information on optimizing the settings in 
trumpet Winsock 2.0 for a SLIP connection over a modem?
Thanks, Hal


-----------[000039][next][prev][last][first]----------------------------------------------------
Date:      Fri, 2 Dec 1994 14:15:04 GMT
From:      iialan@iifeak.swan.ac.uk (Alan Cox)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.unix.solaris,demon.ip.developers,demon.ip.support,demon.ip.support.pc,demon.ip.support.atari,demon.ip.support.unix
Subject:   Re: TCP Protocol problem Solaris<->KA9Q

In article <199411301302.NAA09562@ntl.com> tonymo@ntl.com (Tony Mountifield) writes:
>I am trying to communicate between an Atari ST running the KA9Q NOS package
>and a PC running Sun's Solaris 2, using PPP. Mostly it works. The Solaris
>machine quite happily acts as a gateway, and has no problems accepting TCP
>connections from the KA9Q clients.

Solaris 2.1 on PC's has this urge to protocol error the piggy backed data
stuff done by KA9Q. Now KA9Q is stretching the spec a little, for a massive
performance gain for some operations on slow links (eg amateur radio), but
Solaris is violating the fundamental 'be liberal in what you accept' ethic.

You can either talk to sun or turn the piggy backing of data off on KA9Q.
I can't remember how you do that offhand. 

Alan
-- 
  ..-----------,,----------------------------,,----------------------------,,
 // Alan Cox  //  iialan@www.linux.org.uk   //  GW4PTS@GB7SWN.#45.GBR.EU  //
 ``----------'`--[Anti Kibozing Signature]-'`----------------------------''
One two three: Kibo, Lawyer, Refugee :: Green card, Compaq come read me...

-----------[000040][next][prev][last][first]----------------------------------------------------
Date:      2 Dec 1994 12:43:06 +0100
From:      magri@stud.uni-sb.de (Martin Grimsehl)
To:        comp.protocols.tcp-ip
Subject:   IP- forwarding

hi!


	is there a software packet, that does the folowing:


	bar> telnet foo 2000 	

        and foo:2000 forwards the TCP packet to an given server e.g. at
        myserver.de port 23



						cu martin


-----------[000041][next][prev][last][first]----------------------------------------------------
Date:      2 Dec 1994 14:48:26 GMT
From:      thed@mtek.chalmers.se (Thed Lyngmo)
To:        comp.protocols.tcp-ip
Subject:   Re: local address reuse

Robert Smith (robert@banana.dis.fedex.com) wrote:

: I have a problem with a simple client-server set of programs that are
: being used to test the behaviour of the system when the server side	
: is killed and then needs to be restarted. The problem in general is	
: that the local address stays in a TIME-WAIT condition when the server	
: side is killed. The client side detects connection is broken and exits	
: normally. When I try to restart the server, I get error: bind()
: (125Address already in use). I am unable to restart the server for	
: 4 minutes because of this condition. I have tried using
: setsockopt(SO_LINGER with l_linger=0) and setsockopt(SO_REUSEADDR) in
: both the server and the client as well as no setsockopt calls.

I think the netsoftware is waiting for any living IP-fragments to die
before you can use the same address again.
The time you'll have to wait should be MaxTTL * 2.
TTL = An IP fragments time to live in seconds (0-255).
(multiplied by 2 just to make sure)

I also think this is made to prevent messages sent in the first
connection to be received in the second.

I've found no way to circumvent this behaviour. Please let me know (email)
if you succeed in doing so.

Kind regards,
Ted Lyngmo

-----------[000042][next][prev][last][first]----------------------------------------------------
Date:      Fri, 2 Dec 1994 15:04:28 GMT
From:      zhebu@svusenet.ubs.ch
To:        comp.protocols.tcp-ip
Subject:   Re: IP- forwarding

: 	bar> telnet foo 2000 	
: 
:         and foo:2000 forwards the TCP packet to an given server e.g. at
:         myserver.de port 23

It is definitely not a good idea, but:

Yes, it is called telnet ;-). If you are in Unix, put an entry "myforward" 
in /etc/services and then put a an entry telnet telnet myserver.de 23 in 
your /etc/inetd.conf file:

/etc/services:
..
myforward	2000/tcp

/etc/inetd.conf:
..
myforward stream tcp nowait your_userid /usr/bin/telnet telnet myserver.de 23

You can't do the forwarding in the ip level, because you change the tcp port 
too. 

--
NAME   Urs Eberle
EMAIL  urs.eberle@zhflur.ubs.ubs.ch
PHONE  ++41-1-236-58-08

-----------[000043][next][prev][last][first]----------------------------------------------------
Date:      2 Dec 1994 15:41:44 GMT
From:      kfurge@ysi.com (Kenneth Furge)
To:        comp.protocols.tcp-ip
Subject:   Anyone added FTP's PASV command to KA9Q?

I am trying to add the PASV command to KA9Q for support
of direct FTP access via the Mosaic WWW browser.  I have 
been able to get it to successfully respond to the PASV
request, listen on the port and then connect.  But
when I try to send the file, Mosaic gives me an error and
the connection seems to timeout.  Any clues or partial
source would be greatly appreciated.  My current hacked up
version is based on the NOS-1229 source.  Thanks in advance.

K.C. Furge (N9FKC)
kfurge@ysi.com

-----------[000044][next][prev][last][first]----------------------------------------------------
Date:      Fri, 2 Dec 1994 15:52:43 GMT
From:      tsf@cix.compulink.co.uk ("Steve Scott")
To:        comp.protocols.tcp-ip
Subject:   Re: IP to X.25 Router ?

> Article: 14206 of comp.protocols.tcp-ip
> Newsgroups: comp.protocols.tcp-ip
> Path:
> cix.compulink.co.uk!uknet!sunsite.doc.ic.ac.uk!agate!howland.reston.ans.
> net!math.ohio-state.edu!caen!saimiri.primate.wisc.edu!aplcenmp!kslip4.ap
> l.jhu.edu!userFrom: bobley@mailstorm.dot.gov (Brett Bobley)
> Subject: IP to X.25 Router ?
> Message-ID: <bobley-3011941629490001@kslip4.apl.jhu.edu>
> Sender: usenet@aplcenmp.apl.jhu.edu
> Nntp-Posting-Host: kslip4.apl.jhu.edu
> Organization: US Coast Guard
> Date: Wed, 30 Nov 1994 21:30:09 GMT
> Lines: 39
> > Hi, folks.  Can anyone out there recommend a router (I think that is
 whatI need) for connecting my IP ethernet network to X.25 ?
> > Let me explain some details:
> > Currently, we have a 56K X.25 SprintNet line coming into our 
building. > Itis connected to a Sprint TP-4 switch which has a bank of 
4800 bps
> short-haul modems connected to it.  Then each of these short-haul modems
> is then wired into our phone system so that it can be routed to offices
> onvarious floors.  Each office has their own X.25 "drop" (essentially a
> Motorola/Codex X.25 modem connected up to the Sprint X.25 switch).
> > Under the current situation, I can "dial" out the X.25 drop by
> specifyingthe X.25 address of a remote computer.  However, the drawback
> of thissystem is that the connection between my office and the X.25
> switch isonly 4800 bps.  This is a major bottleneck, particularly if I'm
> running anumber of VC's via the connection.
> > My question is:  is it necessary that we connect the 56K line to these
> short-haul modems?  Would it be possible instead to connect the incoming
> Sprintnet 56K line to some sort of IP/X.25 router?  That way, if I
> wantedto connect to a remote machine, I could just telnet to its domain
> name. This IP/X.25 router would have to be smart enough to map the IP
> address tothe proper X.25 address to that it can set up the virtual
> ciruit to theremote location.
> > By doing this, I'd be taking better advantage of the 56K bandwidth and
> notlimiting myself to 4800 bps.  But is what I just described 
 possible? > Canyou recommend any companies which make such a router?
> > tks,
> > Brett
> > -- > Brett Bobley
> U.S. Coast Guard
> Washington, DC
> Internet: bobley@mailstorm.dot.gov

Software Forge Ltd in the UK (part of the Racal-Airtech group) do a 
LAN-X.25 IP Router that runs in a PC.  The solution includes a sync. 
comms adapter which goes in the PC and can support V.24, X.21 and ISDN 
interfaces to the WAN.  Mapping from IP addresses to X.25 NUAs is via a 
user-configured table.  Mail me for details.


Robin Maidment, Software Forge (Racal-Airtech group)
email:   tsf@cix.compulink.co.uk
phone:   UK (0)734 312477
fax:     UK (0)734 311301

-----------[000045][next][prev][last][first]----------------------------------------------------
Date:      2 Dec 1994 16:06:37 GMT
From:      tjioe@sfsuvax1.sfsu.edu (Joe Chitrady)
To:        comp.protocols.tcp-ip
Subject:   The real meaning ?

Whenever somebody writes in the resume that he/she knows tcp/ip.
What is this really mean ? Is this means he/she know about telnet
ftp etc or KNOWING the DETAIL of how to write program using the
protocol ?

I am just wondering about the above because at least me and my
co-worker did not agree on this.

Joe


-----------[000046][next][prev][last][first]----------------------------------------------------
Date:      2 Dec 1994 16:09:57 GMT
From:      Jeff Porter <jeffp@farallon.com>
To:        comp.protocols.tcp-ip
Subject:   TCP-IP users may qualify for a free twin pak of Farallon's Timbuktu Pro.

If you are currently using TCP/IP on Macs & PC's Farallon will send you a free Timbuktu Pro Twin Pak if you are one of the first 100 respondents to the survey below.

 It's easy. Just answer the following five questions. Be sure to fill out your name, company, and address at the bottom and return it via E-mail to jeffp@farallon.com. Farallon will send the software to your address within six to eight weeks if your one of the first 100 respondents. This offer is limited to one Timbuktu Pro for Macintosh Twin Pak per company or organization.
Farallon Survey


What Mac and PC TCP/IP packages are you using?

What other protocols do you use?

How many networked computers are in your organization?
PC's_____            Mac's______

What WAN/Dial-In products do you use (i.e.PPP,SLIP, ARA)?

What remote control products do you use?


Name:_________________________________
Title___________________________________

Company/Organization_______________________________
Address___________________________________________
City, ST,Zip_________________________________________
Phone______________________

Thank you for your time and attention.

Sincerely,

Jeff Porter
Product Marketing Manager
Voice (510) 814-5361
jeffp@farallon.com


-----------[000047][next][prev][last][first]----------------------------------------------------
Date:      2 Dec 1994 18:20:20 GMT
From:      Charlie Rizzuto <cjrizzuto@ets.org>
To:        comp.protocols.tcp-ip
Subject:   MVS TELNET Application

Is there any software available for MVS TCP/IP that will let me
telnet to a UNIX machine and act like a VT type terminal?

If so, can you tell me where to find it?

Thanks.

Charlie Rizzuto
Educational Testing Service
(cjrizzuto@ets.org)

-----------[000048][next][prev][last][first]----------------------------------------------------
Date:      2 Dec 1994 16:44:16 +0100
From:      teisa@beatles.cselt.stet.it (Claudio Teisa)
To:        comp.protocols.tcp-ip
Subject:   questions on RFC1388 (RIP-2)

I have two questions related to RFC1388:

1) Routing Domains -

Where can I find an example of their usage?

2) Appendix A

(from the RFC ...

   This is a simple example of the use of the next hop field in a rip
   entry.

      -----   -----   -----           -----   -----   -----
      |IR1|   |IR2|   |IR3|           |XR1|   |XR2|   |XR3|
      --+--   --+--   --+--           --+--   --+--   --+--
        |       |       |               |       |       |
      --+-------+-------+---------------+-------+-------+--
        <-------------RIP-2------------->

   Assume that IR1, IR2, and IR3 are all "internal" routers which are
   under one administration (e.g., a campus) which has elected to use
   RIP-2 as its IGP. XR1, XR2, and XR3, on the other hand, are under
   separate administration (e.g., a regional network, of which the
   campus is a member) and are using some other routing protocol (e.g.,
   OSPF).  XR1, XR2, and XR3 exchange routing information among
   themselves such that they know that the best routes to networks N1
   and N2 are via XR1, to N3, N4, and N5 are via XR2, and to N6 and N7
   are via XR3. By setting the Next Hop field correctly (to XR2 for
   N3/N4/N5, to XR3 for N6/N7), only XR1 need exchange RIP-2 routes with
   IR1/IR2/IR3 for routing to occur without additional hops through XR1.
   Without the Next Hop (for example, if RIP-1 were used) it would be

... end)

The example is quite clear on what could be the benefits of the Next Hop
indication in the IR1/2/3. What I can't understand is the relationship
between the OSPF routers and the RIP-2 routers.
It seems that XR1 runs both OSPF and RIP-2, how is this possible?

RFC1247 (OSPF version 2) states that each Autonomous System has a single IGP
and this would mean that IR1/2/3 and XR1/2/3 (as they use different IGPs) 
should be in different AS; in this case they should communicate using EGP 
or BGP, shouldn't them? 

Thank's fro your help!

                                       Claudio T.

--------------------------------------------------------------------------
teisa@beatles.cselt.stet.it



-----------[000049][next][prev][last][first]----------------------------------------------------
Date:      2 Dec 1994 20:40:30 GMT
From:      mcmillan@cps.msu.edu (Eric Marc Mcmillan)
To:        comp.protocols.tcp-ip
Subject:   TCP/IP FAQ

Could somebody send me a tcp/ip faq or
post it.


-----------[000050][next][prev][last][first]----------------------------------------------------
Date:      02 Dec 1994 20:56:24 GMT
From:      croten@eosqa.gsfc.nasa.gov (Charles Roten)
To:        comp.unix.solaris,comp.sys.sun.admin,comp.protocols.tcp-ip
Subject:   Under SunOS 5.3, where to set broadcast address and netmask ??

I'm setting up a SPARCserver 1000 at GSFC, running SunOS 5.3.  Everyything 
is Jake now but the connection to the local network.  Running 'ifconfig le0' 
from the command line tells me why .. the broadcast address is hosed.  

Not too surprising .. it _is_ a new system.  

Under 4.1.3, for a Sun 4/260, I put the line 

    ifconfig ie0 128.183.117.34 netmask 255.255.255.o broadcast 128.183.117.0

in /etc/rc.local.  In SGI System V, /etc/config/ifconfig-1.options reads 

    broadcast 192.107.190.191
    netmask 255.255.255.192

and the same job gets done.  

But Stunned MicrobeSystems doesn't _use_ straight Sys V R4.  So _where_, 
for Pete's sake, do I set the netmask ?  And _HOW_ ??  I've fgrepped the 
whole bloody /etc tree for 'broadcast' and 'netmask'.  Nothing productive.  
'ifconfig' only shows up in a file, /etc/init.d/inetsvc, which seems to 
only be used for NIS configuration.  We don't _use_ NIS here, we use DNS, 
so the file's a template on another Sun running 5.3 at this site.  I'm at 
the end of my rope .. how the _HELL_ do I get the netmask and broadcast 
address right on boot-up ??  

--
Charles D. Roten               | Hughes STX Inc.
croten@nyx.cs.du.edu           | NASA GSFC (Hurrah DAAC!)
croten@eosdata.gsfc.nasa.gov   | (301) 286-4413 (w), (301) 317-0872 (h)

-----------[000051][next][prev][last][first]----------------------------------------------------
Date:      Fri, 2 Dec 1994 21:01:24 GMT
From:      Keith McFarlane <nunuv@your.biz>
To:        comp.protocols.tcp-ip
Subject:   MBONE software for Windows?

Has anyone heard of MBONE drivers/clients/etc. for MS Windows 3.11
and/or NT? Our site is getting hooked up and we have a significant
PC presence.

Keith McFarlane

-----------[000052][next][prev][last][first]----------------------------------------------------
Date:      Fri, 2 Dec 1994 21:44:16 GMT
From:      barnett@grymoire.crd.ge.com (Bruce Barnett)
To:        comp.protocols.tcp-ip
Subject:   Re: FTP, ASCII vs Binary

In article <porter-261194211143@f181-139.net.wisc.edu> porter@chem.wisc.edu (ron porter) writes:

>   I know the difference between binary and ASCII, but have
> been wondering, is it okay to transfer an ASCII file in Binary?
> Maybe you wouldn't do this knowingly, but if you don't know
> or don't change it back to ASCII.


I have done some studies, and in some cases, transfering files in
ASCII takes twice as long as BINARY.

If you select ASCII, the host has to read the entire file, searching
for special characters. If you select BINARY, the file is not
scanned for these special characters.



If the machines are both UNIX, use binary.
I'd say use BINARY whenever you can. 
--
Bruce Barnett <barnett@crd.ge.com> uunet!crdras!barnett

-----------[000053][next][prev][last][first]----------------------------------------------------
Date:      2 Dec 1994 23:46:00 GMT
From:      bsmith@wci.com
To:        comp.protocols.tcp-ip
Subject:   SVR4.2 ping source addr

How does the SVR4.2 ping command choose the source IP address
on a machine with multiple interfaces?  Is there any way to 
control which interface is used as the source?

Thanks!
-- 
Bob Smith
Wireless Connect, Inc
2177 Augusta Place,  Santa Clara, CA  95051-1714
Voice: (408) 296-1546    FAX: (408) 296-1547

-----------[000054][next][prev][last][first]----------------------------------------------------
Date:      03 Dec 1994 00:48:43 GMT
From:      pxl@bifrost.Lanl.GOV (Peter Lomdahl)
To:        comp.unix.solaris,comp.sys.sun.admin,comp.protocols.tcp-ip
Subject:   Re: Under SunOS 5.3, where to set broadcast address and netmask ??

"Charles" == Charles Roten <croten@eosqa.gsfc.nasa.gov> writes:
........[stuff deleted]
Charles> But Stunned MicrobeSystems doesn't _use_ straight Sys V R4.
Charles> So _where_, for Pete's sake, do I set the netmask ?  And
Charles> _HOW_ ??  I've fgrepped the whole bloody /etc tree for
Charles> 'broadcast' and 'netmask'.  Nothing productive.  'ifconfig'
Charles> only shows up in a file, /etc/init.d/inetsvc, which seems to
Charles> only be used for NIS configuration.  We don't _use_ NIS here,
Charles> we use DNS, so the file's a template on another Sun running
Charles> 5.3 at this site.  I'm at the end of my rope .. how the
Charles> _HELL_ do I get the netmask and broadcast address right on
Charles> boot-up ??

Commenting out the NIS based ifconfig line in /etc/init.d/inetsvc and
running with:

/usr/sbin/ifconfig nf0 `uname -n` -trailers up netmask 0xffffff00 broadcast 0x80b716ff

works fine for us at boot-up.

	-Peter

--
-- Peter Lomdahl -- T-11, MS B262 --
-- Los Alamos National Laboratory --
-- Los Alamos,  New Mexico  87545 --
-- pxl@lanl.gov -- (505) 665-0461 --

-----------[000055][next][prev][last][first]----------------------------------------------------
Date:      3 Dec 1994 05:46:41 GMT
From:      Neil Butani <unbutan@pacbell.com>
To:        comp.protocols.tcp-ip
Subject:   RFC1006

Where can I find the source for RFC-1006 ?

Thanks in advance
Neil Butani
unbutan@pacbell.com

-----------[000056][next][prev][last][first]----------------------------------------------------
Date:      Sat, 3 Dec 1994 06:23:15 GMT
From:      adam@comptech.demon.co.uk (Adam Goodfellow)
To:        comp.protocols.tcp-ip
Subject:   Re: bytes received from ping

In article <3blru1$91o@news.xs4all.nl>,
                kaat <kaat@xs1.xs4all.nl> wrote:

>Noah_Davids@vos.stratus.com wrote:
>
>: I've noticed that the local copy of ping reports X+8 bytes received where
>: X is the number of bytes that I requested it send. As far as I can tell the
>: extra 8 bytes represent the ICMP header. I've looked at several systems
>: and Unix Network Programming by R Stevens and they all report the same +8
>: bytes.
 
>: Is there a reason I am missing for including the ICMP header length in the
>: bytes received count? If it includes the ICMP header why not the IP header?
>
>I'll really like to know this too!
>Currently I'm doing a project for school conserning TCP/IP.
>One of the major questions I had was indeed why the extra 8 bytes are
>there?
>


Have just spent the last half hour porting the BSD pinger this over to
my system.

Anyway - from the ping man page:

----

ICMP PACKET DETAILS

An IP header without options is 20 bytes. An ICMP ECHO_REQUEST packet
contains an additional 8 bytes worth of ICMP header followed by an
arbitrary amount of data. When a packetsize is given, this indicates the
size of this extra piece of data (the default is 56). Thus the amount of
data received inside of an IP packet of type ICMP ECHO_REPLY will always
be 8 bytes more than the requested data space (the ICMP header).

If the data space is at least eight bytes large, ping uses the first
eight bytes of this space to include a timestamp which it uses in the
computation of round trip times. If less than eight bytes of pad are
specified, no round trip times are given.

----

Hope this helps.

-- 
Adam

=======================================================================
| Computech  Tel/Fax: 0181 673 7817  email: adam@comptech.demon.co.uk |
=======================================================================

-----------[000057][next][prev][last][first]----------------------------------------------------
Date:      3 Dec 1994 17:57:13 -0500
From:      mcknight@rbdc.rbdc.com (Michael Mcknight)
To:        comp.protocols.tcp-ip
Subject:   Multiple NICs on a server on same segment


We have the need to connect a Sun server to a very busy network.  The idea
so far has been to connect the users and servers, etc. to an etherswitch
to manage the traffic better.  Here's the idea.

We want to have a wide pipeline going from the switch to the server... say
4 30MB pipes.  We want the server to logically be on 1 (one, uno) segment.
I wont go into all of the reasons why, but my question is this.. can we?

For example:

      |--------|               
      |        |---------------|==========|     (no need to draw users)
      | SPARC  |---------------| EtherNet |-------------^
      | SERVER |---------------|  Switch  |-----------------^
      |        |---------------|==========|
      |--------|      ^
                      |-- Cat-5 UTP 10-Base-T Ethernet Cables

Lets say the SPARC has 4 10BT ethernet cards numbered and named as follows:
	le0		sun1a	155.125.10.1
	le1		sun1b	155.125.10.2
	le2		sun1c	155.125.10.3
	le3		sun1d	155.125.10.4

All NICs would be on the same logical & physical network.

Is this a problem?  We invision the clients simply connecting to a different
card depending on their use/location.  This is different than segmentation 
in that we are not really segmenting the network.

It seems like the sun would simply see each NIC as a different card and
not really care about what network it is on.  The users would simply conect
to a specified card without having to really care where it is.

I've been told this can be done and that it cannot be done.  I can see it
as a problem if we were trying to route between cards, but in this case,
we are not routing between them.

If it cannot work, why?  Is it a limit of TCP/IP or the SPARC system or the
SunOS?

As much detail as possible (or a refence document) would be greatly
appreciated.

Please Email all replies.
Thanks to all in advance.
 
-Michael McKnight
 mcknight@rbdc.rbdc.com

-----------[000058][next][prev][last][first]----------------------------------------------------
Date:      3 Dec 1994 13:56:50 GMT
From:      di04@hrz.th-darmstadt.de (A. Schindler)
To:        comp.protocols.tcp-ip
Subject:   Re: Wanted TCP/IP terminal server with one serial port

Bengt Bredenberg (bjb@esmiodt.ahlstrom.com) wrote:
: Perhaps this is not the right forum for requests like this (please inform me 
: which could be the right group), but I have been looking for many years in 
: Finland (without luck) for a product with the following specification:
: - TCP/IP terminal server with
: - Ethernet (thin, TP) connection (Token Ring) 
: - one RS232C serial port (9600 Bd)
: - low price and small size
 
: Devices like this are usually sold with 8 or 16 ports (Bridge etc) and are 
: expensive and big.  Similar devices used as network printer adapters are also 
: available, but these usually include only a parallell port and are not 
: bidirectional.
:  
: What I am looking for is a commercial small cheap product, which I could put 
: inside our concentrator to handle the serial communication to the network.
: The device should be completely bidirectional without any delays in either 
: directions. The TCP/IP address, port number and serial speed should be 
: programmable. (A fixed port number such as 23/telnet can be tolerated).
: If anyone of you have heard or used a device like this could you please inform 
: me so I can contact the right sources.
 
: Bengt Bredenberg
: Oy Esmi Ab, Finland
: email: bjb@esmiodt.ahlstrom.com

There is a Terminal/Print-Server sold by the Californian company LANTRONIX
distributed in Germany by TRANSTEC in Tuebingen. May be contacted via
Internet mail using transtec@transtec.de or mail@transtec.de

The server provides 2 ports, one serial and one parallel (centronics) and is
thus very well suited for a remote terminal/printer station. The device 
is able to autoconfigure but may also be set-up via bootp. It understands
besides TCP/IP the protocols IPX and LAT (DEC). The size is about 2x a cigarettebox, full metal enclosure with separate power supply. The network is attached
by either AUI (15 pin) or twisted-pair. The same vendor supplies a plug on
transceiver.  We use serveral of these for about 1/2 year, no problems. The
device needs no host based software support, but comes with a reverse-telnet
package (sources !) and DOS/Novell test programs on diskette. Prices are
about DM 950,- ($ 600) for the entire package and about DM 95,- ($ 60) for
the thin-wire transceiver.

-----------[000059][next][prev][last][first]----------------------------------------------------
Date:      3 Dec 1994 15:24:39 GMT
From:      drew@taco.cc.ncsu.edu (Drew Wilhite)
To:        comp.protocols.tcp-ip,comp.unix.sys5.r4
Subject:   Slip Error msg


Does anyone know the meaning of the following error message?

WARNING: ipintr: pullupmsg failed

It occurs in sporadic bunchs on 2 machines that are connected by SLIP.
(Both running Dell sysvr4 issue 2.2). 

I apologize if this is a dumb question...

Drew Wilhite
ncsipm1.cropsci.ncsu.edu

-----------[000060][next][prev][last][first]----------------------------------------------------
Date:      3 Dec 1994 17:12:02 GMT
From:      slack@earth (Charles Slack)
To:        comp.protocols.tcp-ip,comp.unix.programmer
Subject:   Re: Address already in use message

This is a common 'bug' for merc/diku muds <grin>.  Perhaps you should ask
for the exact fix on r.g.m.d?

--
/*
 * Charles C. Slack  (slack@usa.net) 
 */

-----------[000061][next][prev][last][first]----------------------------------------------------
Date:      Sat, 3 Dec 1994 18:38:12 GMT
From:      bobley@mailstorm.dot.gov (Brett Bobley)
To:        comp.protocols.tcp-ip,comp.sys.hp.apps,comp.sys.hp.misc
Subject:   Re: POP3 server for HPUX

In article <3bn4jk$opq@sgi.iunet.it>, mm1402@iunet.it (ANDREA SPINELLI) wrote:

> Hi everyone.
> I am looking for a public-domain POP3 server for a HPUX system.
> 
> There is also the possibility that /etc/smtpd will do the job,
> but I am pretty ignorant; will some kind soul answer to me?
> I'll post summary, if I receive answer(s).
> 
> Thanx in advance
> 
> ---------------------------------------------------------------------------
> Andrea Spinelli                 aspinelli@ismes.it
> ISMES SpA                       phone +39-35-307777
> Via Pastrengo 20                fax   +39-35-307710
> Seriate - Italy
> ---------------------------------------------------------------------------

Try FTPing to:

 hpux.csc.liv.ac.uk (formerly ftp.csc.liv.ac.uk) and it's
              mirrors world-wide,

        and iworks.ecn.uiowa.edu

They contain lots of HP-UX software including POP3.

Brett

-- 
Brett Bobley
U.S. Coast Guard
Washington, DC
Internet: bobley@mailstorm.dot.gov

-----------[000062][next][prev][last][first]----------------------------------------------------
Date:      3 Dec 1994 19:16:13 GMT
From:      mrider@ix.netcom.com (Mike Rider)
To:        comp.protocols.tcp-ip
Subject:   Re: DHCP

In <3bj8bh$1c7@hubcap.clemson.edu> gberger@eng.clemson.edu (Gary L. 
Berger) writes: 

>
>
>
>---
>Can anyone tell me about DHCP on NT advanced server?
>--------------------------------

Microsoft has a white paper on DHCP and WINS available on their WWW
server:   http://www.microsoft.com

Mike Rider



-----------[000063][next][prev][last][first]----------------------------------------------------
Date:      3 Dec 1994 22:57:29 GMT
From:      broswell@ix.netcom.com (Bob Roswell)
To:        comp.protocols.tcp-ip
Subject:   2 IP address for 1 DOS Machine?  Is it Possible?

We have two UNIX boxes, one on the Internet and another which is for
internal Use.  At my workstation, (I'm currently using Lan WorkPlace for
DOS)  I can configure LWP for the in-house Class C address or for our 
Internet class C address, but not both at the same time.  We don't want 
routing between the two for security reasons.

I am currently using two cards to solve the problem.  Does anyone have
a better suggestion?


-----------[000064][next][prev][last][first]----------------------------------------------------
Date:      3 Dec 1994 21:31:45 +0100
From:      casper@fwi.uva.nl (Casper H.S. Dik)
To:        comp.unix.solaris,comp.sys.sun.admin,comp.protocols.tcp-ip
Subject:   Re: Under SunOS 5.3, where to set broadcast address and netmask ??

croten@eosqa.gsfc.nasa.gov (Charles Roten) writes:

>But Stunned MicrobeSystems doesn't _use_ straight Sys V R4.  So _where_, 
>for Pete's sake, do I set the netmask ?  And _HOW_ ??  I've fgrepped the 
>whole bloody /etc tree for 'broadcast' and 'netmask'.  Nothing productive.  
>'ifconfig' only shows up in a file, /etc/init.d/inetsvc, which seems to 
>only be used for NIS configuration.  We don't _use_ NIS here, we use DNS, 
>so the file's a template on another Sun running 5.3 at this site.  I'm at 
>the end of my rope .. how the _HELL_ do I get the netmask and broadcast 
>address right on boot-up ??  


You set the netmask in /etc/netmasks.

Casper

-----------[000065][next][prev][last][first]----------------------------------------------------
Date:      4 Dec 1994 02:07:08 GMT
From:      myast3@icarus.lis.pitt.edu (Muh-rong Yang)
To:        comp.protocols.tcp-ip
Subject:   Internet Infrastructure

Dear Nets:
  I would like to know current Internet infrastructrue?  Is it built by
several regional networks and connected one another?  or else ?  Is any
Map I can find via annonymous ftp?  Thanks!

muh-rong 
E-mail: myast3@icarus.lis.pitt.edu
  

-----------[000066][next][prev][last][first]----------------------------------------------------
Date:      Sun, 04 Dec 94 02:15:12 GMT
From:      clyde@hitech.com.au (Clyde Smith-Stubbs)
To:        comp.protocols.tcp-ip
Subject:   POP server for SunOS?

Does anyone know of a POP server that will run under SunOs?

TIA.
--
 Clyde Smith-Stubbs       | HI-TECH Software,       | Voice: +61 7 300 5011
 clyde@hitech.com.au      | P.O. Box 103, Alderley, | Fax:   +61 7 300 5246
 ...!nwnexus!hitech!clyde | QLD, 4051, AUSTRALIA.   | BBS:   +61 7 300 5235
                    HI-TECH C: Compiling the real world...

-----------[000067][next][prev][last][first]----------------------------------------------------
Date:      Sun, 4 Dec 1994 04:29:14 GMT
From:      bpeoples@iglou.iglou.com (Ben Peoples)
To:        comp.protocols.tcp-ip
Subject:   PPP via telnet...

Is it possible to run a ppp (or slip for that matter) connection over a 
telnet connection?  

					Thanks, 
					  Ben

-- 
Ben Peoples					       	bpeoples@iglou.com
Unless otherwise stated, the above opinions are ***WORDS***
"As I said, with this net I will catch them, I bet.  With this net I will catch
those things, yet!"  --- _The Cat in The Hat_

-----------[000068][next][prev][last][first]----------------------------------------------------
Date:      Sun,  4 Dec 1994 07:08:00 GMT
From:      carl.shackford@nuilink.com (Carl Shackford)
To:        comp.protocols.tcp-ip
Subject:   SPX/IPX to TCP/IP Gateway

AIS>Path: kaiwan.com!UB.com!wetware!barrnet.net!agate!howland.reston.ans.net!ne
AIS>From: aisg@gate.net (Advanced Information Systems Group)
AIS>Newsgroups: comp.protocols.tcp-ip
AIS>Subject: SPX/IPX to TCP/IP Gateway
AIS>Date: 22 Nov 1994 22:18:43 GMT
AIS>Lines: 16
AIS>Message-ID: <3atqo3$l5a@tequesta.gate.net>
AIS>NNTP-Posting-Host: inca.gate.net
AIS>X-Newsreader: TIN [version 1.2 PL2]
AIS>Xref: kaiwan.com comp.protocols.tcp-ip:36260


AIS>Has anyone heard about FEL Computing's LANlink TCP product? It acts as a
AIS>gateway for Novel connected PCs to TCP/IP networks. It enables pooling TCP/
AIS>licesnses and negates the need for TCP/IP on the desktop. Client software i
AIS>loaded on the PCs, whioch communicate with the gateway via IPX. The clients
AIS>include telnet, ftp and a winsock driver.

AIS>There is also a simular product by IPSwitch. Any comments would be
AIS>appreciated. Is this a sound approach?


AIS>Thanks,

AIS>John Klann
AIS>407-774-7181
AIS>klann@advinfo.com

John;

I hope to be evaluating a product from firefox called NOVip (i think).
My company is getting into the TCP/IP world and I am hunting for a way
to centrally administer IP address on Novell LANS. The product looks
promising though a bit costly to start with.

The product is an NLM that allows sharing of address without protocol
stacks, IPX is used (sounds like a gateway). This could save me and my
fellows a tremendous amount of time and bookkeeping. By the way with
that phone # you may Know one of my companies as Sea World.

Carl Shackford
BEC Information Systems
314-957-4845
NUINET

---
 * SLMR 2.1a * from the kitchen table to your screen
 * St. Louis NetWare Users Group, St. Louis, MO. (314)544-8948

-----------[000069][next][prev][last][first]----------------------------------------------------
Date:      Sun, 4 Dec 1994 08:13:41 GMT
From:      herbr@netcom.com (Herb Rosenberg)
To:        comp.protocols.tcp-ip
Subject:   Setting up DHCP help?

Subject: Setting up DHCP help?
Newsgroups: comp.unix.admin
Organization: NETCOM On-line Communication Services (408 261-4700 guest)
Summary: 
Keywords: 

Does anyone out there have any information on how to set up a DHCP server 
on a Sun or RS6000 server?  If so, I would appreacite any comments or 
suggestions.

Thanks.



-- 
herbr@netcom.com

-----------[000070][next][prev][last][first]----------------------------------------------------
Date:      4 Dec 1994 22:04:17 -0500
From:      barmar@nic.near.net (Barry Margolin)
To:        comp.protocols.tcp-ip
Subject:   Re: The real meaning ?

In article <3bngme$jfa@news.csus.edu> tjioe@sfsuvax1.sfsu.edu (Joe Chitrady) writes:
>Whenever somebody writes in the resume that he/she knows tcp/ip.
>What is this really mean ? Is this means he/she know about telnet
>ftp etc or KNOWING the DETAIL of how to write program using the
>protocol ?

It means whatever the person who wrote it wants it to mean, since TCP/IP is
a very general thing to "know".  It probably depends on the context as
well; for a sysadmin, it probably means they know how to administer TCP/IP
networks, while for a programmer it should mean that they know how to write
TCP/IP applications.  And if they're applying for a job at a router vendor
it should probably mean that they know the low-level details of the
protocols.
-- 

Barry Margolin
BBN Internet Services Corp.
barmar@near.net

-----------[000071][next][prev][last][first]----------------------------------------------------
Date:      4 Dec 1994 15:55:21 GMT
From:      merle@cube.net (Anselm Schaefer)
To:        comp.protocols.tcp-ip
Subject:   is it possible to share modems under tcp-ip?

can I share a modem in a tcp-ip network? if so, how?
I know that I telnet on a host and use his modem. But I want to use his
modem without having to login on some other host.

thanks for any advice, merle



-----------[000072][next][prev][last][first]----------------------------------------------------
Date:      04 Dec 94 21:26:04
From:      jhk_utrcv1_res_utc_com@holt-systems.com (jhk@utrcv1.res.utc.com)
To:        comp.protocols.tcp-ip
Subject:   Need info on multicast routing protocols

From: jhk@utrcv1.res.utc.com
Subject: Need info on multicast routing protocols
Organization: UTRC

Does anyone have info on where to find RFCs and Internet Drafts on
multicast routing protocols.  I'm looking for DVMRP, MOSPF, and especially
PIM documentation.

Thanks,
Jim Kilroy
jhk@utrc.utc.com
--
|Fidonet:  jhk@utrcv1.res.utc.com 1:3609/504
|Internet: jhk_utrcv1_res_utc_com@holt-systems.com
|
| Standard disclaimer: The views of this user are strictly his own.
 

-----------[000073][next][prev][last][first]----------------------------------------------------
Date:      5 Dec 1994 05:55:41 -0800
From:      cgi@crl.com (Paul Smith)
To:        comp.protocols.tcp-ip
Subject:   Shutdown() block proof?

I'm trying to craft a sockets wrapper library that will not block indeffinately
on a close of a TCP service NOR interact with a child task's use of the fd
after a perent server forks (concurrant TCP server).

1) I've set SO_LINGER to 10 seconds, for a reasonalby short pause for the
other end to read out standing packets.

2) I was setting the fd to O_NONBLOCK prior to calling close to avoid potential
indefinite blocks..

I've discovered the following architectural interactions:

1) For a TCP concurrant server, where the main server forks on every accept()
then closes the new fd, the child closes the main fd.  There can be interaction
from both tasks acting on the same fd control block in the kernel if the library
I wrote sets the fd to O-NONBLOCK prior to calling close. 

In the case of the parent closing the new client accept() fd, it sets the fd
to O_NONBLOCK and the child task's use of this socket, then becomes non-blocking.
OK, ok, I didn't think of that when I coded my close library function.  Then how
to I craft a robust close wrapper given that it will be used for both
sequential and concurrant server and client environments???

Do I need two (or more) close wrapper functions; 1 for orderly shutdown; SO_LINGER
set to some reasonable short time, set to O_NONBLOCK ?? (or is this at odds with
SO_LINGER??) then close.  2) don't fool with it close, just call close, DON't
call shutdown, or set to O_NONBLOCK.

I've got more questions???

THanks.


-----------[000074][next][prev][last][first]----------------------------------------------------
Date:      5 Dec 1994 06:03:05 -0800
From:      cgi@crl.com (Paul Smith)
To:        comp.protocols.tcp-ip
Subject:   Re: Remote TCP/IP to Unixware

Alfred Harding (alfred.harding@windmill.com) wrote:
: I have a Unixware server that connects local workstation over TCP/IP.
 
: We are adding a remote office that has three workstations running
: Windows for Workgroups over Ethernet.
 
: I would like to allow these remote workstation access to the Unixware
: server.  I know that I could install a modem in each workstation and
: allow them SLIP access into the Unixware server.

Install the MS vxb winsock stack on each WFWG box.  Get the share ware
winsock app; comt102.zip or a leter version from your usuall winsock app
ftp archive, places I've found it; archie.au, freebsd.cdrom.com 
ftp.digital.com

Comt provides a comxx redirector to a telnet session over winsock.

I.E. It fools your procom etc opening com5 into thinking it's a local
com port when it really is a telnetsion to a node:telnet port.  Whick
allows you to access remote modems on a terminal server etc...

Good luck.


-----------[000075][next][prev][last][first]----------------------------------------------------
Date:      04 Dec 94 22:39:46
From:      Adam.Goodfellow@holt-systems.com (Adam Goodfellow)
To:        comp.protocols.tcp-ip
Subject:   Re: Slip Error msg

From: adam@comptech.demon.co.uk (Adam Goodfellow)
Subject: Re: Slip Error msg
Organization: Computech

In article <3bq2jn$78b@taco.cc.ncsu.edu>,
                Drew Wilhite <drew@taco.cc.ncsu.edu> wrote:

>
>Does anyone know the meaning of the following error message?
>
>WARNING: ipintr: pullupmsg failed
>
>It occurs in sporadic bunchs on 2 machines that are connected by SLIP.
>(Both running Dell sysvr4 issue 2.2). 
>

I assume "ipintr" refers to lower IP<->IF layer code, the pullupmsg
sounds like a reference to kernel mbuf code. Most likely it sounds like
a problem within SLIP where it probably rearranges mbufs before sending
to the hardware.

As to cause, no idea, but are you very low on memory, or running SLIP
with a huge or very small MTU? Does it typically happen with TCP
connections that are sending alot of data, resulting in a large
amount of data being buffered at the interface?

-- 
Adam

=======================================================================
| Computech  Tel/Fax: 0181 673 7817  email: adam@comptech.demon.co.uk |
=======================================================================
--
|Fidonet:  Adam Goodfellow 1:3609/504
|Internet: Adam.Goodfellow@holt-systems.com
|
| Standard disclaimer: The views of this user are strictly his own.


-----------[000076][next][prev][last][first]----------------------------------------------------
Date:      05 Dec 94 00:40:37
From:      Tony.Mountifield@holt-systems.com (Tony Mountifield)
To:        comp.protocols.tcp-ip
Subject:   Re: TCP Protocol problem Solaris<->KA9Q

From: tonymo@ntl.com (Tony Mountifield)
Subject: Re: TCP Protocol problem Solaris<->KA9Q
Organization: NTL

In article <D06sx4.6q6@info.swan.ac.uk>,
Alan Cox <iialan@iifeak.swan.ac.uk> wrote:
>In article <199411301302.NAA09562@ntl.com> tonymo@ntl.com (Tony Mountifield)
 writes:
>>I am trying to communicate between an Atari ST running the KA9Q NOS package
>>and a PC running Sun's Solaris 2, using PPP. Mostly it works. The Solaris
>>machine quite happily acts as a gateway, and has no problems accepting TCP
>>connections from the KA9Q clients.
>
>Solaris 2.1 on PC's has this urge to protocol error the piggy backed data
>stuff done by KA9Q. Now KA9Q is stretching the spec a little, for a massive
>performance gain for some operations on slow links (eg amateur radio), but
>Solaris is violating the fundamental 'be liberal in what you accept' ethic.
>
>You can either talk to sun or turn the piggy backing of data off on KA9Q.
>I can't remember how you do that offhand. 

Yes, that was the problem. Many thanks to Alan and to Chris Butterworth,
who pointed me to the command TCP SYNDATA OFF in KA9Q. This completely
cured the problem.

Tony

--
|Fidonet:  Tony Mountifield 1:3609/504
|Internet: Tony.Mountifield@holt-systems.com
|
| Standard disclaimer: The views of this user are strictly his own.


-----------[000077][next][prev][last][first]----------------------------------------------------
Date:      4 Dec 1994 04:41:16 +1100
From:      iss20903@loranth.cs.uow.edu.au (Darrick Rochili)
To:        comp.protocols.tcp-ip
Subject:   Having more than one session using SLIP

Hello, could anyone help me with this. I want to get more than
one session when I'm logged in the internet, which uses UNIX.
I'm using msdos and telix. Do I have to get a software for it ?
If anyone knows how to do this, please reply and let me know.
	Thanks


-----------[000078][next][prev][last][first]----------------------------------------------------
Date:      4 Dec 1994 23:59:40 GMT
From:      Friedrich Schotte <schotte@physik.uni-frankfurt.de>
To:        comp.protocols.tcp-ip
Subject:   Re: How to get the ethernet address of the host itself ?

> I want to get the ethernet address of the host itself. My O.S. is UNIX SVR 4.0.

You could try the following:

- Log in to another UNIX host using telnet or rlogin
- Enter the command 'arp -a' 

This gives you a table of Ethernet addresses corresponding to
Internet addresses including your hosts  Ethernet address.


-----------[000079][next][prev][last][first]----------------------------------------------------
Date:      Mon, 5 Dec 1994 01:11:09 GMT
From:      bill@twg.bc.ca (Bill Irwin)
To:        biz.sco.general,comp.protocols.tcp-ip,wimsey.general
Subject:   Fkey characters getting split on frame relay

I have a situation where the characters being sent by a
terminal's function keys are being split into separate packets as
they travel through the network on their way to the host
application.  This usually results in an error "beep", with one
or more straggling characters being displayed at the cursor
position.

I have been told by the router manufacturer (Gandalf) that there
is no way to adjust the router so that it can recognize the Fkey
lead-in character and make sure that it is not split from the
following two characters into different packets.

Although modifying each application to wait longer for Fkey
completion characters may be an option, it seems to me that the
logical place to do this would be at the OS level, either in the
TCP/IP software or the termcap/info databases.  In this way, the
configuration is done once for each type of terminal being
supported, rather than for each application.

Does anyone know of a way to ensure that Fkeys get recognized
reliably by applications running on the host/server?
-- 
Bill Irwin     -      The Westrheim Group     -    Vancouver, BC, Canada
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
uunet!twg!bill              604-431-9600 (voice) |     Your Computer  
bill@twg.bc.ca              604-430-4329 (fax)   |    Systems Partner

-----------[000080][next][prev][last][first]----------------------------------------------------
Date:      05 Dec 94 06:29:56
From:      Steve.Rago@holt-systems.com (Steve Rago)
To:        comp.protocols.tcp-ip
Subject:   Re: Slip Error msg

From: sar@plc.com (Steve Rago)
Subject: Re: Slip Error msg
Organization: Programmed Logic Corporation

In article <3bq2jn$78b@taco.cc.ncsu.edu> drew@taco.cc.ncsu.edu (Drew Wilhite)
writes:
>
>Does anyone know the meaning of the following error message?
>
>WARNING: ipintr: pullupmsg failed
>
>It occurs in sporadic bunchs on 2 machines that are connected by SLIP.
>(Both running Dell sysvr4 issue 2.2). 

It means that IP couldn't concatenate/align all the bytes it
needed to build an IP header, probably because another message
couldn't be allocated.  IP would have then discarded the message.
This is a sign of a low-memory condition.  When it happens, try
running "sar -k 1 1" to see how much memory is allocated by the
dynamic kernel memory allocator, just to verify the hypothesis.
If too much memory is being allocated, you could be hitting one
of two limits.  SYSSEGSZ, defined in /etc/conf/cf.d/mtune, limits
the amount of dynamic virtual memory that the kernel can allocate.
It is expressed in units of pages.  The other limit is the amount
of physical memory you have installed on the machine.

Steve Rago
sar@plc.com
--
|Fidonet:  Steve Rago 1:3609/504
|Internet: Steve.Rago@holt-systems.com
|
| Standard disclaimer: The views of this user are strictly his own.
 

-----------[000081][next][prev][last][first]----------------------------------------------------
Date:      05 Dec 94 06:58:34
From:      root@holt-systems.com (root)
To:        comp.protocols.tcp-ip
Subject:   Re: SLIP pins for RS232

From: root@cacofonix.utr.ac.za (root)
Subject: Re: SLIP pins for RS232
Organization: University of Transkei

css@netcom.com (Chris Schefler) writes:

>I'm setting up a hardwired SLIP connection from my UNIX server to a PC
>using RS232 connectors with modular adapters (RJ25).  
 
>My question is this: which of the 25 pins are required for SLIP?

I am trying to do something similar. Pointers to a faq on
hardware/software configuration of slip links would be much appreciated.
Or if some boff out there could kindly post a 'minimum cable' diagram. (
just 2,3 and 7 work ok for simple login access via an RS232 port on SCO
unix - not sure whether this would also work with Slip connections ? )

Thanks
Neil

--
|Fidonet:  root 1:3609/504
|Internet: root@holt-systems.com
|
| Standard disclaimer: The views of this user are strictly his own.


-----------[000082][next][prev][last][first]----------------------------------------------------
Date:      05 Dec 94 07:40:54
From:      Richard.Holbo@holt-systems.com (Richard Holbo)
To:        comp.protocols.tcp-ip
Subject:   DHCP server for Unix?????

From: HOLBOR@bus.orst.edu (Richard Holbo)
Subject: DHCP server for Unix?????
Organization: Product Testing Lab - Colledge of

Subject says it all.  I am looking for a Unix DHCP server preferrably BSDI or 
HPUX.  I've seen several notes requesting information about DHCP servers here, 
but no replies have been posted to the list.

Richard Holbo
holbor@bus.orst.edu
Product Testing Lab - College of Business - Oregon State University
--
|Fidonet:  Richard Holbo 1:3609/504
|Internet: Richard.Holbo@holt-systems.com
|
| Standard disclaimer: The views of this user are strictly his own.


-----------[000083][next][prev][last][first]----------------------------------------------------
Date:      05 Dec 94 09:17:28
From:      Shai.Fultheim-System.Assistant@holt-systems.com (Shai Fultheim-System Assistant)
To:        comp.protocols.tcp-ip
Subject:   help with RPC brodcast.

Subject: help with RPC brodcast.
From: shai@shekel.jct.ac.il (Shai Fultheim-System Assistant)
Organization: Jerusalem College of Technology

hi all.

i use RPC for runing procedures on more than one host.
i need to know (for the main program initialization) which hosts is
runing the server i wrote, so i use clnt_brodcast() for that.
but now i have problem, i'm getting all the answers from all the hosts
in aprox. 15 seconds, but clnt_brodcast() waiting for aprox. 100 more
seconds.
how can i change (decrease) the time that RPC wait till its going out
with exit status 5 (RPC_TIMEOUT).

my program folows:
-------------------------------------------------------------------------------
#include <stdio.h>
#include <string.h>
#include <sys/types.h>
#include <sys/socket.h>
#include <netinet/in.h>
#include <arpa/inet.h>
#include <rpc/rpc.h>
#include <rpc/pmap_clnt.h>
#include "rportouid.h"
#include "version.h"

#define MAXCOUNT 99 /* maximum_number_of_items_to_return */
struct in_addr names[MAXCOUNT];

extern int errno;
int count=0,lost=0;

int eachresult(char *out, struct sockaddr_in *addr)
{
  int i;
  char tmpaddr[32];
  strcpy(tmpaddr,inet_ntoa(addr->sin_addr));
  for(i=0;i<=count;i++) {
    /*  printf("count:%i i:%i lost:%i tmpaddr:%s */
    /*  names[i]:%s\n",count,i,lost,tmpaddr,inet_ntoa(names[i])); */
    if(!strcmp(inet_ntoa(names[i]),tmpaddr))
      if(lost>count) {
	return 1;
      } else {
	lost++;
	return 0;
      }
  }
  if(i==count+1) {
    names[count]=addr->sin_addr;
    printf("caught: %s\n",inet_ntoa(names[count]));
    count++;
    return 0;
  }
}

main(int argc,char *argv[])
{
  CLIENT *cl;
  ptouid_res *result;
  char *server;
  int message;

  enum clnt_stat clnt_stat;

  printf("getuser version %s\n",VERSION);

  /* the code below is for broadcating: */
  printf("starting broadcast:\n");

  count=0;
  clnt_stat = clnt_broadcast(PTOUIDPROG,PTOUIDVERS,VER_NUM,xdr_void,0,
			     xdr_void,0,eachresult);
  printf("broadcast exit status: %i\n",clnt_stat);

  if(argc!=3){
    printf("usage: %s host msg\n",argv[0]);
    exit(1);
  }

  server=argv[1];
  message=atoi(argv[2]);

  if((cl = clnt_create(server,PTOUIDPROG,PTOUIDVERS,"udp"))==NULL) {
    clnt_pcreateerror(server);
    exit(1);
  }

  if((result=ptouid_1(&message,cl))==NULL) {
    clnt_perror(cl,server);
    exit(1);
  }

  if(result->errno > 0) {
    errno = result->errno-10;
    perror("client error");
    exit(1);
  } else if(result->errno == 0) {
    printf("recived --> %s\n",result->ptouid_res_u.name);
    exit(0);
  } else if(result->errno == -1) {
 printf("recived --> no such user\n");
  }
}
-------------------------------------------------------------------------------

thx. id ad.
Shai Fultheim.

	** PLEASE ANSWER DIRECTLY TO THE OWNER OF THIS ARTICLE **



--
Shai Fultheim    				E-mail: shai@shekel.jct.ac.il
System Assistant				Tel (W):       (972)-2-751160
Jerusalem College of Technology - Israel	Tel (H):       (972)-2-963476
--
|Fidonet:  Shai Fultheim-System Assistant 1:3609/504
|Internet: Shai.Fultheim-System.Assistant@holt-systems.com
|
| Standard disclaimer: The views of this user are strictly his own.


-----------[000084][next][prev][last][first]----------------------------------------------------
Date:      5 Dec 1994 04:58:22 GMT
From:      yanoff@csd4.csd.uwm.edu (Scott A. Yanoff)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.domains
Subject:   Help! Aliasing Addresses


A company wires us up to their T1, and gives us something.com
Assuming we have only 1 computer, can we from there
name this computer   name.something.com

How is this done, if they are just giving us something.com ?

Secondly, suppose I want aliases, so that name.something.com and
gopher.something.com are the same machines?

Please email, thanx,
-Scott
-- 
 _/\ _ !\ _         @          Milwaukee, WI - A Great Place On a Great Lake
!  _! !! ! !_  ~~  @ ~  ~~         
! ! ! !! ! ! !~~__=||_~ ~~~  Info-Media Technology - Mainframe/Unix Consulting
! ! ! _! ! ~~~ ~\____/  ~~~     yanoff@csd.uwm.edu   yanoff@cs.uwm.edu

-----------[000085][next][prev][last][first]----------------------------------------------------
Date:      Mon, 5 Dec 1994 05:26:04 GMT
From:      jhk@utrcv1.res.utc.com
To:        comp.protocols.tcp-ip
Subject:   Need info on multicast routing protocols

Does anyone have info on where to find RFCs and Internet Drafts on
multicast routing protocols.  I'm looking for DVMRP, MOSPF, and especially
PIM documentation.

Thanks,
Jim Kilroy
jhk@utrc.utc.com

-----------[000086][next][prev][last][first]----------------------------------------------------
Date:      05 Dec 94 10:48:33
From:      Keith.Perry@holt-systems.com (Keith Perry)
To:        comp.protocols.tcp-ip
Subject:   Novell LWP PPP Dialup Stack Help Request

From: perryk@ksc.nasa.gov (Keith Perry)
Subject: Novell LWP PPP Dialup Stack Help Request
Organization: MDSS&DS


I'm trying to evaluate Novell's LAN WorkPlace Mobile Dialer along with their
TCPIP Stack on a PPP connection to copmare with Trumpet 2.0B and
Chameleon (Demo) which I have up and running.  Since I'm starting from
scratch, I had to retrieve 3 other patch archive files in addition to
LWP303.EXE:
VLMP1.EXE
LWP207.EXE
LWP208.EXE
I created a Net.cfg file per Dialer.txt and have 3 different boot
configurations
so that only one Winsock is in my path at a time.  I'm trying to figure out
which files are required from each archive for the dialup stack to function
properly,  i.e. I don't want to put anything on my hard drive that I don't
need.
I'm able to run the Windows Dialer, and view the modem types.  I stopped
at this point because I get a TSR error when launching windows (WfW 3.11)
and want to clean this up before I proceed.  I have not yet tried loading all
the LWP207 and LWP208 files to my /bin directory, and will try if I don't get
any feedback on this.  Another concern is the [Protocol TCPIP] section
of my Net.cfg, in particular, the "tcp-cfg, script and profile" path
statements.
Do I need these directories (or the info in them) on my hard drive?  Can I
delete reference to these?

Thanks in advance,

Keith

Below is some pertinent info:

 



Error message generated when launching Windows:
-------------------------------------------------------
	   TSR Support Problem
"Cannot Run Application Specified by SLIPP_PPP"



Text dumped to screen during Autoexec.bat execution
-------------------------------------------------------
The configuration file used was c:\internet\novell\bin\net.cfg
Max Boards 4, Max Stacks 4
Buffers 8,  Buffer Size 1500 bytes, Memory Pool 4076 bytes

Novell Asynchronous SLIPP_PPP Driver Dual Mode
(R42-5, 940715)  Copyright 1992 Novell Inc.
Int 4, Port 3F8
Max Frame 1500 bytes
Board 1, Frame PPP

Novell TCP/IP Transport v4.2 (940802)
Copyright 92,93,94 Novell Inc
Network Name: IP_NET	Bind: SLIPP_PPP
IP Address: 0.0.0.0	Board Number: 1
Subnet Mask: 255.0.0.0  Frame: PPP




NET.CFG
-------------------------------------------------------
Link Support
  buffers 8 1500
  mempool 4096 

Link Driver SLIP_PPP
  FRAME  PPP
  ACCM   00000000
  INT    4           
  Port   3F8                 
  TCPIPComp VJ    
  COMMOWNER NO      
     
Protocol TCPIP
   path lwp_cfg c:\internet\novell\hstacc
   path tcp_cfg c:\internet\novell\tcp
   path script c:\internet\novell\script
   path profile c:\internet\novell\profile
   tcp_window 4096       


Applicable lines in Autoexec.bat:
-------------------------------------------------------
SET NET=C:\INTERNET\NOVELL
CD C:\INTERNET\NOVELL\BIN
LSL
SLIP_PPP
TCPIP
PATH C:\HERCULES;C:\WINDOWS;C:\DOS;C:\MOUSE;C:\;C:\PGP;%NET%


Directory of C:\INTERNET\NOVELL
------------------------------------------------------- 
LWP207   TXT      7387 08-31-94   5:18p
PPPFAQ   WRI     71062 09-06-94   4:09p
PPP      TXT     17922 08-05-93   2:51p
DIALER   TXT     12782 08-30-94   5:49p
LWP203   TXT      4907 09-02-94  11:15a
BIN          <DIR>     11-27-94  10:41p
HSTACC       <DIR>     11-27-94  10:42p
LWP208   TXT     13823 08-31-94   5:27p
NOVFILES TXT         0 11-27-94  10:57p
       11 file(s)     127883 bytes

Directory of C:\INTERNET\NOVELL\BIN

CONFIG   EXE    104720 07-20-94  10:57a
CONFIG   HLP     23147 07-20-94  10:57a
DIALER   EXE    292704 07-20-94  10:57a
DIALER   HLP    222959 07-20-94  10:57a
MDMAPI   DLL     54808 07-20-94  10:57a
SLIP_PPP COM     42046 07-20-94  10:57a
SLPAPI   DLL     24388 07-20-94  10:57a
SLPMON   EXE     30640 07-20-94  10:57a
LSL      COM     18309 09-22-94   7:48p
LWPUTIL  DLL     35936 03-21-94   4:20a
NOVASYNC EXE      4672 07-20-94  10:57a
TCPIP    EXE     43652 08-02-94   2:39p
VTCPIP   386     10721 04-04-94   4:20a
WINSOCK  DLL     35120 07-20-94  10:57a
WLIBSOCK DLL     47846 04-04-94   4:20a
NET      CFG       435 11-27-94  10:47p
       18 file(s)     992103 bytes

Directory of C:\INTERNET\NOVELL\HSTACC

AIOMDMS  MDC    172824 07-28-94   4:53p
CODES    DLR      5412 07-20-94  10:57a
        4 file(s)     178236 bytes

Total files listed:
       33 file(s)    1298222 bytes
                    
-- 
_____________________________________
KEITH PERRY                  
Communication Systems Engineer	
McDonnell Douglas Space Systems	
Kennedy Space Center, Florida	
						
Office: perryk@sugarloaf.ksc.nasa.gov	
Home:   k+rperry@bb.iu.net	
______________________________________
-- 
_____________________________________
KEITH PERRY                  
Communication Systems Engineer	
McDonnell Douglas Space Systems	
Kennedy Space Center, Florida	
						
Office: perryk@sugarloaf.ksc.nasa.gov	
Home:   k+rperry@bb.iu.net	
______________________________________

--
|Fidonet:  Keith Perry 1:3609/504
|Internet: Keith.Perry@holt-systems.com
|
| Standard disclaimer: The views of this user are strictly his own.


-----------[000087][next][prev][last][first]----------------------------------------------------
Date:      Mon, 5 Dec 1994 06:39:46 GMT
From:      adam@comptech.demon.co.uk (Adam Goodfellow)
To:        comp.protocols.tcp-ip,comp.unix.sys5.r4
Subject:   Re: Slip Error msg

In article <3bq2jn$78b@taco.cc.ncsu.edu>,
                Drew Wilhite <drew@taco.cc.ncsu.edu> wrote:

>
>Does anyone know the meaning of the following error message?
>
>WARNING: ipintr: pullupmsg failed
>
>It occurs in sporadic bunchs on 2 machines that are connected by SLIP.
>(Both running Dell sysvr4 issue 2.2). 
>

I assume "ipintr" refers to lower IP<->IF layer code, the pullupmsg
sounds like a reference to kernel mbuf code. Most likely it sounds like
a problem within SLIP where it probably rearranges mbufs before sending
to the hardware.

As to cause, no idea, but are you very low on memory, or running SLIP
with a huge or very small MTU? Does it typically happen with TCP
connections that are sending alot of data, resulting in a large
amount of data being buffered at the interface?

-- 
Adam

=======================================================================
| Computech  Tel/Fax: 0181 673 7817  email: adam@comptech.demon.co.uk |
=======================================================================

-----------[000088][next][prev][last][first]----------------------------------------------------
Date:      Mon, 5 Dec 1994 15:40:54 LOCAL
From:      HOLBOR@bus.orst.edu (Richard Holbo)
To:        comp.protocols.tcp-ip
Subject:   DHCP server for Unix?????

Subject says it all.  I am looking for a Unix DHCP server preferrably BSDI or 
HPUX.  I've seen several notes requesting information about DHCP servers here, 
but no replies have been posted to the list.

Richard Holbo
holbor@bus.orst.edu
Product Testing Lab - College of Business - Oregon State University

-----------[000089][next][prev][last][first]----------------------------------------------------
Date:      Mon, 5 Dec 1994 08:40:37 +0000
From:      tonymo@ntl.com (Tony Mountifield)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.unix.solaris,demon.ip.developers,demon.ip.support,demon.ip.support.pc,demon.ip.support.atari,demon.ip.support.unix
Subject:   Re: TCP Protocol problem Solaris<->KA9Q

In article <D06sx4.6q6@info.swan.ac.uk>,
Alan Cox <iialan@iifeak.swan.ac.uk> wrote:
>In article <199411301302.NAA09562@ntl.com> tonymo@ntl.com (Tony Mountifield) writes:
>>I am trying to communicate between an Atari ST running the KA9Q NOS package
>>and a PC running Sun's Solaris 2, using PPP. Mostly it works. The Solaris
>>machine quite happily acts as a gateway, and has no problems accepting TCP
>>connections from the KA9Q clients.
>
>Solaris 2.1 on PC's has this urge to protocol error the piggy backed data
>stuff done by KA9Q. Now KA9Q is stretching the spec a little, for a massive
>performance gain for some operations on slow links (eg amateur radio), but
>Solaris is violating the fundamental 'be liberal in what you accept' ethic.
>
>You can either talk to sun or turn the piggy backing of data off on KA9Q.
>I can't remember how you do that offhand. 

Yes, that was the problem. Many thanks to Alan and to Chris Butterworth,
who pointed me to the command TCP SYNDATA OFF in KA9Q. This completely
cured the problem.

Tony


-----------[000090][next][prev][last][first]----------------------------------------------------
Date:      Mon, 05 Dec 94 13:48:33 EST
From:      perryk@ksc.nasa.gov (Keith Perry)
To:        comp.protocols.tcp-ip
Subject:   Novell LWP PPP Dialup Stack Help Request


I'm trying to evaluate Novell's LAN WorkPlace Mobile Dialer along with their
TCPIP Stack on a PPP connection to copmare with Trumpet 2.0B and
Chameleon (Demo) which I have up and running.  Since I'm starting from
scratch, I had to retrieve 3 other patch archive files in addition to LWP303.EXE:
VLMP1.EXE
LWP207.EXE
LWP208.EXE
I created a Net.cfg file per Dialer.txt and have 3 different boot configurations
so that only one Winsock is in my path at a time.  I'm trying to figure out
which files are required from each archive for the dialup stack to function
properly,  i.e. I don't want to put anything on my hard drive that I don't need.
I'm able to run the Windows Dialer, and view the modem types.  I stopped
at this point because I get a TSR error when launching windows (WfW 3.11)
and want to clean this up before I proceed.  I have not yet tried loading all
the LWP207 and LWP208 files to my /bin directory, and will try if I don't get
any feedback on this.  Another concern is the [Protocol TCPIP] section
of my Net.cfg, in particular, the "tcp-cfg, script and profile" path statements.
Do I need these directories (or the info in them) on my hard drive?  Can I
delete reference to these?

Thanks in advance,

Keith

Below is some pertinent info:

 



Error message generated when launching Windows:
-------------------------------------------------------
	   TSR Support Problem
"Cannot Run Application Specified by SLIPP_PPP"



Text dumped to screen during Autoexec.bat execution
-------------------------------------------------------
The configuration file used was c:\internet\novell\bin\net.cfg
Max Boards 4, Max Stacks 4
Buffers 8,  Buffer Size 1500 bytes, Memory Pool 4076 bytes

Novell Asynchronous SLIPP_PPP Driver Dual Mode
(R42-5, 940715)  Copyright 1992 Novell Inc.
Int 4, Port 3F8
Max Frame 1500 bytes
Board 1, Frame PPP

Novell TCP/IP Transport v4.2 (940802)
Copyright 92,93,94 Novell Inc
Network Name: IP_NET	Bind: SLIPP_PPP
IP Address: 0.0.0.0	Board Number: 1
Subnet Mask: 255.0.0.0  Frame: PPP




NET.CFG
-------------------------------------------------------
Link Support
  buffers 8 1500
  mempool 4096 

Link Driver SLIP_PPP
  FRAME  PPP
  ACCM   00000000
  INT    4           
  Port   3F8                 
  TCPIPComp VJ    
  COMMOWNER NO      
     
Protocol TCPIP
   path lwp_cfg c:\internet\novell\hstacc
   path tcp_cfg c:\internet\novell\tcp
   path script c:\internet\novell\script
   path profile c:\internet\novell\profile
   tcp_window 4096       


Applicable lines in Autoexec.bat:
-------------------------------------------------------
SET NET=C:\INTERNET\NOVELL
CD C:\INTERNET\NOVELL\BIN
LSL
SLIP_PPP
TCPIP
PATH C:\HERCULES;C:\WINDOWS;C:\DOS;C:\MOUSE;C:\;C:\PGP;%NET%


Directory of C:\INTERNET\NOVELL
------------------------------------------------------- 
LWP207   TXT      7387 08-31-94   5:18p
PPPFAQ   WRI     71062 09-06-94   4:09p
PPP      TXT     17922 08-05-93   2:51p
DIALER   TXT     12782 08-30-94   5:49p
LWP203   TXT      4907 09-02-94  11:15a
BIN          <DIR>     11-27-94  10:41p
HSTACC       <DIR>     11-27-94  10:42p
LWP208   TXT     13823 08-31-94   5:27p
NOVFILES TXT         0 11-27-94  10:57p
       11 file(s)     127883 bytes

Directory of C:\INTERNET\NOVELL\BIN

CONFIG   EXE    104720 07-20-94  10:57a
CONFIG   HLP     23147 07-20-94  10:57a
DIALER   EXE    292704 07-20-94  10:57a
DIALER   HLP    222959 07-20-94  10:57a
MDMAPI   DLL     54808 07-20-94  10:57a
SLIP_PPP COM     42046 07-20-94  10:57a
SLPAPI   DLL     24388 07-20-94  10:57a
SLPMON   EXE     30640 07-20-94  10:57a
LSL      COM     18309 09-22-94   7:48p
LWPUTIL  DLL     35936 03-21-94   4:20a
NOVASYNC EXE      4672 07-20-94  10:57a
TCPIP    EXE     43652 08-02-94   2:39p
VTCPIP   386     10721 04-04-94   4:20a
WINSOCK  DLL     35120 07-20-94  10:57a
WLIBSOCK DLL     47846 04-04-94   4:20a
NET      CFG       435 11-27-94  10:47p
       18 file(s)     992103 bytes

Directory of C:\INTERNET\NOVELL\HSTACC

AIOMDMS  MDC    172824 07-28-94   4:53p
CODES    DLR      5412 07-20-94  10:57a
        4 file(s)     178236 bytes

Total files listed:
       33 file(s)    1298222 bytes
                    
-- 
_____________________________________
KEITH PERRY                  
Communication Systems Engineer	
McDonnell Douglas Space Systems	
Kennedy Space Center, Florida	
						
Office: perryk@sugarloaf.ksc.nasa.gov	
Home:   k+rperry@bb.iu.net	
______________________________________
-- 
_____________________________________
KEITH PERRY                  
Communication Systems Engineer	
McDonnell Douglas Space Systems	
Kennedy Space Center, Florida	
						
Office: perryk@sugarloaf.ksc.nasa.gov	
Home:   k+rperry@bb.iu.net	
______________________________________


-----------[000091][next][prev][last][first]----------------------------------------------------
Date:      05 Dec 94 15:01:09
From:      Dan.Lanciani@holt-systems.com (Dan Lanciani)
To:        comp.protocols.tcp-ip
Subject:   Re: Shutdown() block proof?

From: ddl@harvard.edu (Dan Lanciani)
Subject: Re: Shutdown() block proof?
Organization: Internet 

In article <3bv64t$n2e@crl6.crl.com>, cgi@crl.com (Paul Smith) writes:
|I'm trying to craft a sockets wrapper library that will not block
 indeffinately
|on a close of a TCP service NOR interact with a child task's use of the fd
|after a perent server forks (concurrant TCP server).
|
|1) I've set SO_LINGER to 10 seconds, for a reasonalby short pause for the
|other end to read out standing packets.
|
|2) I was setting the fd to O_NONBLOCK prior to calling close to avoid
 potential
|indefinite blocks..
|
|I've discovered the following architectural interactions:
|
|1) For a TCP concurrant server, where the main server forks on every accept()
|then closes the new fd, the child closes the main fd.  There can be
 interaction
|from both tasks acting on the same fd control block in the kernel if the
|library
|I wrote sets the fd to O-NONBLOCK prior to calling close. 
|
|In the case of the parent closing the new client accept() fd, it sets the fd
|to O_NONBLOCK and the child task's use of this socket, then becomes
|non-blocking.
|OK, ok, I didn't think of that when I coded my close library function. Then
 how
|to I craft a robust close wrapper given that it will be used for both
|sequential and concurrant server and client environments???
|
|Do I need two (or more) close wrapper functions; 1 for orderly shutdown;
|SO_LINGER
|set to some reasonable short time, set to O_NONBLOCK ?? (or is this at odds
|with
|SO_LINGER??) then close.  2) don't fool with it close, just call close, DON't
|call shutdown, or set to O_NONBLOCK.

I quoted the whole message because I think the problem isn't well captured
in any one part of the question.  The important thing to realize is that
most/all of the operations you are using act on the underlying object
(the socket) and not on the descriptor you are using to reference it.  This
certainly applies to shutdown() and the SO_LINGER socket option.  In most
systems this also applies to the blocking state, though I have heard of
cases where such state is maintained pre-file-descriptor.  So, anything
you do is going to change the socket's characteristics as seen by your
child process.

If your main concern is that the parent never be allowed to block on
a close, you might want to consider some way of synchronizing the child
and parent.  Understand that the only way the close() call can block
on a socket is if this is the final close(), i.e., there are no other
file descriptors associated with the socket.  In any other case, close()
merely decrements a reference count.  Unless your child dies or deliberately
closes its descriptor before the parent can close its own, you are safe.
I think most implementations consider it safe enough that the parent closes
immediately after the fork().  You could add a sleep in the child to further
bias the odds in your favor.  To be certain, you could employ some IPC
mechanism
to block the child until the parent has completed the close().  Ultimately,
I suspect this is all overkill.  The circumstances under which the final
close() of a newly accept()ed socket can block for a significant time
are rather obscure, perhaps even bugs. :)

				Dan Lanciani
				ddl@harvard.*
--
|Fidonet:  Dan Lanciani 1:3609/504
|Internet: Dan.Lanciani@holt-systems.com
|
| Standard disclaimer: The views of this user are strictly his own.


-----------[000092][next][prev][last][first]----------------------------------------------------
Date:      5 Dec 1994 10:29:21 GMT
From:      thed@mtek.chalmers.se (Thed Lyngmo)
To:        comp.protocols.tcp-ip
Subject:   Re: PPP via telnet...

Ben Peoples (bpeoples@iglou.iglou.com) wrote:
: Is it possible to run a ppp (or slip for that matter) connection over a 
: telnet connection?  

I only know how SLIP works and I can't see why that shouldn't be possible.
I guess the same goes for PPP.

What do you want to achieve by doing that?

Kind regards,
Ted Lyngmo

-----------[000093][next][prev][last][first]----------------------------------------------------
Date:      5 Dec 1994 19:38:18 -0500
From:      msbeebe@mtu.edu (Matt Beebe)
To:        comp.protocols.tcp-ip
Subject:   Domain help

I know that this is probably in the wrong newsgroup, however,

Where do I apply to register my.company.com as a domain w/ IP address?


		Thanks,

					-Matt


-----------[000094][next][prev][last][first]----------------------------------------------------
Date:      Mon, 5 Dec 1994 11:30:00 GMT
From:      ddv@unix.brighton.ac.uk (Dom De Vitto)
To:        comp.object,comp.protocols.tcp-ip,comp.unix.solaris,comp.lang.c++,comp.client-server
Subject:   Re: ACE version 2.15.5 now available

: : It is not clear to me why you would 'naturally' not be responsible for
: : any problems caused by using this library.  What if a problem within this
: : library caused all of the Iridium satellites to become useless?  I'm not
: : saying you have full responsibility by any means, but to claim a desire
: : of a comprehensive and robust library without responsibility for any
: : lack of those characteristics seems like a conflict.
 
: Don't look a gift horse in the mouth!
 
: If you want someone to take responsibility fork out the $$$$. The above 
: is simply a standard disclaimer that anyone in their right mind would put 
: with free software.

Look at a comercial software package sometime - they say *exactly* the same
thing...:(

Dom

-----------[000095][next][prev][last][first]----------------------------------------------------
Date:      Mon, 5 Dec 1994 10:34:29 MET
From:      bohac@feld.cvut.cz (Leos Bohac Ing. K332)
To:        comp.protocols.tcp-ip
Subject:   TCP/IP protocol stack needed

Hi everyone,
	I am looking for TCP/IP stack source in C language to experiment
with the network using packet driver. Does anybody knows a shareware
source where I can retrieve that. 	
				Thanks in advance
					LEO
DATA.15

-----------[000096][next][prev][last][first]----------------------------------------------------
Date:      5 Dec 1994 12:03:25 GMT
From:      jyl@yiscgw.yonsei.ac.kr (Lee JaiYong)
To:        comp.protocols.tcp-ip
Subject:   CFP ICCC95


Following is the SECOND CALL FOR PAPER for ICCC'95(revised version) to be held in Seoul
Korea 1995.

Publicity Chair, 
ICCC'95
-----------------cut here---------------------------------------------
----------------------------------------------------------------------

                     	  CALL FOR PAPERS

                       	     ICCC '95 

       "Information Highways for a Smaller World  & Better Living" 
			  Seoul, Korea
                     August 21 - 24, 1995

-----------------------------------------------------------------------------------

    The ICCC, the International Council for Computer Communication (ICCC), 
    founded in 1972,  is an Affiliate Member of the International Federation 
    for Information Processing (IFIP). 

    Its purposes are to foster:   
        · scientific research and the development of computer communication;
        · progress in the evaluation of applications of computer communication 
	  to educational, scientific, medical, economic, legal, cultural and 
	  other peaceful purposes;
        · study of the potential social and economic impacts of computer 
	  communcation and of policies which influence those impacts. 

    This 12th conference aims at providing a forum to exchange ideas, discuss 
    key issues and to present  the late research results for "Information Highways 
    for a Smaller World  & Better Living."  The main program includes technical 
    presentations, invited talks, tutorials, and technical visits. 


    TOPICS :  Areas of interest include but are not limited to

	. Strategies, Policies, and User       . Wireless Communications
	  Perspectives of Information          . Intelligent Networks		
	  Superhighways                        . Personal Communications Systems
        . Social and Economical Impacts        . Broadband Communication
	  of Information Superhighways         . ATM Switching
        . Computer Communication for           . International Emergencies
	  Developing Countries                 . Distance Learning
        . Network Planning                     . Optical Communications
	. Security and Privacy in Computer     . Multimedia Communication and its
	  Communications                         Applications
        . Evolution towards the High-Speed     . High-Speed Protocols
	  Networks including Frame Relay       . Network Management  
	  and SMDS                             . Protocol Engineering     
	. Packet Radio Technologies
	. Satellite Communications


    SUBMISSION OF PAPERS  

        Prospective authors should send 5 copies of a full paper to the following 
	address;
		ICCC'95
		Dr. Seon Jong Chung
		ICCC'95 Technical Program Chairman
		ETRI,  Yusong P.O.Box 106, Taejon, Korea, 305-606
		Tel: +82-42-860-8630
		Fax: +82-42-860-6465
		E-mail: iccc95@giant.etri.re.kr


        The manuscript should not exceed 4000 words in length and should include 
	author's name, affiliation, and addresses(telephone, e-mail, fax), and 
	150-200 words abstracts in the title page. Also, authors are encouraged 
	to send a Postscript version of their full paper to the Technical Program 
	Committee Chairman by e-mail iccc95@giant.etri.re.kr

                             |-------------------------------|
                             |  Important Dates              |
                             |    Submission of Paper        |
                             |      February 1st, 1995       |
                             |    Notification of Acceptance |
                             |      May 1st, 1995            |
                             |    Camera-ready Papers        |
                             |      June 15th, 1995          |
                             |-------------------------------|

-----------------------------------------------------------------------------------
    Sponsored by
	The International Council for Computer Communication

    Hosted by
	Electronics and Telecommunications Research Institute
	Korea Information Science Society

    Under the Patronage of
	Ministry of Communication, The Republic of Korea


    Conference Governor                          
	Ronald P.Uhlig, Northern Telecom, U.S.A.
					       

    Conference Organizing Committee
	Chair : Chongsun Hwang, KISS, Korea
	Co-Chair : Seungtaik Yang, ETRI, Korea

    Local Arrangement 
	Dongho Lee, Kwangwoon Unvi., Korea

    Publication 
	Keosang Lee, Dacom, Korea
    
    Publicity
	Jaiyong Lee, Yon-Sei Univ., Korea

    Registration 
	Samyoung Suh, NCA, Korea

    Treasurer 
	Seungkyu Park, Ajou Univ., Korea

    Tutorial 
	Sunshin An, Korea Univ., Korea

    Social Program 
	Nosik Kim, KTRC, Korea

    Secretariate 
	Yanghee Choi, SNU, Korea
        Jinpyo Hong, ETRI, Korea

    Technical Program 
        Chair : Seonjong Chung, ETRI, Korea
	Co-Chairs : Serge Fdida, MASI, France
		    Nicholas Georganas, Univ. of Ottawa, Canada
		    Roger Needham, Univ. of Cambridge, U.K.
		    Otto Spaniol, Aachen Tech. Univ., Germany
		    Hideyoshi Tominaga, Waseda Univ., Japan
		    Pramode Verma, AT&T, U.S.A.

        Members : Sunshin An, Korea Univ., Korea
		  Yanghee Choi, SNU, Korea
		  Jin Pyo Hong, ETRI/PEC, Korea
		  Byungchul Shin, KAIST, Korea
		  Yongjin Park, Hanyang Univ., Korea
		  Donggyoo Kim, Ajou Univ., Korea
		  Seungkyu Park, Ajou Univ., Korea
		  Dongho Lee, Kwangwoon Univ., Korea
		  Kwangsue Chung, Kwangwoon Univ., Korea
		  Daeyoung Kim, Cheoungnam National Univ., Korea
		  Ilyoung Chung, ETRI, Korea
		  Chimoon Han, ETRI, Korea
		  Woojik Chon, ETRI, Korea
		  Hoon Choi, ETRI, Korea
		  Jaiyong Lee, Yonsei Univ., Korea
		  Tadao Saito, Tokyo Univ., Japan
                  Tahahiko Kamae, HP Lab., Japan
		  Reigo Yatsuboshi, Fujitsu Lab., Japan
		  Kinji Ono, NACSIS, Japan
		  Michel Diaz, LAAS-CNRS, France
		  Christophie Diot, INRIA, France
		  Jean-Yves Le Boudec, IBM, Zurich Lab., Swiss
		  Georgio Ventre, Univ. di Napoli, France
		  David Hutchison, Lanchaster Univ., U.K.
		  Augusto Casaca, INES,Portugal 
		  Martina Zitterbart, Univ. of Karlsruhe, Germany
		  Ulf Koerner, Lund Univ., Sweden
		  David J. Farber, Univ. of Pennsylvania, USA
		  Reg A. Kaenel, Marcicopa-County Comm. College, USA
		  Ira Cotton, USA
		  Martin E. Silveretoin, USA
		  Albert Kuendig, Swiss Federal Inst. of Tech., Swiss

      
-----------------------------------------------------------------------------------

-----------[000097][next][prev][last][first]----------------------------------------------------
Date:      Mon, 5 Dec 1994 14:29:56 GMT
From:      sar@plc.com (Steve Rago)
To:        comp.protocols.tcp-ip,comp.unix.sys5.r4
Subject:   Re: Slip Error msg

In article <3bq2jn$78b@taco.cc.ncsu.edu> drew@taco.cc.ncsu.edu (Drew Wilhite) writes:
>
>Does anyone know the meaning of the following error message?
>
>WARNING: ipintr: pullupmsg failed
>
>It occurs in sporadic bunchs on 2 machines that are connected by SLIP.
>(Both running Dell sysvr4 issue 2.2). 

It means that IP couldn't concatenate/align all the bytes it
needed to build an IP header, probably because another message
couldn't be allocated.  IP would have then discarded the message.
This is a sign of a low-memory condition.  When it happens, try
running "sar -k 1 1" to see how much memory is allocated by the
dynamic kernel memory allocator, just to verify the hypothesis.
If too much memory is being allocated, you could be hitting one
of two limits.  SYSSEGSZ, defined in /etc/conf/cf.d/mtune, limits
the amount of dynamic virtual memory that the kernel can allocate.
It is expressed in units of pages.  The other limit is the amount
of physical memory you have installed on the machine.

Steve Rago
sar@plc.com

-----------[000098][next][prev][last][first]----------------------------------------------------
Date:      5 Dec 1994 12:52:52 +0100
From:      jj@ekf.werries.de ("J. Jansen")
To:        comp.protocols.tcp-ip,comp.realtime,comp.sources.wanted,comp.arch.bus.vmebus
Subject:   Re: TCP/IP communications via a bus or shared memory?

bennett@cs.unc.edu (Robert Bradley Bennett) writes:


>Does anybody have source for TCP/IP communications driver using shared
>memory or a communications bus as the network media?
 
>Any pointer would be greatly appreciated...

EKF provides OS-9/RAMNet for their CPU boards. This protocol matches
the specifications of OS-9/Net. The TCP/IP backplane net version will be
ported within the next month by EKF.

Anybody interested in a free data sheet feel free to write to

                            info@ekf.werrries.de

Please do not forget to include your full postal address, thank you!


--------------------- EKF Elektronik Messtechnik GmbH ---------------------
Joachim Jansen                                 e-mail:    jj@ekf.werries.de
Philipp Reis Str. 4                            phone:   ++49 (0)2381-6890-0
D-59065 Hamm (Germany)                         fax:    ++49 (0)2381-6890-90
------------------- Advanced VMEbus Hardware + Software -------------------


-- 
---------------------------------------------------------------------------
                              jj@ekf.werries.de
---------------------------------------------------------------------------


-----------[000099][next][prev][last][first]----------------------------------------------------
Date:      Mon, 5 Dec 1994 19:43:08
From:      Timo.Vanttinen@ntc.nokia.com (Timo Vänttinen)
To:        comp.protocols.tcp-ip
Subject:   Problem with PC/TCP's ftp_open()


Hello,

I am using PC/TCP's Dev Kit 2.3 in order to integrate certain
ftp functions into my application. The problem is that if
ftp_open() gets an address to which it can't connect it hangs
the application.

When ftp_connect() is called with nonexistent address, it never
finishes and the application starts to process Windows messages.
I am able to use menus of my application etc. But the first time
I click another window the program hangs.

I have similar problem with FTP Software's ftp-application.
If I try to connect it into a nonexistent machine, program hangs
with similar symptoms.

I have PC/TCP 2.3, WfW 3.11 and Lan Manager and I use Borland
C++ for windows v 4.02. Does anybody out there have any idea how
to solve the problem or what might cause it?



Thanks,
__________________________________________________________________
Timo Vanttinen                 Email: Timo.Vanttinen@ntc.nokia.com
Nokia Telecommunications Oy    Tel:   +358 0 51141
P.O. Box 12, FIN-02611 Espoo   Fax:   +358 0 5117432
Finland


-----------[000100][next][prev][last][first]----------------------------------------------------
Date:      5 Dec 1994 15:10:54 GMT
From:      jonesj@renoir (Jeff Jones)
To:        comp.protocols.tcp-ip
Subject:   Source Code

	I have just been tasked with writing code to allow the out systems
	to do some data transfer and I was wondering if anyone could point
	me to an FTP site where there might be some example code. 

	I have some sniglets of code, but they are very basic (no drop
	detection or error handling).



-----------[000101][next][prev][last][first]----------------------------------------------------
Date:      6 Dec 1994 04:55:11 -0800
From:      cgi@crl.com (Paul Smith)
To:        comp.protocols.tcp-ip
Subject:   Re: Source Code

Jeff Jones (jonesj@renoir) wrote:
: 	I have just been tasked with writing code to allow the out systems
: 	to do some data transfer and I was wondering if anyone could point
: 	me to an FTP site where there might be some example code. 
 
: 	I have some sniglets of code, but they are very basic (no drop
: 	detection or error handling).

Get a copy of both of the R. Stevens TCP/IP programming books.  There are 
ftp sites (all, choose one; ftp.uu.net etc): ../books/.. where the src
in these and other books is available.

Don't forget that ncftp (src on the net) is a scriptable ftp tool.  That
in .ncftprc can be a whole file transfer and logic script.  Why write
C and debug when simple scripts will do..  Also, theres a C callable
ftp library in the comp.misc.* src archives; libftp.partXX.

Good luck.




-----------[000102][next][prev][last][first]----------------------------------------------------
Date:      Mon, 5 Dec 1994 17:17:28 GMT
From:      shai@shekel.jct.ac.il (Shai Fultheim-System Assistant)
To:        comp.protocols.tcp-ip
Subject:   help with RPC brodcast.

hi all.

i use RPC for runing procedures on more than one host.
i need to know (for the main program initialization) which hosts is
runing the server i wrote, so i use clnt_brodcast() for that.
but now i have problem, i'm getting all the answers from all the hosts
in aprox. 15 seconds, but clnt_brodcast() waiting for aprox. 100 more
seconds.
how can i change (decrease) the time that RPC wait till its going out
with exit status 5 (RPC_TIMEOUT).

my program folows:
-------------------------------------------------------------------------------
#include <stdio.h>
#include <string.h>
#include <sys/types.h>
#include <sys/socket.h>
#include <netinet/in.h>
#include <arpa/inet.h>
#include <rpc/rpc.h>
#include <rpc/pmap_clnt.h>
#include "rportouid.h"
#include "version.h"

#define MAXCOUNT 99 /* maximum_number_of_items_to_return */
struct in_addr names[MAXCOUNT];

extern int errno;
int count=0,lost=0;

int eachresult(char *out, struct sockaddr_in *addr)
{
  int i;
  char tmpaddr[32];
  strcpy(tmpaddr,inet_ntoa(addr->sin_addr));
  for(i=0;i<=count;i++) {
    /*  printf("count:%i i:%i lost:%i tmpaddr:%s */
    /*  names[i]:%s\n",count,i,lost,tmpaddr,inet_ntoa(names[i])); */
    if(!strcmp(inet_ntoa(names[i]),tmpaddr))
      if(lost>count) {
	return 1;
      } else {
	lost++;
	return 0;
      }
  }
  if(i==count+1) {
    names[count]=addr->sin_addr;
    printf("caught: %s\n",inet_ntoa(names[count]));
    count++;
    return 0;
  }
}

main(int argc,char *argv[])
{
  CLIENT *cl;
  ptouid_res *result;
  char *server;
  int message;

  enum clnt_stat clnt_stat;

  printf("getuser version %s\n",VERSION);

  /* the code below is for broadcating: */
  printf("starting broadcast:\n");

  count=0;
  clnt_stat = clnt_broadcast(PTOUIDPROG,PTOUIDVERS,VER_NUM,xdr_void,0,
			     xdr_void,0,eachresult);
  printf("broadcast exit status: %i\n",clnt_stat);

  if(argc!=3){
    printf("usage: %s host msg\n",argv[0]);
    exit(1);
  }

  server=argv[1];
  message=atoi(argv[2]);

  if((cl = clnt_create(server,PTOUIDPROG,PTOUIDVERS,"udp"))==NULL) {
    clnt_pcreateerror(server);
    exit(1);
  }

  if((result=ptouid_1(&message,cl))==NULL) {
    clnt_perror(cl,server);
    exit(1);
  }

  if(result->errno > 0) {
    errno = result->errno-10;
    perror("client error");
    exit(1);
  } else if(result->errno == 0) {
    printf("recived --> %s\n",result->ptouid_res_u.name);
    exit(0);
  } else if(result->errno == -1) {
 printf("recived --> no such user\n");
  }
}
-------------------------------------------------------------------------------

thx. id ad.
Shai Fultheim.

	** PLEASE ANSWER DIRECTLY TO THE OWNER OF THIS ARTICLE **



--
Shai Fultheim    				E-mail: shai@shekel.jct.ac.il
System Assistant				Tel (W):       (972)-2-751160
Jerusalem College of Technology - Israel	Tel (H):       (972)-2-963476

-----------[000103][next][prev][last][first]----------------------------------------------------
Date:      5 Dec 1994 04:43:13 +0800
From:      peter@haywire.DIALix.COM (Peter Wemm)
To:        comp.protocols.tcp-ip,comp.unix.sys5.r4
Subject:   Re: Slip Error msg

drew@taco.cc.ncsu.edu (Drew Wilhite) writes:

>Does anyone know the meaning of the following error message?
 
>WARNING: ipintr: pullupmsg failed
 
>It occurs in sporadic bunchs on 2 machines that are connected by SLIP.
>(Both running Dell sysvr4 issue 2.2). 
 
>I apologize if this is a dumb question...
 
>Drew Wilhite
>ncsipm1.cropsci.ncsu.edu

It means that your slip driver passed a packet up to the IP layer that
was so short, that it wasn't even long enough to have a valid IP
header.   The kernel, after recieving a packet, makes a call to
pullupmsg() to make sure that there are at least 20 bytes in the first
message block.  The message is saying that this failed.

Usually this is caused by bad slip framing characters or something
like that at startup - I never figured out for sure.

Incidently, a bug was recently discovered in pullupmsg() - it does not
protect itself from streams interrupts and can cause serious
corruption and panics.  It's been fixed on unixware with a PTF, but I
dont think there are any available for the older SVR4
releases.. Fortunately, it's very rare that pullupmsg() is called from
an interrupt handler.

-Peter

-----------[000104][next][prev][last][first]----------------------------------------------------
Date:      Mon, 5 Dec 1994 19:00:20 GMT
From:      patrick@vega.oes.amdahl.com (Patrick Horgan)
To:        comp.unix.programmer,comp.protocols.tcp-ip
Subject:   Re: Address already in use message

In article <NEWTNews.18249.786316206.wbell@interramp.com>, Bill Bell <pp001529@interramp.com> writes:
|> 
|> I have read Rago and Stevens books and I cannot find a good explanation of the 
|> socket options.  WHat does SO_LINGER and SO_REUSEADDR do?
|> 

SO_REUSEADDR

An address associated with a  TCP socket doesn't become available for reuse
when you close the socket. TCP keeps it available for about twice the maximum
time for a packet to live out on the net.  This is because TCP doesn't
really know when a packet is lost, it just sets a timeout, and if a reply
doesn't appear before the timeout's over, it assumes that the packet is lost.
This implies that in general, at any given moment, there are packets floating
around out there that we sent, or replies to them might be floating around.
They might be caught in a short lived routing loop, or just be delayed.  Be-
cause of this, the TPC software keeps that ip-address/port number pair reserved
for about twice the max TTL for a packet, just to make sure that the next
process wanting to bind a socket to that ipaddr/port combo won't get your
traffic.  You can see how this could be a security breach.  Of course sometimes
you don't care, and usually when writing daemons you don't care.  In that
case, you use setsockopt to pull up the boolean SO_REUSEADDR to tell the TCP
software that you want to allow a bind to that address no matter what.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#include <sys/types.h>
#include <sys/socket.h>

. . .

    int reuse=1

    setsockopt(socket,SO_REUSEADDR,(const char *)&reuse);

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
SO_LINGER

SO_LINGER applies to information you've already "sent" when you want to close
a socket, but that is still in the queue...i.e. it's been sent by your code, but
not yet by TCP.  Normally when you do a close on a socket, it returns right
away, but SO_LINGER can be used to say that we want our choice of two other
behaviors.

1) Linger time non-zero, we want the close() to wait until all pending data is
   sent and acknowledged, or until the linger time set in the linger struct
   has passed.  (Most current implementations don't actually use the time,
   they just note whether it's non-zero despite what I just said, and the
   comment below says.)

   struct  linger {
        int     l_onoff;                /* option on/off */
        int     l_linger;               /* linger time */
   };

2) Linger time zero. On close() abort the connection, discard all pending 
   packets, and send a flush down the connection.  This is actually the most
   common use.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#include <sys/types.h>
#include <sys/socket.h>

. . .

    struct linger l;

    l.l_onoff=0;

    setsockopt(socket,SO_REUSEADDR,(const char *)&l);

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
I hope this is helpful.

--

     These opinions are mine, and not Amdahl's (except by coincidence;).

    ~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~  
  /                        |                           | (\               \
 |  Patrick J. Horgan      |  Amdahl Corporation       |  \\    Have       |
 |  patrick@oes.amdahl.com |  1250 East Arques Avenue  |   \\  _ Sword     |
 |  Phone : (408)992-2779  |  P.O. Box 3470 M/S 316    |    \\/    Will    |
 |  FAX   : (408)773-0833  |  Sunnyvale, CA 94088-3470 |   _/\\     Travel |
  \                        |  O16-2294                 |      \)          /
   ~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~

-----------[000105][next][prev][last][first]----------------------------------------------------
Date:      05 Dec 1994 19:47:47 GMT
From:      croten@eosqa.gsfc.nasa.gov (Charles Roten)
To:        comp.unix.solaris,comp.sys.sun.admin,comp.protocols.tcp-ip
Subject:   OK .. now, how to get the SPARCserver 1000 to see it's router ??

In article <CROTEN.94Dec2155624@eosqa.gsfc.nasa.gov> croten@eosqa.gsfc.nasa.gov (Charles Roten) writes:

>I'm setting up a SPARCserver 1000 at GSFC, running SunOS 5.3.  Everyything 
>is Jake now but the connection to the local network.  Running 'ifconfig le0' 
>from the command line tells me why .. the broadcast address is hosed.  
 
>Not too surprising .. it _is_ a new system.  
 
>Under 4.1.3, for a Sun 4/260, I put the line 
 
>    ifconfig ie0 128.183.117.34 netmask 255.255.255.o broadcast 128.183.117.0
 
>in /etc/rc.local.  In SGI System V, /etc/config/ifconfig-1.options reads 
 
>    broadcast 192.107.190.191
>    netmask 255.255.255.192
 
>and the same job gets done.  
 
>But Stunned MicrobeSystems doesn't _use_ straight Sys V R4.  So _where_, 
>for Pete's sake, do I set the netmask ?  And _HOW_ ??  I've fgrepped the 
>whole bloody /etc tree for 'broadcast' and 'netmask'.  Nothing productive.  
>'ifconfig' only shows up in a file, /etc/init.d/inetsvc, which seems to 
>only be used for NIS configuration.  We don't _use_ NIS here, we use DNS, 
>so the file's a template on another Sun running 5.3 at this site.  I'm at 
>the end of my rope .. how the _HELL_ do I get the netmask and broadcast 
>address right on boot-up ??  

OK .. I figured it out.  The file is /etc/netmasks, and the format for the 
broadcast address is _wierd_.  

Check this out ... from another SunOS 5.3 host ...

    cddev1.gsfc.nasa.gov% cat /etc/netmasks
    #
    # The netmasks file associates Internet Protocol (IP) address
    # masks with IP network numbers.
    # 
    #       network-number  netmask
    #
    # Both the network-number and the netmasks are specified in
    # "decimal dot" notation, e.g:
    #
    #               128.32.0.0 255.255.255.0
    #
    128.183.0.0     255.255.255.0
    ^^^^^^^^^^^
    
Then .. 
    
    cddev1.gsfc.nasa.gov% /sbin/ifconfig ie0
    ie0: flags=863<UP,BROADCAST,NOTRAILERS,RUNNING,MULTICAST> mtu 1500
            inet 128.183.117.37 netmask ffffff00 broadcast 128.183.117.255
                                                           ^^^^^^^^^^^^^^^
Wierd ...  

Now, 'ifconfig le0' shows the new interface to be righteous.  But the new 
machine _cannot_ find it's router.  The twisted pair cable _is_ plugged in 
to the proper places at both ends.  

I know the gateway for this box ('spsosun') should be 128.183.117.1.  It's 
_own_ IP address is 128.183.117.20.  But 'netstat -r' gives the following ...


    Destination      Gateway            Flags    MTU    RTT RTTvar    Use Interface
    localhost        localhost          UH         0      0      0   5695  lo0
    128.183.117.0    spsosun            U          ........................le0


'ping' from spsosun fails for both named hosts _and_ hosts specified by IP 
number.  

Can someone enlighten me about this?  What in chaos is going on?  The system's 
IP number and name are in /etc/hosts.  /etc/nsswitch.conf is set up for DNS.  
What have I overlooked?  Thanks.  

--
Charles D. Roten               | Hughes STX Inc.
croten@nyx.cs.du.edu           | NASA GSFC (Hurrah DAAC!)
croten@eosdata.gsfc.nasa.gov   | (301) 286-4413 (w), (301) 317-0872 (h)

-----------[000106][next][prev][last][first]----------------------------------------------------
Date:      5 Dec 1994 16:58:34 +0200
From:      root@cacofonix.utr.ac.za (root)
To:        comp.protocols.tcp-ip
Subject:   Re: SLIP pins for RS232

css@netcom.com (Chris Schefler) writes:

>I'm setting up a hardwired SLIP connection from my UNIX server to a PC
>using RS232 connectors with modular adapters (RJ25).  
 
>My question is this: which of the 25 pins are required for SLIP?

I am trying to do something similar. Pointers to a faq on
hardware/software configuration of slip links would be much appreciated.
Or if some boff out there could kindly post a 'minimum cable' diagram. (
just 2,3 and 7 work ok for simple login access via an RS232 port on SCO
unix - not sure whether this would also work with Slip connections ? )

Thanks
Neil


-----------[000107][next][prev][last][first]----------------------------------------------------
Date:      5 Dec 1994 20:59:47 GMT
From:      ais007@news.salford.ac.uk (Richard Letts)
To:        comp.protocols.tcp-ip
Subject:   Re: IP multicasting

In article <3bhrp7$7ui@ubszh.fh.zh.ubs.com> bsz@dla831.ubszh.net.ch (Carl Binding) writes:
>From: bsz@dla831.ubszh.net.ch (Carl Binding)
>Subject: IP multicasting
>Date: 30 Nov 1994 12:39:03 GMT
>Keywords: IP multicasting, routers
>I'm looking for the relevant information regd. IP multicasting, i.e. could anyone please quote me the list of relevant RFCs? 
>Kindly also point me into the direction of an ftp-mail server where I can fetch these
>from.
>Lastly: which routers currently support the IP multicast protocol? There is a big
>debate here, claiming that for instance CISCO does currently not support IP m-cast in
>their products. Is that correct? 
>Please reply directly to bsz@ubszh.net.ch.
>Thanks for any info.
>Carl
>

cisco IOS 10.2 supports muticast  - either PIM or DVMRP

if you are lucky enough to have a router that can run it ...

Richard
Richard Letts  
-------------------------------------------------------------------------------
Network Manager                               mail:    R.J.Letts@salford.ac.uk
University of Salford                         phone:     +44 61 745 5252
Great Britain                                 fax:       +44 61 745 5888

-----------[000108][next][prev][last][first]----------------------------------------------------
Date:      Tue, 6 Dec 1994 12:05:14 -0800
From:      Bjorn Eng <bjorn@tetons.jpl.nasa.gov>
To:        comp.protocols.tcp-ip
Subject:   T3000 settings for SLIP?

Hi,

   I'm trying to use a Telebit T3000 modem at the server side of a CSLIP 
connection. Does anyone have a good initialization string for SLIP?

TIA,

	Bjorn Eng 
        Jet Propulsion Laboratory


-----------[000109][next][prev][last][first]----------------------------------------------------
Date:      Tue, 06 Dec 1994 07:15:52 -0500
From:      droms@bucknell.edu (Ralph Droms)
To:        comp.protocols.tcp-ip
Subject:   Re: DHCP server for Unix?????

In article <HOLBOR.13.0515D827@bus.orst.edu>, HOLBOR@bus.orst.edu (Richard
Holbo) wrote:

> Subject says it all.  I am looking for a Unix DHCP server preferrably BSDI or 
> HPUX.  I've seen several notes requesting information about DHCP servers
 here, 
> but no replies have been posted to the list.

Sorry - there don't seem to be any freely available DHCP servers available
as yet.  There are some commercial versions; I'll know more after this
week's IETF meeting.

There is a mailing list for DHCP-related issues:
host-conf@sol.cs.bucknell.edu (host-conf-request for adds/deletes).  The
availability of implementations, as well as technical issues and status of
DHCP as an Internet standard protocol, is discussed there.

-----------[000110][next][prev][last][first]----------------------------------------------------
Date:      5 Dec 94 23:01:09 GMT
From:      ddl@harvard.edu (Dan Lanciani)
To:        comp.protocols.tcp-ip
Subject:   Re: Shutdown() block proof?

In article <3bv64t$n2e@crl6.crl.com>, cgi@crl.com (Paul Smith) writes:
|I'm trying to craft a sockets wrapper library that will not block indeffinately
|on a close of a TCP service NOR interact with a child task's use of the fd
|after a perent server forks (concurrant TCP server).
|
|1) I've set SO_LINGER to 10 seconds, for a reasonalby short pause for the
|other end to read out standing packets.
|
|2) I was setting the fd to O_NONBLOCK prior to calling close to avoid potential
|indefinite blocks..
|
|I've discovered the following architectural interactions:
|
|1) For a TCP concurrant server, where the main server forks on every accept()
|then closes the new fd, the child closes the main fd.  There can be interaction
|from both tasks acting on the same fd control block in the kernel if the
|library
|I wrote sets the fd to O-NONBLOCK prior to calling close. 
|
|In the case of the parent closing the new client accept() fd, it sets the fd
|to O_NONBLOCK and the child task's use of this socket, then becomes
|non-blocking.
|OK, ok, I didn't think of that when I coded my close library function. Then how
|to I craft a robust close wrapper given that it will be used for both
|sequential and concurrant server and client environments???
|
|Do I need two (or more) close wrapper functions; 1 for orderly shutdown;
|SO_LINGER
|set to some reasonable short time, set to O_NONBLOCK ?? (or is this at odds
|with
|SO_LINGER??) then close.  2) don't fool with it close, just call close, DON't
|call shutdown, or set to O_NONBLOCK.

I quoted the whole message because I think the problem isn't well captured
in any one part of the question.  The important thing to realize is that
most/all of the operations you are using act on the underlying object
(the socket) and not on the descriptor you are using to reference it.  This
certainly applies to shutdown() and the SO_LINGER socket option.  In most
systems this also applies to the blocking state, though I have heard of
cases where such state is maintained pre-file-descriptor.  So, anything
you do is going to change the socket's characteristics as seen by your
child process.

If your main concern is that the parent never be allowed to block on
a close, you might want to consider some way of synchronizing the child
and parent.  Understand that the only way the close() call can block
on a socket is if this is the final close(), i.e., there are no other
file descriptors associated with the socket.  In any other case, close()
merely decrements a reference count.  Unless your child dies or deliberately
closes its descriptor before the parent can close its own, you are safe.
I think most implementations consider it safe enough that the parent closes
immediately after the fork().  You could add a sleep in the child to further
bias the odds in your favor.  To be certain, you could employ some IPC mechanism
to block the child until the parent has completed the close().  Ultimately,
I suspect this is all overkill.  The circumstances under which the final
close() of a newly accept()ed socket can block for a significant time
are rather obscure, perhaps even bugs. :)

				Dan Lanciani
				ddl@harvard.*

-----------[000111][next][prev][last][first]----------------------------------------------------
Date:      Mon, 5 Dec 1994 23:28:21 GMT
From:      vjs@calcite.rhyolite.com (Vernon Schryver)
To:        comp.protocols.tcp-ip
Subject:   Re: SLIP pins for RS232

In article <3bv9qq$h2@cacofonix.utr.ac.za> root@cacofonix.utr.ac.za (root) writes:
>css@netcom.com (Chris Schefler) writes:
>
>>I'm setting up a hardwired SLIP connection from my UNIX server to a PC
>>using RS232 connectors with modular adapters (RJ25).  
 
>>My question is this: which of the 25 pins are required for SLIP?
>
>I am trying to do something similar. Pointers to a faq on
>hardware/software configuration of slip links would be much appreciated.
>Or if some boff out there could kindly post a 'minimum cable' diagram. (
>just 2,3 and 7 work ok for simple login access via an RS232 port on SCO
>unix - not sure whether this would also work with Slip connections ? )
> ...

You also need RTS and CTS (pins 4 and 5 on a DB-25) to use "hardware flow
control", which is practically required at speeds above 4800 bit/sec.

You probably also need DTR and DCD (pins 20 and 8 on a DB-25) if you
care to know when the link has gone down to shut down the PPP or SLIP link
when the modem hangs up, or to make the modem hang up when you shut down
PPP or SLIP.


Vernon Schryver    vjs@rhyolite.com

-----------[000112][next][prev][last][first]----------------------------------------------------
Date:      Mon, 5 Dec 1994 23:48:26 GMT
From:      gnn@netcom.com (George Neville-Neil)
To:        comp.protocols.tcp-ip
Subject:   FAQ for December



Archive-name:tcp-ip/FAQ
Last-modified:  1994/12/5

Internet Protocol Frequently Asked Questions

Maintained by: George V. Neville-Neil (gnn@netcom.com)
Contributions from:
Ran Atkinson
Stephane Bortzmeyer
Rodney Brown
Dr. Charles E. Campbell Jr.
Phill Conrad 
Alan Cox
Rick Jones
Jon Kay 
Jay Kreibrich
William Manning
Barry Margolin 
Jim Muchow
Subu Rama
W. Richard Stevens 
 
Version 1.8


************************************************************************

	The following is a list of Frequently Asked Questions, and
their answers, for people interested in the Internet Protocols,
including TCP, UDP, ICMP and others.  Please send all additions,
corrections, complaints and kudos to the above address.  This FAQ will
be posted on or about the first of every month.

	This FAQ is available for anonymous ftp from :
ftp.netcom.com:/pub/gnn/tcp-ip.faq .  You may get it from my home page at
ftp://ftp.netcom.com/pub/gnn/gnn.html

************************************************************************

Table of Contents:
Glossary
1) Are there any good books on IP?
2) Where can I find example source code for TCP/UDP/IP?
3) Are there any public domain programs to check the performance of an
IP link? 
4) Where do I find RFCs?
5) How can I detect that the other end of a TCP connection has
crashed?  Can I use "keepalives" for this?
6) Can the keepalive timeouts be configured?
7) Can I set up a gateway to the Internet that translates IP
addresses, so that I don't have to change all our internal addresses 
to an official network? 
8) Are there object-oriented network programming tools?
9) What other FAQs are related to this one?
10) What newsgroups contain information on networks/protocols?

Glossary:

I felt this should be first given the plethora of acronyms used in the
rest of this FAQ.

IP: Internet Protocol.  The lowest layer protocol defined in TCP/IP.
This is the base layer on which all other protocols mentioned herein
are built.  IP is often referred to as TCP/IP as well.

UDP: User Datagram Protocol.  This is a connectionless protocol built
on top of IP.  It does not provide any guarantees on the ordering or
delivery of messages.  This protocol is layered on top of IP.

TCP: Transmission Control Protocol.  TCP is a connection oriented
protocol that guarantees that messages are delivered in the order in
which they were sent and that all messages are delivered.  If a TCP
connection cannot deliver a message it closes the connection and
informs the entity that created it.  This protocol is layered on top
of IP.

ICMP:  Internet Control Message Protocol.  ICMP is used for
diagnostics in the network.  The Unix program, ping, uses ICMP
messages to detect the status of other hosts in the net.  ICMP
messages can either be queries (in the case of ping) or error reports,
such as when a network is unreachable.

RFC: Request For Comment.  RFCs are documents that define the
protocols used in the IP Internet.  Some are only suggestions, some
are even jokes, and others are published standards.  Several sites in
the Internet store RFCs and make them available for anonymous ftp.

SLIP:  Serial Line IP.  An implementation of IP for use over a serial
link (modem).  CSLIP is an optimized (compressed) version of SLIP that
gives better throughput.

Bandwidth:  The amount of data that can be pushed through a link in
unit time.  Usually measured in bits or bytes per second.

Latency:  The amount of time that a message spends in a network going
from point A to point B.

Jitter:  The effect seen when latency is not a constant.  That is, if
messages experience a different latencies between two points in a
network.

RPC:  Remote Procedure Call.  RPC is a method of making network access
to resource transparent to the application programmer by supplying a
"stub" routine that is called in the same way as a regular procedure
call.  The stub actually performs the call across the network to
another computer.

Marshalling:  The process of taking arbitrary data (characters,
integers, structures) and packing them up for transmission across a
network.

MBONE: A virtual network that is a Multicast backBONE.  It is still a
research prototype, but it extends through most of the core of the
Internet (including North America, Europe, and Australia).  It uses IP
Multicasting which is defined in RFC-1112.  An MBONE FAQ is available
via anonymous ftp from: ftp.isi.edu" There are frequent broadcasts of
multimedia programs (audio and low bandwidth video) over the MBONE.
Though the MBONE is used for mutlicasting, the long haul parts of the
MBONE use point-to-point connections through unicast tunnels to
connect the various multicast networks worldwide.


1) Are there any good books on IP?

A) Yes.  Please see the following:

Internetworking with TCP/IP Volume I
(Principles, Protocols, and Architecture)
Douglas E. Comer
Prentice Hall 1991 ISBN 0-13-468505-9

This volume covers all of the protocols, including IP, UDP, TCP, and
the gateway protocols.  It also includes discussions of higher level
protocols such as FTP, TELNET, and NFS.

Internetworking with TCP/IP Volume II
(Design, Implementation, and Internals)
Douglas E. Comer / David L. Stevens
Prentice Hall 1991  ISBN 0-13-472242-6

Discusses the implementation of the protocols and gives numerous code
examples.

Internetworking with TCP/IP Volume III (BSD Socket Version)
(Client - Server Programming and Applications)
Douglas E. Comer / David L. Stevens
Prentice Hall 1993  ISBN 0-13-474222-2

This book discusses programming applications that use the internet
protocols.  It includes examples of telnet, ftp clients and servers.
Discusses RPC and XDR at length.

TCP/IP Illustrated, Volume 1: The Protocols, 
W. Richard Stevens
(c) Addison-Wesley, 1994  ISBN 0-201-63346-9

An excellent introduction to the entire TCP/IP protocol suite,
covering all the major protocols, plus several important applications.

Unix Network Programming
W. Richard Stevens
Prentice Hall 1990  ISBN 0-13-949876

An excellent introduction to network programming under Unix.

The Design and Implementation of the 4.3 BSD Operating System
Samuel J. Leffler, Marshall Kirk McKusick, Michael J. Karels, John S.
Quarterman 
Addison-Wesley 1989  ISBN 0-201-06196-1

Though this book is a reference for the entire operating system, the
eleventh and twelfth chapters completely explain how the networking
protocols are implemented in the kernel.


2)  Where can I find example source code for TCP/UDP/IP?

A)  Code from the Internetworking with TCP/IP Volume III is available
for anonymous ftp from:

arthur.cs.purdue.edu:/pub/dls

Code used in the Net-2 version of Berkeley Unix is available for
anonymous ftp from:

ftp.uu.net:systems/unix/bsd-sources/sys/netinet 

and

gatekeeper.dec.com:/pub/BSD/net2/sys/netinet

Code from Richard Steven's book is available on:
ftp.uu.net:/published/books/stevens.*

Example source code and libraries to make coding quicker is available
in the Simple Sockets Library written at NASA.  The Simple Sockets
Library makes sockets easy to use!  And, it comes as source code.  It
has been tested on: Unix (SGI, DecStation, AIX, Sun 3, Sparcstation;
version 2.02+: Solaris 2.1, SCO), VMS, and MSDOS (client only since
there's no background there).  It is provided in source code form, of
course, and sits atop Berkeley sockets and tcp/ip.

You can order the "Simple Sockets Library" from

                           Austin Code Works
                          11100 Leafwood Lane
                       Austin, TX 78750-3464 USA
                         Phone (512) 258-0785
                 
Ask for the "SSL - The Simple Sockets Library".  Last I checked, they
were asking $20 US for it.


For DOS there is WATTCP.ZIP (numerous sites): 

WATTCP is a DOS TCP/IP stack derived from the NCSA Telnet program and
much enhanced. It comes with some example programs and complete source
code. The interface isn't BSD sockets but is well suited to PC type
work. It is also written so that it can be used and memory
allocation).

3)  Are there any public domain programs to check the performance of
an IP link?

A)  

TTCP:  Available for anonymous ftp from....

Host gatekeeper.dec.com

    Location: /.0/BSD/NetBSD/NetBSD-current/othersrc
      DIRECTORY dr-xr-xr-x        512  Apr  8 09:57  ttcp
    Location: /.0/BSD/NetBSD/NetBSD-current/othersrc/ttcp
           FILE -r--r--r--       3885  Nov  7 03:35  ttcp.1
           FILE -r--r--r--      19225  Nov  7 03:35  ttcp.c

Host world.std.com

    Location: /src/wuarchive/graphics/graphics/mirrors/sgi.com/sgi/src/ttcp
           FILE -r--r--r--       3885  Oct  4 1991  ttcp.1
           FILE -r--r--r--      19170  May 17 1993  ttcp.c
           FILE -r--r--r--      13033  Sep  5 1989  ttcp.c-brl

On ftp.sgi.com are netperf (from Rick Jones at HP) and nettest
(from Dave Borman at Cray).  ttcp is also availabel at ftp.sgi.com.

Netperf can also be fotten using the Web from:
ftp://ftp.cup.hp.com/dist/networking/benchmarks

as well as:

ftp://col.hp.com/dist/networking/benchmarks



There is suite of Bandwidth Measuring programs from gnn@netcom.com.
Available for anonymous ftp from ftp.netcom.com in
~ftp/gnn/bwmeas-0.3.tar.Z These are several programs that meausre
bandwidth and jitter over several kinds of IPC links, including TCP
and UDP.


4) Where do I find RFCs?

A)  This is the latest info on obtaining RFCs:
Details on obtaining RFCs via FTP or EMAIL may be obtained by sending
an EMAIL message to rfc-info@ISI.EDU with the message body 
help: ways_to_get_rfcs.  For example:

        To: rfc-info@ISI.EDU
        Subject: getting rfcs

        help: ways_to_get_rfcs

The response to this mail query is quite long and has been omitted.

RFCs can be obtained via FTP from DS.INTERNIC.NET, NIS.NSF.NET,
NISC.JVNC.NET, FTP.ISI.EDU, WUARCHIVE.WUSTL.EDU, SRC.DOC.IC.AC.UK,
FTP.CONCERT.NET, or FTP.SESQUI.NET.


Using Web, WAIS, and gopher:

Web:

http://web.nexor.co.uk/rfc-index/rfc-index-search-form.html

WAIS access by keyword:

wais://wais.cnam.fr/RFC

Excellent presentation with a full-text search too:

http://www.cis.ohio-state.edu/hypertext/information/rfc.html

With Gopher:

gopher://r2d2.jvnc.net/11/Internet%20Resources/RFC
gopher://muspin.gsfc.nasa.gov:4320/1g2go4%20ds.internic.net%2070%201%201/.ds/
.internetdocs



5) How can I detect that the other end of a TCP connection has crashed?
Can I use "keepalives" for this?

A) Detecting crashed systems over TCP/IP is difficult.  TCP doesn't require
any transmission over a connection if the application isn't sending
anything, and many of the media over which TCP/IP is used (e.g. ethernet)
don't provide a reliable way to determine whether a particular host is up.
If a server doesn't hear from a client, it could be because it has nothing
to say, some network between the server and client may be down, the server
or client's network interface may be disconnected, or the client may have
crashed.  Network failures are often temporary (a thin ethernet will appear
down while someone is adding a link to the daisy chain, and it often takes
a few minutes for new routes to stabilize when a router goes down), and TCP
connections shouldn't be dropped as a result.

Keepalives are a feature of the sockets API that requests that an empty
packet be sent periodically over an idle connection; this should evoke an
acknowledgement from the remote system if it is still up, a reset if it has
rebooted, and a timeout if it is down.  These are not normally sent until
the connection has been idle for a few hours.  The purpose isn't to detect
a crash immediately, but to keep unnecessary resources from being allocated
forever.

If more rapid detection of remote failures is required, this should be
implemented in the application protocol.  There is no standard mechanism
for this, but an example is requiring clients to send a "no-op" message
every minute or two.  An example protocol that uses this is X Display
Manager Control Protocol (XDMCP), part of the X Window System, Version 11;
the XDM server managing a session periodically sends a Sync command to the
display server, which should evoke an application-level response, and
resets the session if it doesn't get a response (this is actually an
example of a poor implementation, as a timeout can occur if another client
"grabs" the server for too long).

6) Can the keepalive timeouts be configured?

A) I know they can on many systems, but I don't know the details.

7) Can I set up a gateway to the Internet that translates IP addresses, so
that I don't have to change all our internal addresses to an official
network?

A) There's no general solution to this.  Many protocols include IP
addresses in the application-level data (FTP's "PORT" command is the most
notable), so it isn't simply a matter of translating addresses in the IP
header.  Also, if the network number(s) you're using match those assigned
to another organization, your gateway won't be able to communicate with
that organization (RFC 1597 proposes network numbers that are reserved for
private use, to avoid such conflicts, but if you're already using a
different network number this won't help you).

However, if you're willing to live with limited access to the Internet from
internal hosts, the "proxy" servers developed for firewalls can be used as
a substitute for an address-translating gateway. See the firewall FAQ.

8) Are there object-oriented network programming tools?

A) Yes, and one such system is called ACE (ADAPTIVE Communication
Environment).  Here is how to get more information and the software:

OBTAINING ACE

An HTML version of this README file is available at URL
http://www.cs.wustl.edu/~schmidt/ACE.html.  All software and
documentation is available via both anonymous ftp and the Web.

ACE is available for anonymous ftp from the ics.uci.edu (128.195.1.1)
host in the gnu/C++_wrappers.tar.Z file (approximately .5 meg
compressed).  This release contains contains the source code,
documentation, and example test drivers for C++ wrapper libras.

9) What other FAQs might you want to look in?
comp.protocols.tcp-ip.ibmpc
   Aboba, Bernard D.(1994) "comp.protocols.tcp-ip.ibmpc Frequently
    Asked Questions (FAQ)" Usenet news.answers, available via
    file://ftp.netcom.com/pub/mailcom/IBMTCP/ibmtcp.zip,
    57 pages.

comp.protocols.ppp
   Archive-name: ppp-faq/part[1-8]
   URL: http://cs.uni-bonn.de/ppp/part[1-8].html

comp.dcom.lans.ethernet
   ftp site: dorm.rutgers.edu, pub/novell/DOCS
   Ethernet Network Questions and Answers
   Summarized from UseNet group comp.dcom.lans.ethernet

10) What other newsgroups deal with networking?

comp.dcom.cabling       Cabling selection, installation and use.
comp.dcom.isdn          The Integrated Services Digital Network
			(ISDN).
comp.dcom.lans.ethernet Discussions of the Ethernet/IEEE 802.3
			protocols.
comp.dcom.lans.fddi     Discussions of the FDDI protocol suite.
comp.dcom.lans.misc     Local area network hardware and software.
comp.dcom.lans.token-ring       Installing and using token ring
				networks.
comp.dcom.servers       Selecting and operating data communications
			servers.
comp.dcom.sys.cisco     Info on Cisco routers and bridges.
comp.dcom.sys.wellfleet Wellfleet bridge & router systems hardware &
			software.
comp.protocols.ibm      Networking with IBM mainframes.
comp.protocols.iso      The ISO protocol stack.
comp.protocols.kerberos The Kerberos authentication server.
comp.protocols.misc     Various forms and types of protocol.
comp.protocols.nfs      Discussion about the Network File System
			protocol.
comp.protocols.ppp      Discussion of the Internet Point to Point
			Protocol.
comp.protocols.smb      SMB file sharing protocol and Samba SMB
			server/client.
comp.protocols.tcp-ip   TCP and IP network protocols.
comp.protocols.tcp-ip.ibmpc     TCP/IP for IBM(-like) personal
				computers.
comp.security.misc      Security isuipment for the PC.
comp.os.ms-windows.networking.misc      Windows and other networks.
comp.os.ms-windows.networking.tcp-ip    Windows and TCP/IP networking.
comp.os.ms-windows.networking.windows   Windows' built-in networking.
comp.os.os2.networking.misc     Miscellaneous networking issues of
				OS/2.
comp.os.os2.networking.tcp-ip   TCP/IP under OS/2.
comp.sys.novell         Discussion of Novell Netware products.
-- 
gnn@netcom.com

Law is there to clean up etiquette's failures.
					Ms. Manners

-----------[000113][next][prev][last][first]----------------------------------------------------
Date:      6 Dec 1994 00:06:00 GMT
From:      dries@elvis.rowan.edu (PowerSlave)
To:        comp.os.ms-windows.apps.comm,comp.os.ms-windows.networking.tcp-ip,comp.protocals.tcp-ip,alt.dcom.slip-emulators
Subject:   Twinsock 1.2 on IRIX 5.2

Hello All.

I've been trying to get a successful TSHOST going since TwinSock v1.0,
to no avail.

I've been able to get tshost v1.1 and 1.2 to compile fine (1.1 after a
bit of editing) on a SunOS machine (SPARC), however, my local host is an
SGI Indy running IRIX 5.2. I've been able to compile tshost 1.1 and 1.2
find on the Indy, with just a couple of edits (on 1.1, 1.2 compiles
fine) ... The problem I have is after the TwinSock gets connected,
running applications:

I've tried the following under OS/2 v2.10 WinOS2, OS/2 v2.11 WinOS2, and
Windows for Workgroups v3.11, I get the same result in all
circumstances:

Twinsock starts `normally', without any errors or incident. I fire up
Netscape 0.94, and on the SunOS built tshost, everything works fine.
Under the IRIX 5.2 built tshost, I always get a `Twinsock caused a GPF
in WINSOCK.DLL at address 0001:14AC' .. the address almost never
changes, although I have gotten a 0002:xxxx once.

Unless the SGI Indy is 64bit... then perhaps that is my problem. Anyone
working on a 64bit port?

-----------[000114][next][prev][last][first]----------------------------------------------------
Date:      6 Dec 1994 00:39:06 GMT
From:      dries@elvis.rowan.edu (PowerSlave)
To:        comp.os.ms-windows.apps.comm,comp.os.ms-windows.networking.tcp-ip,comp.protocols.tcp-ip,alt.dcom.slip-emulators
Subject:   Re: Twinsock 1.2 on IRIX 5.2

dries@elvis.rowan.edu (PowerSlave) writes:

>Hello All.
 
>I've been trying to get a successful TSHOST going since TwinSock v1.0,
>to no avail.
 
>I've been able to get tshost v1.1 and 1.2 to compile fine (1.1 after a
>bit of editing) on a SunOS machine (SPARC), however, my local host is an
>SGI Indy running IRIX 5.2. I've been able to compile tshost 1.1 and 1.2
>find on the Indy, with just a couple of edits (on 1.1, 1.2 compiles
>fine) ... The problem I have is after the TwinSock gets connected,
>running applications:
 
>I've tried the following under OS/2 v2.10 WinOS2, OS/2 v2.11 WinOS2, and
>Windows for Workgroups v3.11, I get the same result in all
>circumstances:
 
>Twinsock starts `normally', without any errors or incident. I fire up
>Netscape 0.94, and on the SunOS built tshost, everything works fine.
>Under the IRIX 5.2 built tshost, I always get a `Twinsock caused a GPF
>in WINSOCK.DLL at address 0001:14AC' .. the address almost never
>changes, although I have gotten a 0002:xxxx once.
 
>Unless the SGI Indy is 64bit... then perhaps that is my problem. Anyone
>working on a 64bit port?

To follow up on my own post, no Scarlet, the SGI Indy is not 64bit... It
outputs the same size ints as the SPARC running SunOS, so that isn't the
reason. There goes that theory. 

So, can anyone else suggest a reason that Twinsock likes SunOS and not
IRIX?

-----------[000115][next][prev][last][first]----------------------------------------------------
Date:      6 Dec 1994 18:30:00 -0800
From:      heffron@falstaff.css.beckman.com (Matt Heffron)
To:        comp.protocols.tcp-ip
Subject:   Query: Internal net with multiple Internet connections. Routing?

Beckman currently has a class B network that has a single Internet connection
through a T1 connected to a firewall (router):

[Eurocenter]------[internal router]----------[firewall]--------[Internet]...
             56kb   |               Ethernet     |         T1
                    |__other subnets             |__public server(soon)

Part of the internal net is a 56kb line to the Beckman Eurocenter in Nyon.
The teleprocessing group is investigating adding an Internet connection in the
Eurocenter so that external e-mail (customers,...) to the Eurocenter doesn't have
to come in through Southern California, then back to Europe over the 56kb line.

[Eurocenter]------[internal router]----------[firewall]--------[Internet]...
  |          56kb   |               Ethernet     |         T1
  |                 |__other subnets             |__public server(soon)
  |_[firewall]---[Internet]...

The question arises.  Can the Internet providers actually route the traffic so
that the traffic to the Eurocenter enters via Nyon, and the traffic to the rest
of Beckman enters through the Southern California connection.  Especially
troublesome seems to be the public server which must be accessed only through the
Southern California connection (due to the firewalls preventing public traffic
from traversing the internal net).

The only idea that I have been able to come up with is to get the Eurocenter
assigned a new Class C address group.  This would keep the external traffic
"localized" to the appropriate Internet connection.  Also, make sure that the
internal routing information between the SoCal class B and the Eurocenter
class C is not exported, so that the "outside world" doesn't know the two nets
are connected.

My questions here are:
   Are there other options that don't require a separate class C net?
   What issues have I overlooked?
   What critical wrong assumptions have I made?
   ...

Thanks in advance for any and all advice,

Matt Heffron
-- 
Matt Heffron           heffron@falstaff.css.beckman.com
Beckman Instruments, Inc.         voice: (714) 961-3128
2500 N. Harbor Blvd. MS X-10, Fullerton, CA 92634-3100
I don't speak for Beckman Instruments (or CRFG) unless they say so.

-----------[000116][next][prev][last][first]----------------------------------------------------
Date:      Tue, 6 Dec 1994 05:31:51 GMT
From:      wfp@world.std.com (William F Phillips)
To:        comp.os.ms-windows.apps.comm,comp.os.ms-windows.networking.tcp-ip,comp.protocals.tcp-ip,alt.dcom.slip-emulators
Subject:   Re: Twinsock 1.2 on IRIX 5.2

In article <3c09t8$ich@gboro.rowan.edu>,
PowerSlave <dries@elvis.rowan.edu> wrote:
:Hello All.
:
:I've been trying to get a successful TSHOST going since TwinSock v1.0,
:to no avail.
:
:I've been able to get tshost v1.1 and 1.2 to compile fine (1.1 after a
:bit of editing) on a SunOS machine (SPARC), however, my local host is an
:SGI Indy running IRIX 5.2. I've been able to compile tshost 1.1 and 1.2
:find on the Indy, with just a couple of edits (on 1.1, 1.2 compiles
:fine) ... The problem I have is after the TwinSock gets connected,
:running applications:
...

Similar experience here; tho' the SGI is an 8-cpu Challenge.  I'd really
like to see this work, but ...

I've tried a number of winsock clients with it, but the most spectacular
results are with NCSA Mosaic, which occasionally COLD-boots my machine.
Cute.  Other than that, no app seems to be able to do anything.
-- 
   ___                ___    _                        |     Bill Phillips
  (/__) .  /) /)     (/__) /_   .  /) /) .  _   _     |  ShoeString Projects
  /__) (__(__(___    /    /  )_(__(__(__(_ /_)_/_)_   | =DTP=Graphic Design=
 (                  (                     (           | =Computer Consulting=

-----------[000117][next][prev][last][first]----------------------------------------------------
Date:      6 Dec 1994 14:08:50 -0500
From:      rjm@panix.com (rjm)
To:        comp.protocols.tcp-ip
Subject:   dynamic routing & nfs

 I have a c/s application that requires static routing between the 
 server and its clients. between the server and the clients is router A
 cisco configured dynamically dual T1 lines , then router B then the clients.
 
ex: 

  server<------>router A<----> T1<---->router B<--->clients

  the telcom (network owners) are telling me that these twp routers
are ciscos configured to be dynamic routers. Is there any way that a 
static route can be configured along side the dynamic route?
or will I have to get them to assign me s compleatly independent
route with my own routers?  Env: AIX 3.2.5 ethernet applicationmakes extensive
use of NFS.

thanks,

hurring up and waiting,

rjm@panix.com 

-----------[000118][next][prev][last][first]----------------------------------------------------
Date:      6 Dec 1994 06:05:57 GMT
From:      whalenm@ulabsgi (Matthew Whalen)
To:        comp.protocols.tcp-ip,comp.unix.solaris,comp.unix.sys5.r4
Subject:   TLI t_snd flow restriction

I'm writing an application that listens and sends data on a TLI connection
endpoint simultanously.  I have one (solaris) thread that listens
on the endpoint for data to arrive using t_rcv, while another
thread wants to send data on the endpoint using t_snd.  It seems
that the t_snd gets blocked by a "flow control restriction" on the
endpoint, and so both frequently sit there blocked.  Short of making
the endpoint asynchronous, is there a way get around this flow control
restriction?  It seems silly to resort to asynchronous communication
on a TLI endpoint when I have a threads library.  Any ideas?  Thanks.

-matthew
whalenm@tis.telos.com _or_ whalenm@ulabsgi.gsfc.nasa.gov


-----------[000119][next][prev][last][first]----------------------------------------------------
Date:      Tue, 6 Dec 1994 06:42:36 GMT
From:      georgeb@netcom.com (George H. Bosworth)
To:        ba.seminars,comp.protocols.tcp-ip,comp.protocols.snmp,dcom.net-management
Subject:   SEF/UniForum Open Systems SIG, Tue. Dec. 13: Network Mgt.

                  SEF/UniForum Open Systems SIG
  
                    Network/System Management
                    -=-=-=-=-=-=-=-=-=-=-=-=-
  
  Network and system management have been at the center of considerable
  attention and investment throughout the industry over the last 4
  years.  How far have we come?  Are we meeting expectations?  What does
  the future look like? What is the formula or strategy for success in
  this environment?
  
  Andre Schwager, President of NetLabs, will discuss these issues. 
  NetLabs is a significant developer of products that help
  administrators to manage the collection of powerful communication and
  computer equipment as an integrated environment.
  
  Place:        Digital Equipment Corporation
                130 Lytton Street, Palo Alto CA
                (Corner of Alma, 1 block N. of University)
  
  Date & Time:  Tuesday, December 13, 1994, 7:00-9:00 p.m.
  
  Cost:         Free for SEF and full UniForum members and DEC employees;
                $5 for trial UniForum members; $10 for others.
                No reservation required; limit of 90 attendees.
  
  Information:  George Bosworth, 415/851-3304, georgeb@netcom.com
  
  Andre Schwager has 25 years of experience in data communications and
  networks.  He was GM of a network division at Hewlett-Packard and
  president of several start-up companies, and has been President of
  NetLabs for 4 years.
  
  SEF, the Software Entrepreneurs' Forum, started in 1983, is a leading
  Silicon Valley-based non-profit organization dedicated to software
  professionals, with over 900 members. SEF informs and educates its
  members on all facets of the software industry.  SEF sponsors 13
  other SIGs, which meet once a month: Business Operations, Client
  Server, Intelligent Systems, International, Internet, Macintosh,
  Marketing, Multimedia, Networking, Pen/Mobile, Visual Basic, Windows,
  and Wireless. Call 415/854-7219 for more SEF information.
  
  UniForum, The International Association of Open Systems Professionals,
  is a vendor-independent, not-for-profit professional association that
  helps individuals and their organizations increase their Information
  Systems effectiveness through the use of open systems, based on shared
  industry standards. Central to UniForum's mission is the delivery of
  high quality educations programs, trade shows and conferences,
  publications, on-line services, and peer group interactions.  Call
  800/255-5620 for more UniForum information.

-----------[000120][next][prev][last][first]----------------------------------------------------
Date:      Tue, 06 Dec 94 15:08:34 PDT
From:      Frank Belland <fbelland@dmewrk1.orl.mmc.com>
To:        mmc.orl.general,mmc.den.general,mmc.announce,comp.unix.osf.misc,comp.sys.hp.hpux,comp.sys.hp.hardware,comp.sys.hp.apps,comp.protocols.tcp-ip,comp.protocols.snmp
Subject:   Call for Papers 1995 OpenView Forum Conference



_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_
/_/_/_/_/
 
       _/_/_/                             _/          _/
    _/     _/                             _/        _/ _/
   _/     _/  _/_/_/    _/_/   _/     _/  _/      _/       _/_/  _/  
     _/
  _/     _/ _/    _/ _/   _/  _/_/   _/   _/    _/  _/  _/   _/  _/  
    _/
 _/     _/ _/    _/ _/_/_/_/ _/  _/ _/    _/  _/   _/  _/_/_/_/  _/  
_/ _/
 _/    _/ _/    _/ _/       _/   _/_/     _/_/    _/  _/         
_/_/_/_/
 _/_/_/  _/_/_/    _/_/_/  _/     _/      _/     _/    _/_/_/    _/  
 _/
        _/
       _/          _/_/_/_/_/
      _/          _/
                 _/        _/_/    _/_/_/  _/     _/  _/    _/
                _/_/_/  _/    _/  _/   _/ _/     _/  _/_/ _/_/
               _/      _/    _/  _/      _/     _/  _/  _/  _/
              _/      _/    _/  _/      _/     _/  _/       _/
             _/       _/_/_/   _/       _/_/_/    _/        _/
 
_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_
/_/_/_/_/
 
 
    _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
   _/
  _/              C A L L   F O R   P A P E R S
 _/
_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
                                                             _/
                                                            _/
         June 19-23, 1995 in Seattle Washington.           _/
                                                          _/
                                                         _/
 _/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
 
  Attention OpenView Users and Developers!!!!!!
 
The 1995 OpenView Forum Program Committee would like to announce
that it is accepting papers for presentation at the Third OpenView 
Forum
User/Developers conference on Integrated Network and Systems
Management (OVF '95) will be held from  June 19-23, 1995 in
Seattle Washington.
 
 
The OpenView Forum 1995 conference on Integrated Network and
Systems Management using the OpenView Framework will build on the
successes of OVF 93 & OVF 94 conferences as the central technical 
exchange
forum for the research, standards, development, systems integrator, 
vendor
and user community for OpenView network and systems management 
frameworks.
 
The OpenView Forum community exemplifies the increasing interest
in overall Enterprise management solutions across all types of 
networks,
supporting enterprise communication systems, distributed computing 
systems,
database management system, automated Operation systems, facilities 
management
systems, help desk management systems and client/server application
management.
 
 
Authors are invited to submit unpublished papers, as well as 
proposals for
tutorials, panel discussions, vendor/user/developer demonstrations, 
or
birds-of-a-feather sessions (informal discussion groups), in the 
following
areas:
 
       Presentations may be about any topic relevant to OpenView use 
on any
       of the OpenView platforms: UNIX DM, UNIX SNMP, or Windows.
 
       Suggestions for topics could include but are not limited to 
the
       following list:
 
         OpenView use in current business applications
 
         OpenView tutorials - either advanced or beginner levels
 
         Application of OpenView for systems management integration
 
         Customization of OpenView - benefits and examples
 
         Integration examples of OpenView and  legacy management 
products
 
         Monitoring of Database systems using RDBMS mib.
 
         Monitoring Email systems using MADMAN mib.
 
         Smart SNMP Agents, the next generation.
 
         Enterprize management for Automated Operations
 
                System Administration
 
                Automated Inventory (DMI)
 
                Software Distribution
 
                Physical Cable Plant and facilities management
                   Next Generation
 
          OpenView for the Telecommunication industry
 
          OpenView use in WAN management
 
          OpenView DM development - benefits and examples
 
 
Please send your abstracts in English to the address listed below.
A cover page should accompany the abstract and should contain the
following information:
 
               author's full name
               company/affiliation,
               address,
               email information ,
               telephone number,
               fax numbers
               intended audience:
                       OpenView novice user ___ ,  advanced user 
_____
 
               OpenView Platform applicable:
                       UNIX DM ____, UNIX SNMP _____, WINDOWS ____
 
Abstracts for presentation papers should be received by
Monday, Jan. 16, 1995 at the address shown below.  ( Final papers
will be due on March 31, 1995.  )
 
Send papers and abstracts to :
 
 OpenView Forum
 1995 Program Committee
 Attn: Sandra W. Potter
 P.O. Box 77046
 San Francisco  CA  94107

email amotive@mcimail.com
 
 
Suggestions for Tutorials, Panel Discussions, Birds-of-a-Feather 
Sessions, and
additional Conference Topics should be submitted to the 1995 program 
Committee
Chair.
 
If you have any questions please feel free to contact me at:
 
1995 Program Committee Chair
Sandra W. Potter
Air Products and Chemicals, Inc.
7201 Hamilton Blvd.
Allentown, PA 18195-1501
 
Tel: 610-481-5598
Fax: 610-481-3300
Internet: pottersw@ttown.apci.com
 
 
 
 
 
All submissions will be carefully reviewed by our Program Committee 
and
returned to the author(s) with comments to incorporate any suggested
revisions. The authors of accepted papers will receive the suggested
modifications made by the reviewer(s) for inclusion in the widely 
distributed,
Conference Proceedings.
 
The final camera-ready copy should be no longer than 12 single-spaced 
pages.
Final papers arriving too late will be will not be published, and may 
be
removed from the conference presentation. The authors of accepted 
papers
must guarantee that their paper will be presented at the conference.
 
A limited number of stipends are available only to College students 
unable to
obtain funding to attend the conference. Students whose papers are 
accepted
and who will present the paper themselves are encouraged to apply if 
such
assistance is needed. Requests for stipends should be addressed to:
 
OpenView Forum
Frank Belland,
VP of Technical Operations
313 Green Reed Road
De Bary Fla 32713
 
Phone: (407) 826-7299
Fax:   (407) 826-1530
Email: FBELLAND@DMEWRK1.ORL.MMC.COM

 
PLEASE SEND COMPLETE PAPERS TO:


 OpenView Forum
 1995 Program Committee
 Attn: Sandra W. Potter
 P.O. Box 77046
 San Francisco  CA  94107 

 email : amotive@mcimail.com

 
Papers should be submitted as soon as possible,
for inclusion in our conference schedule.
 
Proposals accepted starting:          December 1, 1994
Deadline for Receipt of Papers:       January 16, 1995
 
Notification of Acceptance Mailed:    Febuary  1, 1995
Final Camera Ready Papers Due:        March   31, 1995
-----------------------------------------------------
 
 
 
Deadline for Receipt of Proposals for Tutorials,
               Feb  1, 1995
Panel Discussions, Birds-of-a-Feather Sessions,
               March  1, 1995
------------------------------------------------------
 
ORGANIZING COMMITTEE:
 
 
Complete/Check items below:
 
 [ ]  I am interested in attending.
 
 [ ]  I intend to submit a paper.
      Provisional Title: ____________________________________________
 
 [ ]  I intend to submit a poster.
      Provisional Title: ____________________________________________
 
 [ ]  I plan to submit a tutorial proposal.
      Provisional Title: ____________________________________________
 
 [ ]  My company is interested in participating in the Vendor 
Program.
 
Please send me more information regarding:
 
 [ ]  Technical Program
 
 [ ]  Tutorial Program
 
 [ ]  Vendor Program
 
 [ ]  Accommodation
 
 [ ]  Other (Please 
Specify)__________________________________________
 
 [ ]  Membership in the OpenView Forum
 
 
 [ ]  Volunteer activities for conferences
 
Name:
 
Address:
 
City:
 
Post/State:
 
Zip Code:
 
Country:
 
Phone:
 
Fax:
 
Email Internet Addr:
 
_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_
/
 


Sandra W. Potter
Air Products and Chemicals, Inc.
7201 Hamilton Blvd.
Allentown, PA 18195-1501

Tel: 610-481-5598
Fax: 610-481-3300
Internet: pottersw@ttown.apci.com






-----------[000121][next][prev][last][first]----------------------------------------------------
Date:      Tue, 6 Dec 1994 15:10:51
From:      langen@pols.ucl.ac.be (LANGEN Gabriel)
To:        comp.protocols.tcp-ip
Subject:   Packet Driver - PCMCIA Ethernet - IBM Thinpad


I'm looking for a packet driver for a IBM PCMCIA Ethernet adapter for
the IBM Thinpad.  

Thanks,

LANGEN G.

-----------[000122][next][prev][last][first]----------------------------------------------------
Date:      Tue, 6 Dec 1994 10:36:01 GMT
From:      iialan@iifeak.swan.ac.uk (Alan Cox)
To:        comp.unix.solaris,comp.sys.sun.admin,comp.protocols.tcp-ip
Subject:   Re: Under SunOS 5.3, where to set broadcast address and netmask ??

In article <CROTEN.94Dec2155624@eosqa.gsfc.nasa.gov> croten@eosqa.gsfc.nasa.gov (Charles Roten) writes:
>Under 4.1.3, for a Sun 4/260, I put the line 
>    ifconfig ie0 128.183.117.34 netmask 255.255.255.o broadcast 128.183.117.0

Minor point but broadcast addresses end with all 1 bits not all 0's. This
was a very old BSDism that got fixed many years ago. It's a good idea to
fix it properly while you are playing with this as some systems don't let
you set .0 broadcasts and some routers don't handle them right (eg they
route them)

Alan

-- 
  ..-----------,,----------------------------,,----------------------------,,
 // Alan Cox  //  iialan@www.linux.org.uk   //  GW4PTS@GB7SWN.#45.GBR.EU  //
 ``----------'`--[Anti Kibozing Signature]-'`----------------------------''
One two three: Kibo, Lawyer, Refugee :: Green card, Compaq come read me...

-----------[000123][next][prev][last][first]----------------------------------------------------
Date:      Tue, 6 Dec 1994 10:37:39 GMT
From:      iialan@iifeak.swan.ac.uk (Alan Cox)
To:        comp.unix.solaris,comp.unix.programmer,comp.protocols.tcp-ip
Subject:   Re: Can RPC be done through FIFOs?

In article <D06u7r.C55@world.std.com> aep@world.std.com (Andrew E Page) writes:
>I'm working on a daemon/server that will only be used on it's local
>host, therefore I would like to optimize it's communication between
>processes as much as possible.  It was suggested to me that FIFOs
>would be the best way to do that.
>
>  On the other hand, I would really really like to experiment with
>using RPC calls.  However, traditionally that would require that I go
>through at least udp via sockets.  

Not really - RPC is transport independant, and on SYS5.4 cleanly so. You can
easily use Unix datagrams or pipes with a bit of care and attention. In fact
you can (and I have) used RPC over a serial line.

Alan
-- 
  ..-----------,,----------------------------,,----------------------------,,
 // Alan Cox  //  iialan@www.linux.org.uk   //  GW4PTS@GB7SWN.#45.GBR.EU  //
 ``----------'`--[Anti Kibozing Signature]-'`----------------------------''
One two three: Kibo, Lawyer, Refugee :: Green card, Compaq come read me...

-----------[000124][next][prev][last][first]----------------------------------------------------
Date:      6 Dec 1994 11:48:50 GMT
From:      root@PROBLEM_WITH_INEWS_DOMAIN_FILE (root)
To:        comp.protocols.tcp-ip
Subject:   bootpclient for MS-DOS/Windows

Hi, Guys

I am looking for a bootp client running under MS-Dos and Windows, either one 
executable which can be used by both or seperate ones.

Anybody know where I can find ?

Please respond to my e-mail address.

TIA

PETER KOOIMAN                 | Fax  : ++27-12-342-1754
Paradigm Systems Technology   |
Pretoria, South Africa        | DNS  : pko@paradigm.co.za
++27-12-342-1145              | UUCP : pko@silmaril.UUCP

-----------[000125][next][prev][last][first]----------------------------------------------------
Date:      6 Dec 1994 13:46:24 GMT
From:      prichard@axion.bt.co.uk (Paul Richardson)
To:        comp.protocols.tcp-ip,demon.ip.unix,comp.unix.solaris
Subject:   64K leased line peripherals for Sun

Hi,
Can anyone recommend a device that will allow a 64K leased line IP feed to be
connected directly to a Sun SPARCstation.  The physical connection is X21.

Any experiences doing something similar would be gratefully received.

Thanks,

Paul Richardson

-----------[000126][next][prev][last][first]----------------------------------------------------
Date:      6 Dec 94 14:31:31 GMT
From:      olafund@vax.sbu.ac.uk (olafund@vax.sbu.ac.uk)
To:        comp.protocols.tcp-ip
Subject:   PC TCP connectivity problem

Hello there..


I would appreciate some help with the following problem, as I seem to be 
getting increasingly out of my depth !!!!

I have recently taken delivery of a Versa V50 notebook to replace my old 
Dell 320 N, and have moved all relevant files to the new machine, including 
the FTP - PCTCP software. As the new machine will be used mainly with its 
docking station, I have installed a 3Com 3C509-Combo adapter in the station
, wheras I had previously used an Accton EtherPocket Adapter. I found some 
difficulty installing the adapter software, as there appeared to be no 
drivers specifically for PCTCP. Eventually, after consulting 3Com direct, I 
managed a connection by installing the NDIS driver software supplied by 3Com
, and setting this up as per the PCTCP manual.

The relevant settings from my set-up files are as follows:

From autoexec.bat:

c:\pctcp\netbind
set ftp_config=c:\pctcp\ftp\ftp.cfg
set PCTCP=c:\pctcp\ftp\ftp.cfg
rem C:\PCTCP\PKTDRV\PEPKTDRV 0X60
C:\PCTCP\ETHDRV

From config.sys:

device=c:\pctcp\protman.sys
device=c:\pctcp\elnk3.dos
device=c:\pctcp\dis_pkt.gup
DEVICEHIGH = C:\PCTCP\IPCUST.SYS
DEVICEHIGH = C:\PCTCP\IFCUST.SYS

From \lanman\protocol.ini:

[elnk]
    DRIVERNAME = ELNK3$
;   DRIVERNAME = ELNK3$
[PKTDRV]
    DRIVERNAME = PKTDRV
    BINDINGS = elnk
    INTVEC = 0x65

My problem is that some of my comms. applications now work satisfactorily, 
whilst others do not work at all. These can be summarised as follows:

i)  The 'tn' command now results in the PC locking completely - it can only 
be freed by re-booting. The 'tnglass' command, however, works as normal.

ii) ST320 works in windows, although it no longer seems capable of 
supporting multiple sessions simultaneously (it reports a conflict over 
'com1', although I am not using that port). In DOS, however, ST320 fails to 
make a connection.

iii) KEA TM! (a Windows-only product), works perfectly, allowing up to 4 
simultaneous sessions as expected.

As I am expected to support users on all 3 applications, I need to resolve 
these problems fairly urgently. Unfortunately, after the best part of two 
days spent with my head in a variety of manuals, I seem no nearer to a 
solution.

Any suggestions would be greatly appreciated.

Regards

Paul

PLEASE REPLY TO : olafund@sbu.ac.uk

+----------------------------------+-------------------------------------+
| Dotun Olaf.  [*dot*]	           | Information systems Engineering Dpt.|
|                                  |        South Bank University        |
|netmail   -> olafund@sbu.ac.uk    |_____________.London.________________|
|voicemail -> +44 0181-*BARRED*(ho)| Remploy Limited, Criclewood NW2     |
|          -> +44 0181-235-0506(of)|              London                 |
 +----------------------------------+-------------------------------------+

-----------[000127][next][prev][last][first]----------------------------------------------------
Date:      6 Dec 1994 14:33:00 GMT
From:      pww@entrust.com (Peter Whittaker)
To:        comp.unix.misc,comp.sys.hp.misc,comp.protocols.tcp-ip
Subject:   Summary: How to detect dead socket under HP-UX 9.03 (and more generally)?

I received a small number of replies to my original post, about half of
which pointed out why the original method - the one I was trying to
replace - would not work, the rest of which used various combinations of
recv() and polling - some using asynchronous sockets - to accomplish the
same thing.

In the end, I used the method I proposed below, and found it to work
very well in practice.

Why this works:  if the select call returns with a non-zero return code,
either a socket is bad, or there is activity on a socket, so check the
sockets of interest:  if the ioctl() fails, then assume the socket is
bad (a more than fair assumption:->); if the ioctl() succeeds and
indicates that there is nothing to read (numbytes==0), assume that the
activity that caused select() to set the bit for the socket was the
socket's untimely death.  Otherwise, the ioctl() succeeds and indicates
that there is data on the socket - numbytes > 0 - so invoke the routine
that handles client requests.

Note that this method depends on the use of blocking sockets.

In article <3bgvvj$doc@bcarh8ab.bnr.ca>, Peter Whittaker <pww@bnr.ca> wrote:
>I have an application that acts as a proxy for clients apps, submitting
>their requests to a server they cannot reach.  The proxy accepts a
>client connection, reads the client's request, submits it to the server,
>then uses select to wait for activity from either the client or the
>server ("select(0, &readfds, 0, 0, &timeout)").
>
>I thought of doing the following:
>
>    fd_set(s, &readfds);
>    select(0, &readfds, 0, 0, &timeout);
>    if (fd_isset(s, &readfds) {
>	ioctl(s, FIONREAD, &numbytes);
>	if (numbytes == 0) {
>	    /* there are no bytes to read, yet the "s" bit is set */
>	    exit;
>	} else {
>	    /* there really is stuff to read on s, so it must be good */
>	    recv(s, ....
>	}
>    }
>
>My reasoning is that if select() sets the bit corresponding to "s", then
>either 1) there is data to read on "s", or 2) "s" is no longer useful.
>If the ioctl(FIONREAD) shows that there is no data to read on "s",
>assume the latter (bad s) and exit.


-- 
Peter Whittaker      [~~~~~~~~~~~~~~~~~~~~~~~~~~]   NT Secure Networks
pww@entrust.com      [                          ]   P.O. Box 3511, Station C
Ph: +1 613 765 2064  [                          ]   Ottawa, Ontario, Canada
FAX:+1 613 765 3520  [__________________________]   K1Y 4H7

-----------[000128][next][prev][last][first]----------------------------------------------------
Date:      6 Dec 1994 14:42:03 GMT
From:      sd03@roger.gte.com (Shuang Deng)
To:        comp.dcom.lans.ethernet,comp.protocols.tcp-ip
Subject:   Ref. on Intel Streaming Multicast?

Intel Streaming (tm) Multicast was mentioned in Intel's
marketing literature as a multicast technique at Ethernet
mac layer that was developed by Intel. But it gives little 
technical details.  I am wondering what it is, and where 
I can find some techical descriptions of it.  Any pointer 
will be greatly apprecaited.


-Shuang Deng (sdeng@gte.com)






-----------[000129][next][prev][last][first]----------------------------------------------------
Date:      Tue, 6 Dec 1994 14:42:16 GMT
From:      aep@world.std.com (Andrew E Page)
To:        comp.unix.solaris,comp.unix.programmer,comp.protocols.tcp-ip
Subject:   Choosing RPC


  Thanks to all who mailed me some excellent perspectives on using RPC
and fifos.  I've decided to go ahead and start building the
applicaiton using RPC.  At first I will use whatever interface it
defaults to.  Someone pointed out to me that most OS's will optimize
for the loopback address anyway.  Once the interfaces are working well
enough I may fish around and see if I can put that to the test.  If
I come up with any startling optimizations I'll post them.  Thanks 
again.


-- 
Andrew E. Page   (Warrior Poet) |   Decision and Effort The Archer and Arrow
Mac Consultant                  |     The difference between what we are
Macintosh and DSP Technology    |           and what we want to be.

-----------[000130][next][prev][last][first]----------------------------------------------------
Date:      6 Dec 1994 16:51:07 GMT
From:      joe@eln.utovrm.it (Giovanni Schiavon)
To:        comp.protocols.tcp-ip,comp.sys.hp.apps,comp.sys.hp.misc
Subject:   Re: POP3 server for HPUX

Try at: ftp.csc.liv.ac.uk (138.253.42.172).
You'll find popper-1.831beta.tar.gz, under /hpux9/Networking.
I didn't try it, but all other public domain software found there was ok.
Note that you can find almost everything there, already customized for HP-PA
architecture.
There are also some mirror sites:
		hpux.cict.fr			192.70.79.53
		hpux.ask.uni-karlsruhe.de	129.13.200.57
		hpux.ced.tudelft.nl		130.161.140.100
		ftp.cae.wisc.edu		144.92.4.62
They also offer WWW access:
		http://hpux.csc.liv.ac.uk/
		http://hpux.cict.fr/
		http://hpux.ask.uni-karlsruhe.de/
		http://www.cae.wisc.edu/

-----------[000131][next][prev][last][first]----------------------------------------------------
Date:      Tue, 6 Dec 1994 18:03:30 GMT
From:      bruce@holmes.bed.ns.doe.ca (Bruce Macdonald)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   NFS/SLIP

Okay - tell me I'm crazy - we're trying to use NFS over SLIP...

It's slow, very slow, excruciatingly slow - it's much slower than
say FTP, httpd or gopher for transferring files - should I expect
otherwise?  What is it about the NFS protocol that would cause such
degradation in performance?


Bruce Macdonald
bruce@ns.doe.ca

-----------[000132][next][prev][last][first]----------------------------------------------------
Date:      Tue, 6 Dec 1994 18:13:28 GMT
From:      evansmp@mb5198.aston.ac.uk (Mark Evans)
To:        comp.unix.solaris,comp.sys.sun.admin,comp.protocols.tcp-ip
Subject:   Re: Under SunOS 5.3, where to set broadcast address and netmask ??

Alan Cox (iialan@iifeak.swan.ac.uk) wrote:
: In article <CROTEN.94Dec2155624@eosqa.gsfc.nasa.gov> croten@eosqa.gsfc.nasa.gov (Charles Roten) writes:
: >Under 4.1.3, for a Sun 4/260, I put the line 
: >    ifconfig ie0 128.183.117.34 netmask 255.255.255.o broadcast 128.183.117.0
 
: Minor point but broadcast addresses end with all 1 bits not all 0's. This
: was a very old BSDism that got fixed many years ago. It's a good idea to
: fix it properly while you are playing with this as some systems don't let
: you set .0 broadcasts and some routers don't handle them right (eg they
: route them)

While you are at it you might also like to check your routing of
127.X.X.X quite a few configurations have been know to send any of these
addresses, except 127.0.0.1, via the default route.

-----------[000133][next][prev][last][first]----------------------------------------------------
Date:      6 Dec 1994 18:57:35 GMT
From:      colin@mayfield.hp.com (Colin Wynd)
To:        comp.protocols.tcp-ip
Subject:   Re: TCP/IP to SNA

Mr Andrew Lee (alee@bootsgt1.demon.co.uk) wrote:
: I am interested in products that allow terminals(PC's) on a TCP/IP
: network access to a SNA Mainframe without the need for TCP/IP 
: software on the Mainframe.

Try openconnect systems. They have a TCP<->SNA gateway and what
looks like a bunch of software to allow you to connect.

The only number i have for them is:

Open Connect Systems:

   PHONE:       (USA)-214-888-0433
   UK (Croydon)       081-781-6978
                 FAX  081-781-6975
   Contact Vic Baldorino


They have an email for their newsletter: newsltr-l@oc.com

I've never used their product!

Cheers
Colin
-//-



-----------[000134][next][prev][last][first]----------------------------------------------------
Date:      6 Dec 1994 20:03:31 GMT
From:      Marc Koenig <koenig@mirlink.wustl.edu>
To:        comp.protocols.tcp-ip
Subject:   Winsock connection problem. Help please.

    I recently obtained a SLIP account and am having a difficult 
time obtaining a consistant connection. I am using Winsock.dll 
2.0B with Netscape 94 with a new USRobotics 28.8 V.FC V.34 
internal modem.  I still have not figured out how to login with 
a script so I have been doing it manually. Here is what occurs:
     I open Winsock, choose Manual Login from the pull down menu, 
type AT&F1 (I have tried leaving this out), it says, "OK", I then 
type ATDT followed by my dial-up number (only one number has been provided for me). It always dials and connects. It will usually connect and give me my connection speed. At that point I hit return twice and it is suppose to give me sequential prompts for User ID 
and password.  Usually, when it connects and says 26.4 speed, it 
will respond and let me continue on as discussed above. However, 
the majority of the time it will connect and state, "24000/ARQ/V34/LAPM/V42BIS" and when that happens it will not 
respond at all when I hit Enter twice or try anything else. E.G., 
two nights ago I attempted to dial in about about 40 times, all but 
2 times it connected at 24000 (a few times lower) and I had to 
abort because it would not respond. Even the 2 times it connected 
at 26,400 it let me go on through the whole sequence and start 
SLIP but then Netscape was unable to make any connections (but that 
is problably a different problem since previously, it usually works fine if I connect at 26400). Last night it connected at 24000 every time so I was again never able to access the net.   Does anyone 
have any suggestions? I'm getting very frustrated. I would suspest 
the new modem except for the fact that I was connecting fairly consistantly previously before. I was wondering if it had to do 
with my IP address. When I do succeed at a full connection, it 
tells me my IP address. I notice that the last digit in the address frequently varies, i.e. I am being assigned an address on the fly, 
I guess. I was entering a specific IP address, the one I noted I 
first connected at. With this present problem, I tried changing the 
IP Address choice in the setup page to 0.0.0.0 but it is still 
doing the same thing. Any ideas? How is the best way to set up 
Winsock when the IP address is dynamically assigned? I really have 
no idea if the address issue has anything to do with the connection problem but I thought I would mention it.
    I know that the best response is always, "ask your internet provider". However, I have been given a account which is new for 
my institution (PC in the midst of a MAC world) and the support 
staff is just not able to help me any more on problems that come 
up with the setup. Therefor, I am counting on you folks. Any help would be very much appreciated. Thank you very much. 
          
 Marc 

P.S.  I am also trying to work out some E-mail issues, if my 
Internet address, Koenig@mirlink.wustl.edu, does not work, try 72672.712@compuserve.com.


-----------[000135][next][prev][last][first]----------------------------------------------------
Date:      Tue, 6 Dec 1994 20:11:48 GMT
From:      ko@komac.knoware.nl (kS)
To:        comp.protocols.tcp-ip,comp.unix.aux
Subject:   [Q] unix slip client and dynamic addressing?

I want to connect my UNIX system to an internet provider that assigns me
an IP address dynamicaly (each time I connect and login it is different)

Normally I have another internet provider and a fixed internet address;

what I do then is;

%: upsl0 -b38400 -S /etc/slip.dial.knoware komac beasty

slattach: sl0
 at 38400 baud, remote beasty (193.78.120.5) local komac (193.78.122.105)
exec: /usr/etc/route add default beasty 2
add net default: gateway beasty:

komac and beasty are in /etc/hosts

But what to do with an internetprovider that assigns me a different
address each time I login?

slattach ?
slattconf ?
ifconfig ?
route ?

I run A/UX

-- 
V2S
Holland

-----------[000136][next][prev][last][first]----------------------------------------------------
Date:      6 Dec 1994 20:14:34 GMT
From:      somani@sun2.cs.wisc.edu (Amit Somani)
To:        comp.protocols.tcp-ip
Subject:   Available buffer space in a UDP socket ??

Hi, 

Is there any way to find out the amount of available (free) space
associated w/ a UDP socket ? The UDP socket has intitialized its 
buffer space by a call to setsockopt().

I'd apreciate it if you could e-mail me at somani@cs.wisc.edu. 

Thanks,
--Amit

-----------[000137][next][prev][last][first]----------------------------------------------------
Date:      Tue, 6 Dec 1994 20:27:06 GMT
From:      jacobsen@itex.jct.ac.il (Joel Jacobsen)
To:        comp.protocols.tcp-ip
Subject:   TCP/IP for msdos?

Is there any good free or even Shareware TCP/IP package for msdos?

    Joel
--
------------------------------------------------------------------------
Joel Jacobsen       | INTERNET: jacobsen@itex.jct.ac.il
P.O.B 953           | --------------------------------------------------
Kiryat Arba 90100   | "la la la la la "  - (The Mamas & The Papas)

-----------[000138][next][prev][last][first]----------------------------------------------------
Date:      7 Dec 1994 12:20:35 -0800
From:      wolfgang@wsrcc.com (Wolfgang Rupprecht)
To:        comp.protocols.tcp-ip
Subject:   Re: Fkey characters getting split on frame relay

bill@twg.bc.ca (Bill Irwin) writes:
>I have a situation where the characters being sent by a
>terminal's function keys are being split into separate packets as
>they travel through the network on their way to the host
>application.  This usually results in an error "beep", with one
>or more straggling characters being displayed at the cursor
>position.

Delivery timing is not guarenteed.  It would be incorrect for your app
to assume function key escape sequences will arrive in one packet.
Why not have your app parse the key sequence the same way no matter if
it arrived in 50ms or 50 seconds?

-wolfgang
-- 
Wolfgang Rupprecht <wolfgang@wsrcc.com>  <http://www.wsrcc.com/>
News Flash:  Hayes files chapter 11    +++ATH0

-----------[000139][next][prev][last][first]----------------------------------------------------
Date:      Tue, 6 Dec 1994 23:22:47 GMT
From:      bpeoples@iglou.iglou.com (Ben Peoples)
To:        comp.protocols.tcp-ip
Subject:   Re: PPP via telnet...

Ok, I've got conflicting reports.  One person said that version 2.1.2 can be
used via telnet, and that they use it nearly everyday, another says that it
cannot be done.  I'll trust the first :)
: What do you want to achieve by doing that?
The reason for this, is that I can get a free PPP through one site, but not my 
dial-in site.  

				 		Ben

-- 
Ben Peoples					       	bpeoples@iglou.com
Unless otherwise stated, the above opinions are ***WORDS***
"As I said, with this net I will catch them, I bet.  With this net I will catch
those things, yet!"  --- _The Cat in The Hat_

-----------[000140][next][prev][last][first]----------------------------------------------------
Date:      7 Dec 1994 09:11:08 -0500
From:      rjm@panix.com (rjm)
To:        comp.protocols.tcp-ip
Subject:   Re: dynamic routing & nfs

Tony Li (tli@cisco.com) wrote:
: In article <3c2cs2$8hp@panix2.panix.com> rjm@panix.com (rjm) writes:
:      I have a c/s application that requires static routing between the 
:      server and its clients. between the server and the clients is router A
:      cisco configured dynamically dual T1 lines , then router B then the
:      clients. 
:      
:     ex: 
:     
:       server<------>router A<----> T1<---->router B<--->clients
:     
:       the telcom (network owners) are telling me that these twp routers
:     are ciscos configured to be dynamic routers. Is there any way that a 
:     static route can be configured along side the dynamic route?
:     or will I have to get them to assign me s compleatly independent
:     route with my own routers?  Env: AIX 3.2.5 ethernet applicationmakes
:     extensive use of NFS.
:     
: Yes, you can have dynamic routing active and configure static routing at
: the same time.  However, why would you want to?  I can't believe that the
: application can tell the difference between a static route and dynamic
: routing.
 
: Tony
 
  rjm replies:

  My vendor is telling me that nfs keeps timing out across the routers
  (70 blocks) and that there is too much broadcast traffic across the
  network (non-ip stuff). When we do a netstat we get indications that
  the routers were are assigned to are not giving us a direct route between
  them. we have to cross about ten subnets to get to the router we want t
  to get to. They are also telling me that the routers can be configured
  to filter out non ip traffic but I don't see how that can help me
  if non ip users need to use the same routers.


   
  i

-----------[000141][next][prev][last][first]----------------------------------------------------
Date:      Wed, 7 Dec 1994 01:18:56 GMT
From:      vjs@calcite.rhyolite.com (Vernon Schryver)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

In article <D0EI5u.B2r@cid.aes.doe.ca> bruce@holmes.bed.ns.doe.ca (Bruce Macdonald) writes:
>Okay - tell me I'm crazy - we're trying to use NFS over SLIP...
>
>It's slow, very slow, excruciatingly slow - it's much slower than
>say FTP, httpd or gopher for transferring files - should I expect
>otherwise?  What is it about the NFS protocol that would cause such
>degradation in performance?

In effect, NFS has very little windowing.  Except the "read-ahead" or
"write-behind" typical in a UNIX file system, data is moved purely
"request-response."  The NFS client makes one request, waits for the
response, and then makes another request.  While the request is going
from the client to the server, the wire is idle in the other direction,
which clobbers performance.  If you use a big MTU (at least 1006) then
the cost of the IP headers on the fragments is not bad and the lack of
header compression is insignificant.  I'd expect NFS to be about 60% as
fast as FTP.

Some NFS clients have very aggressive read-ahead mechanisms.  That helps
some purely sequential read benchmarks over Ethernet and FDDI, but hurts
random reads on LANs.  Aggressive read-ahead would help pure NFS/SLIP
file transfers, but I shudder to think of the consequences if you are
doing anything random.  The trouble is that reading-ahead is a guess
that the next application request will be for the next block.  That is
good when the guess is right, but wasting 3 or 4 seconds of bandwidth
when the guess is wrong is painful.

I now do a lot of NFS over PPP.  I once did a lot over SLIP.  If you
just want to transfer files, if you do not have some other good reason
to use NFS, then stick to FTP or rcp.


Vernon Schryver    vjs@rhyolite.com

-----------[000142][next][prev][last][first]----------------------------------------------------
Date:      Wed, 7 Dec 1994 02:00:08 GMT
From:      nac@sirius.com (Nancy Cedeno)
To:        comp.protocols.tcp-ip
Subject:   Re: PPP via telnet...

In article <3buq21$mop@nyheter.chalmers.se> thed@mtek.chalmers.se (Thed Lyngmo) writes:

>: Is it possible to run a ppp (or slip for that matter) connection over a 
>: telnet connection?  
 
>I only know how SLIP works and I can't see why that shouldn't be possible.
>I guess the same goes for PPP.
 
>What do you want to achieve by doing that?

I think it would be quite useful if you're stuck behind a firewall, and the 
only way you can get out is via a telnet session to a host server.  

That's the setup we have at work right now.  The only machine you can ftp from 
is that firewall server.  If we could run PPP from our MS Windows machines to 
that host, we could use something like ws_ftp and wouldn't need to deal with 
the middleman.


--
Nancy Cedeno
Writer, Netsurfer           "I'm not procrastinating -- I'm working!"
nac@well.sf.ca.us
nac@sirius.com

-----------[000143][next][prev][last][first]----------------------------------------------------
Date:      Wed, 7 Dec 1994 02:04:05 GMT
From:      ljdejan@pb2esac.com (Leo Dejan)
To:        comp.os.linux.help,comp.protocols.tcp-ip,comp.dcom.lans.ethernet
Subject:   Unable to ifconfig 3C509 card UP. Help!


I am trying to set up two machines on a personnal network:

Machine One: 486 DX2 50 EISA, 16M ram, 1G HD, Linux 1.1.59
			 with 3Com 3C509 ISA  ethernet card.
  CARD IS CONFIGURED AND UP!

Machine Two: Pentium 90 PCI, 16M ram, 2G HD, Linux 1.1.59
			 with 3Com 3C509 ISA ethernet card.
   CARD IS CONFIGURED BUT NOT UP!

I've swapped the two cards with same results. When I use ifconfig
to configure the IP address I get 'SIOCSIFFLAGS: Try again'. The
address gets configured. I try again with the same results. I've
rebuilt the kernel twice. Read volumes of docs!

What have I missed? Please help!

Leo

-----------[000144][next][prev][last][first]----------------------------------------------------
Date:      7 Dec 1994 14:07:38 -0700
From:      pete@dswi.com (Pete Kruckenberg)
To:        comp.os.linux.development,comp.os.linux.admin,comp.os.linux.misc,comp.protocols.tcp-ip,comp.protocols.tcp-ip.domains
Subject:   Multi-homed (multiple IP addresses) Linux hosts?

I saw a reference in another newsgroup to the capabilities of some new
*nix OS's that allow a host to have multiple IP addresses on a single
network interface (using an "ifconfig alias" command). 

Does Linux have this capability, or anything similar to it? What other
OS's (free or commercial) support this capability? What do I need to
get/have to have this capability?

Thanks.
Pete Kruckenberg
pete@dswi.com


-----------[000145][next][prev][last][first]----------------------------------------------------
Date:      7 Dec 1994 04:02:08 GMT
From:      statham@bga.com (Perry Statham)
To:        comp.protocols.tcp-ip
Subject:   Re: TCP/IP protocol stack needed

In article <bohac.1.0@feld.cvut.cz>, bohac@feld.cvut.cz (Leos Bohac Ing. K332) says:
>        I am looking for TCP/IP stack source in C language to experiment
>with the network using packet driver. Does anybody knows a shareware
>source where I can retrieve that.       

Check out the WatTCP package on dorm.rutgers.edu.

Perry Statham
perry@crs.com

-----------[000146][next][prev][last][first]----------------------------------------------------
Date:      Wed, 7 Dec 1994 04:48:55 GMT
From:      robertsr@helios.usq.EDU.AU (roger roberts)
To:        comp.protocols.tcp-ip
Subject:   Re: Lanwatch Diags Software

fks@ftp.com  (Frances K. Selkirk) writes:


>LANWatch is available with several different driver interfaces. It is
>even still available with card-specific drivers for a few network
>interface cards, but now most people do get it the portable version,
>which can run over a packet driver, NDIS driver, or ODI driver.
>However, in 4.0, support for each driver type is direct, rather than
>through shims, so error counts are not limited to those known to the
>packet driver specification. I believe NDIS drivers keep all of these
>error counts; I'm not sure about ODI drivers. You are correct that
>packet drivers do not.

I'm thinking of purchasing Lanwatch V4.0 and I'd like to be able to observe
 error counts caused by or from a selected workstation.I would like to 
use an NDIS driver .Could you please advise if the NDIS driver reports all
errors as accurately  as the card specific versions of Lanwatch.
 This is a request for information only.                                     
        
Thank-you for your assistance.
Roger Roberts
University of Southern Queensland
Australia      mail : robertsr@usq.edu.au




-----------[000147][next][prev][last][first]----------------------------------------------------
Date:      7 Dec 1994 13:59:23 -0500
From:      walt@phoenix.oki.com (Walt Weber)
To:        comp.protocols.tcp-ip
Subject:   Re: help me, please w/SLIP (Q)

John.Jamerson@launchpad.unc.edu (John Jamerson) writes:
>
>PC (stand alone or WFWG - same symptoms) running Trumpet Winsock 2.0b
>dials server (Livingston Port Master hooked to UNIX box running
>SunOS 4.1.3.  After login and password, SLIP enables (Portmaster does
>this), I enter correct IP addr in Trumpet, minimize trumpet, then
>attempt to use a client.  Ping works OK with normal reports, but
>nothing else I attempt does (Mosaic, Cello, telnet, ftp, etc.).

Is this ping by IP address , or ping by hostname?
Is this a ping of the default gateway (the Portmaster?) or ping of
a machine beyond the gateway?

Reasoning:

ping of the gateway by number checks the SLIP link, nothing more.
good first step.

ping of a machine beyond the gateway by IP address checks the routing
tables - success indicates that you've set Trumpet up to know (and use)
the correct default gateway, and the Portmaster is correctly propagating
the route back to your dialup link to other machines on the net.

ping of a machine by name checks that name to address lookup is working.
Trumpet uses a HOSTS file, as well as supporting DNS.

>e-mail preferred, but post if ya gotta.

Sorry, my worldview is "posted questions deserve posted answers", unless
the list/newsgroup is screaming stop! stop!

...walt...
-- 
--
Walt Weber                         OKI Advanced Products Division
walt@oapd.oki.com                  Marlboro, MA 01752 USA
508-624-7000 x8637                 FAX: 508-480-9635

-----------[000148][next][prev][last][first]----------------------------------------------------
Date:      Wed, 7 Dec 1994 06:24:57 GMT
From:      aboba@netcom.com (Bernard Aboba)
To:        comp.protocols.tcp-ip
Subject:   Re: PC as nameserver?

The DIS version of KA9Q (Large model) can act as a nameserver. However, 
be aware that it will only answer queries for locally stored records. It
will not recurse if requested to do so; rather than querying the root servers
for a record it doesn't have, it will return an error. 

Moral: point your resolvers somewhere else; use KA9Q to  answer queries from
the outside world. 

This is assuming that you don't have even  a 386/16 with 4 MB RAM; KA9Q will
run in 1 MB, even on a 286. If you have a 386/16 wtih 4 MB, you can run
named under Linux or FreeBSD instead. 




-----------[000149][next][prev][last][first]----------------------------------------------------
Date:      7 Dec 1994 20:17:13 -0800
From:      cgi@crl.com (Paul Smith)
To:        comp.protocols.tcp-ip
Subject:   Re: Address already in use message

Patrick Horgan (Patrick.Horgan@holt-systems.com) wrote:
: From: patrick@vega.oes.amdahl.com (Patrick Horgan)
: Subject: Re: Address already in use message
: Organization: Amdahl Corporation
 [snip]
: 1) Linger time non-zero, we want the close() to wait until all pending data is
:    sent and acknowledged, or until the linger time set in the linger struct
:    has passed.  (Most current implementations don't actually use the time,
:    they just note whether it's non-zero despite what I just said, and the
:    comment below says.)

Does setting linger to !=0, invite a potential indefinate block on a close()??

Or does it time out?  preferably in the time frame I set in the linger struct..

...


-----------[000150][next][prev][last][first]----------------------------------------------------
Date:      7 Dec 1994 19:22:09 -0700
From:      pete@dswi.com (Pete Kruckenberg)
To:        comp.os.linux.development,comp.os.linux.admin,comp.os.linux.misc,comp.protocols.tcp-ip,comp.protocols.tcp-ip.domains
Subject:   Re: Multi-homed (multiple IP addresses) Linux hosts?

Pete Kruckenberg (pete@dswi.com) wrote:
: I saw a reference in another newsgroup to the capabilities of some new
: *nix OS's that allow a host to have multiple IP addresses on a single
: network interface (using an "ifconfig alias" command). 
 
: Does Linux have this capability, or anything similar to it? What other
: OS's (free or commercial) support this capability? What do I need to
: get/have to have this capability?

Someone mailed me saying that any BSD4.4-based system will have this
capability using "ifconfig ... alias", and I read a post in another
newsgroup which said that BSD and Linux have this capability. So, I'm
assuming I can do this, but I just need to know how. Can anyone help?

Pete.



-----------[000151][next][prev][last][first]----------------------------------------------------
Date:      Wed, 7 Dec 1994 09:31:15 GMT
From:      gdmr@dcs.ed.ac.uk (George Ross)
To:        biz.sco.general,comp.protocols.tcp-ip,wimsey.general
Subject:   Re: Fkey characters getting split on frame relay

In article <D0BCMr.FG3@twg.bc.ca>, bill@twg.bc.ca (Bill Irwin) writes:
> Although modifying each application to wait longer for Fkey
> completion characters may be an option, it seems to me that the
> logical place to do this would be at the OS level, either in the
> TCP/IP software or the termcap/info databases.  In this way, the
> configuration is done once for each type of terminal being
> supported, rather than for each application.

Perhaps turning down the notional baud-rate on the pty would help.  But in any
case it's inherent in TCP's design that it's a bytestream protocol, not a
record-stream, and that the user's data can be split into chunks for
transmission in whatever way it feels like at the time.
-- 
George D M Ross, Department of Computer Science, University of Edinburgh
     Kings Buildings, Mayfield Road, Edinburgh, Scotland, EH9 3JZ
Mail: gdmr@dcs.ed.ac.uk   Voice: +44 131 650 5147   Fax: +44 131 667 7209

-----------[000152][next][prev][last][first]----------------------------------------------------
Date:      7 Dec 1994 09:45:56 GMT
From:      tli@cisco.com (Tony Li)
To:        comp.protocols.tcp-ip
Subject:   Re: dynamic routing & nfs

In article <3c2cs2$8hp@panix2.panix.com> rjm@panix.com (rjm) writes:
     I have a c/s application that requires static routing between the 
     server and its clients. between the server and the clients is router A
     cisco configured dynamically dual T1 lines , then router B then the
     clients. 
     
    ex: 
    
      server<------>router A<----> T1<---->router B<--->clients
    
      the telcom (network owners) are telling me that these twp routers
    are ciscos configured to be dynamic routers. Is there any way that a 
    static route can be configured along side the dynamic route?
    or will I have to get them to assign me s compleatly independent
    route with my own routers?  Env: AIX 3.2.5 ethernet applicationmakes
    extensive use of NFS.
    
Yes, you can have dynamic routing active and configure static routing at
the same time.  However, why would you want to?  I can't believe that the
application can tell the difference between a static route and dynamic
routing.

Tony


-----------[000153][next][prev][last][first]----------------------------------------------------
Date:      Wed, 7 Dec 1994 10:52:49 GMT
From:      ko@komac.knoware.nl (kS)
To:        comp.protocols.tcp-ip,vpro.technisch,comp.unix.aux
Subject:   [Q] Slip client under UNIX and dynamic addressing?

What to do with 

slattconfig
slattach
ifconfig
route

If I dial into a Slip server that does dynamic addressing
(dynamic addressing meaning my system gets assigned a different IP address
from the server each time I have a slip-connect )

-- 
V2S
Holland

-----------[000154][next][prev][last][first]----------------------------------------------------
Date:      07 Dec 1994 13:09:43 GMT
From:      jim@cs.strath.ac.uk (Jim Reid)
To:        comp.protocols.tcp-ip
Subject:   Re: POP server for SunOS?

In article <clyde.786507279@trojan> clyde@hitech.com.au (Clyde Smith-Stubbs) writes:

   Does anyone know of a POP server that will run under SunOs?

The one that comes with MH works just fine. We've been using it for years.

-----------[000155][next][prev][last][first]----------------------------------------------------
Date:      7 Dec 1994 21:32:22 -0500
From:      barmar@nic.near.net (Barry Margolin)
To:        comp.protocols.tcp-ip
Subject:   Re: dynamic routing & nfs

In article <3c4fps$t0g@panix2.panix.com> rjm@panix.com (rjm) writes:
>  My vendor is telling me that nfs keeps timing out across the routers
>  (70 blocks) and that there is too much broadcast traffic across the
>  network (non-ip stuff). When we do a netstat we get indications that
>  the routers were are assigned to are not giving us a direct route between
>  them. we have to cross about ten subnets to get to the router we want t
>  to get to.

If the dynamic routing protocol is choosing a long, slow route rather than
the short one, there's a problem with the routing protocol.  Even a simple
protocol like RIP won't choose 10 hops when there's a direct connection.
What routing protocol are you using on your network?
-- 

Barry Margolin
BBN Internet Services Corp.
barmar@near.net

-----------[000156][next][prev][last][first]----------------------------------------------------
Date:      07 Dec 1994 15:22:59 GMT
From:      nelson@crynwr.crynwr.com (Russell Nelson)
To:        comp.protocols.tcp-ip
Cc:        pko@paradigm.co.za
Subject:   Re: bootpclient for MS-DOS/Windows

In article <3c1j32$o57@hermes.is.co.za> root@PROBLEM_WITH_INEWS_DOMAIN_FILE (root) writes:

   I am looking for a bootp client running under MS-Dos and Windows, either one 
   executable which can be used by both or seperate ones.

What do you plan to do with the data that the bootp client returns?
It's not much good unless you're going to use it with a TCP/IP package...

--
-russ <nelson@crynwr.com>    http://www.crynwr.com/crynwr/nelson.html
Crynwr Software   | Crynwr Software sells packet driver support | ask4 PGP key
11 Grant St.      | +1 315 268 1925 (9201 FAX)  | What is thee doing about it?
Potsdam, NY 13676 | What part of "Congress shall make no law" eludes Congress?

-----------[000157][next][prev][last][first]----------------------------------------------------
Date:      07 Dec 1994 15:24:58 GMT
From:      nelson@crynwr.crynwr.com (Russell Nelson)
To:        comp.protocols.tcp-ip
Subject:   Re: Packet Driver - PCMCIA Ethernet - IBM Thinpad

In article <langen.10.000F2EC5@pols.ucl.ac.be> langen@pols.ucl.ac.be (LANGEN Gabriel) writes:

   I'm looking for a packet driver for a IBM PCMCIA Ethernet adapter for
   the IBM Thinpad.  

Sorry, as far as I know, you're out of luck.  But please ask IBM for
one, because that's how they know which drivers to write.  And feel
free to tell them that Crynwr Software writes packet drivers...

--
-russ <nelson@crynwr.com>    http://www.crynwr.com/crynwr/nelson.html
Crynwr Software   | Crynwr Software sells packet driver support | ask4 PGP key
11 Grant St.      | +1 315 268 1925 (9201 FAX)  | What is thee doing about it?
Potsdam, NY 13676 | What part of "Congress shall make no law" eludes Congress?

-----------[000158][next][prev][last][first]----------------------------------------------------
Date:      Wed, 7 Dec 1994 15:29:29 GMT
From:      "Chris Barrett" <chrisb@twg.com>
To:        comp.protocols.tcp-ip
Subject:   PPP Server for RS/6000

Would someone please tell me where I can find a PPP server for an IBM RS/6000? 
Commercially available is best, but I will consider anything.  Feel
free to include your opinion of the products quality.  Post replies to this
group or to me directly at chrisb@twg.com.  Thanks in advance.

Regards,

Chris Barrett
415-962-7143

-----------[000159][next][prev][last][first]----------------------------------------------------
Date:      7 Dec 1994 13:50:36 +0100
From:      acser@goliat.eik.bme.hu (Cser Andras)
To:        comp.protocols.tcp-ip
Subject:   Traceroute with -g for Solaris 2.3

Dear Guys,

Please send us information on how to get traceroute -g able
traceroute for Solaris 2.3 or send me compiled binary.

Any help would be appreciated,
Greeting

Andras Cser
Sysadm, Center of Information Systems
Technical University of Budapest
Hungary
-- 
Kill the body, the head will die.
------------------------------------------------------------------
Andras Cser				acser@goliat.eik.bme.hu
Technical University of Budapest	Tel.:(voice) +36-1-1812172

-----------[000160][next][prev][last][first]----------------------------------------------------
Date:      07 Dec 1994 15:35:13 GMT
From:      nelson@crynwr.crynwr.com (Russell Nelson)
To:        comp.sys.ibm.pc.hardware.networking,comp.dcom.lans.ethernet,comp.protocols.tcp-ip.ibmpc,comp.protocols.tcp-ip
Subject:   Re: DOS ping for packet driver.

In article <jfauerba.150.000F6FCF@ncdc.noaa.gov> jfauerba@ncdc.noaa.gov (John Fauerbach) writes:

   Anyone know where I can find a DOS base Ping for TCP/IP packet Driver?
   This is to be used on a PC without windows.

Jan Engvald's pdtstnet is absolutely the best
ping/traceroute/address-checker/clock-setter anywhere.  World's
fastest traceroute.  Does multiping so that you can check throughput
along a path for multimedia applications.

Active development:

ftp://ftp.lu.se/pub/network/pdclkset/pdtst303.zip

Stable:

ftp://oak.oakland.edu/pub/msdos/pktdrvr/pdtst217.zip

--
-russ <nelson@crynwr.com>    http://www.crynwr.com/crynwr/nelson.html
Crynwr Software   | Crynwr Software sells packet driver support | ask4 PGP key
11 Grant St.      | +1 315 268 1925 (9201 FAX)  | What is thee doing about it?
Potsdam, NY 13676 | What part of "Congress shall make no law" eludes Congress?

-----------[000161][next][prev][last][first]----------------------------------------------------
Date:      7 Dec 1994 16:05:31 GMT
From:      balenson@tis.com (David M. Balenson)
To:        alt.privacy,alt.security,alt.security.pgp,alt.security.ripem,comp.protocols.kerberos,comp.protocols.tcp-ip,comp.security.misc,comp.security.unix,ieee.announce,sci.crypt
Subject:   Program Announcement: ISOC '95 Symp. Netw. & Distr. Sys. Security

==============================================================================

                    THE INTERNET SOCIETY SYMPOSIUM ON
                 NETWORK AND DISTRIBUTED SYSTEM SECURITY

                         16-17 FEBRUARY 1995

                 CATAMARAN HOTEL - SAN DIEGO, CALIFORNIA

The symposium will bring together people who are building software
and/or hardware to provide network and distributed system security
services.  The symposium is intended for those interested in the more
practical aspects of network and distributed system security, focusing
on actual system design and implementation, rather than in theory.  We
hope to foster the exchange of technical information that will
encourage and enable the Internet community to apply, deploy and
advance the state of the available security technology.

==============================================================================

                 P R E L I M I N A R Y   P R O G R A M

WEDNESDAY, FEBRUARY 15

6:00 P.M. - 8:00 P.M.
REGISTRATION AND RECEPTION

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 

THURSDAY, FEBRUARY 16

7:30 A.M.
CONTINENTAL BREAKFAST

8:30 A.M.
OPENING REMARKS

9:00 A.M.
SESSION 1: DIVERSE APPROACHES TO SECURITY AT THE NETWORK LAYER
Chair: Stephen T. Kent (Bolt, Beranek and Newman, USA)

    Multicast-Specific Security Threats and Counter-Measures, Tony
    Ballardie and Jon Crowcroft (University College London, United
    Kingdom).

    Design of a Key Agile Cryptographic System for OC-12c Rate ATM,
    Daniel Stevenson, Nathan Hillery, Greg Byrd, and Dan Winkelstein
    (Microelectronics Center of North Carolina - MCNC, USA).

    IpAccess: An Internet Service Access System for Firewall
    Installations, Steffen Stempel (University of Karlsruhe, Germany).

10:30 A.M.
BREAK

11:00 A.M.
SESSION 2: PANEL: SECURITY ARCHITECTURE FOR THE INTERNET INFRASTRUCTURE
Chair: Robert W. Shirey (The MITRE Corporation, USA)

    Security for the Internet Protocol (IP) and IP Next Generation,
    Paul A. Lambert (Motorola, USA).

    Security for the Internet Domain Name System, James M. Galvin
    (Trusted Information Systems, USA).

    Security of Routing Protocols in the Internet, Gary Scott Malkin
    (Xylogics, USA).

    Security Approaches to Routing in the Internet, Sandra L. Murphy
    (Trusted Information Systems, USA).

12:30 P.M.
LUNCH

2:00 P.M.
SESSION 3: OFF-LINE OBJECT DISTRIBUTION SECURITY
Chair: Jeffrey I. Schiller (Massachusetts Institute of Technology, USA)

    Trusted Distribution of Software Over the Internet, Aviel D. Rubin
    (Bellcore, USA).

    Location-Independent Information Object Security, John Lowry (Bolt
    Beranek and Newman, USA).

3:00 P.M.
BREAK

3:30 P.M.
SESSION 4: INTERNET PAYMENTS
Chair: Ravi Ganesan (Bell Atlantic, USA)

    Electronic Cash on the Internet, Stefan Brands (Centrum voor
    Wiskunde en informatica - CWI, The Netherlands).

    PANEL: Internet Payment Mechanisms - Requirements and Architecture
    Chair: Ravi Ganesan (Bell Atlantic, USA)
    Panelists: B. Clifford Neuman (Information Sciences Institute, USA), 
    David Crocker (Brandenburg Consulting, USA), and others TBD

7:00 P.M.
DINNER BANQUET

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 

FRIDAY, FEBRUARY 17

7:30 A.M.
CONTINENTAL BREAKFAST

8:30 A.M.
SESSION 5: SECURITY MONITORING TOOLS - PRACTICE AND EXPERIENCE
Chair: Michael St. Johns (Advanced Research Projects Agency, USA)

    NERD: Network Event Recording Device: An Automated System for
    Network Anomaly Detection and Notification, David G. Simmons and
    Ronald Wilkins (Los Alamos National Laboratory, USA).

    An Overview of SNIF: A Tool for Surveying Network Information Flow,
    Jim Alves-Foss (University of Idaho, USA).

    Distributed Audit Trail Analysis, Abdelaziz Mounji, Baudouin Le
    Charlier, Denis Zampunieris and Naji Habra (Facultes Universitaires
    de Namur - FUNDP, Belgium).

10:00 A.M.
BREAK

10:30 A.M.
SESSION 6: AUTHENTICATION AND AUTHORIZATION
Chair: B. Clifford Neuman (Information Sciences Institute, USA)

    SESAME V2 Public Key and Authorisation Extensions to Kerberos,
    Piers McMahon (ICL, United Kingdom).

    Yaksha: Augmenting Kerberos with Public Key Cryptography, 
    Ravi Ganesan (Bell Atlantic, USA).

    GSS-API Security for ONC RPC, Barry Jaspan (OpenVision
    Technologies, USA).

12:00 NOON - 1:30 P.M.
LUNCH

1:30 P.M.
SESSION 7: MECHANISMS OF IDENTITY - THE CERTIFICATE INFRASTRUCTURE
Chair: Hilarie Orman (University of Arizona, USA)

    A Certificate Management System: Structure, Functions and
    Protocols, Nada Kapidzic and Alan Davidson (Stockholm University &
    Royal Institute of Technology, Sweden).

    PEMToolKit: Building a Top-Down Certification Hierarchy for PEM
    from the Bottom Up, Alireza Bahreman (Bellcore, USA).

    A New Approach to the X.509 Framework: Allowing a Global
    Authentication Infrastructure Without a Global Trust Model, Suzan
    Mendes (TS-E3X - Research and Development Center, France) and Christian
    Huitema (INRIA, France).

3:00 P.M.
BREAK

3:30 P.M.
SESSION 8: PANEL: SECURITY ISSUES FOR MOSAIC AND THE WORLD WIDE WEB
Chair: Fred Avolio (Trusted Information Systems, USA)
Panelists: Peter J. Churchyard (Trusted Information Systems, USA),
  Allan M. Schiffman (Enterprise Integration Technologies, USA), and
  Bill Cheswick (AT&T Bell Laboratories, USA)

------------------------------------------------------------------------------

GENERAL CHAIR

    James T. Ellis, CERT Coordination Center, Carnegie Mellon University

PROGRAM CO-CHAIRS

    David M. Balenson, Trusted Information Systems
    Robert W. Shirey, The MITRE Corporation

PROGRAM COMMITTEE

    Thomas A. Berson, Anagram Laboratories
    Matt Bishop, University of California at Davis
    Ravi Ganesan, Bell Atlantic
    Stephen T. Kent, Bolt, Beranek and Newman
    Paul A. Lambert, Motorola
    John Linn, OpenVision Technologies
    B. Clifford Neuman, Information Sciences Institute
    Hilarie Orman, University of Arizona
    Michael Roe, University of Cambridge (UK)
    Robert Rosenthal, U.S. National Institute of Standards and Technology
    Jeffrey I. Schiller, Massachusetts Institute of Technology
    Peter Yee, U.S. National Aeronautics and Space Administration
    Roberto Zamparo, Telia Research (Sweden)

PUBLICATIONS CHAIR

    Terry Mayfield, Institute for Defense Analyses

REGISTRATIONS CHAIR

    Gloria Carrier, The MITRE Corporation

LOCAL ARRANGEMENTS CHAIR

    Thomas Hutton, San Diego Supercomputer Center

STEERING GROUP

    Internet Research Task Force, Privacy and Security Research Group

------------------------------------------------------------------------------

BEAUTIFUL SAN DIEGO

The Symposium venue is the Catamaran Resort Hotel, providing 7 acres of  
gorgeous surroundings, facing Mission Bay and only 100 yards from 
beautiful Pacific Ocean beaches.  Spouses and family members can catch a  
convenient Harbor Hopper for a quick trip to Sea World.  After the 
Symposium, plan to spend the weekend visiting La Jolla, the world 
famous San Diego Zoo or Mexico, only 30 minutes by car or Trolley.

A limited number of rooms have been reserved at the Catamaran for the
very special rate of $71.56 single, $88 double.  Reservations, on a
space available basis, can be made by calling (800)-288-0770 and
indicating you are attending the ISOC Security Symposium, or by FAXing
the hotel registration form attached below.  Reservations must be made
before Jan. 15, 1995 to ensure the special rate.

CLIMATE

February weather in San Diego is normally very pleasant.  Early morning 
temperatures average 55 degrees while afternoon temperatures average 67
degrees.  Generally, a light jacket or sweater is adequate during February;
although, occasionally it rains.

TRANSPORTATION

San Diego International Airport is 10 miles (approx. 15 minutes) from
the Catamaran Hotel.  Cloud9 shuttle operates a continuous service
between the airport and the hotel: fare is $6.00.  When you arrive at
the airport, go to the shuttle loading area at either terminal and ask
the attendant to radio for a Cloud9 shuttle to the Catamaran.  Taxi
fare between the airport and the hotel is approx. $20.  The Catamaran
charges $6 per day for parking.

REGISTRATION FEES

Postmarked        Subsequent
by Jan. 6         registration

$320              $365

REGISTRATION INCLUDES

- Attendance        - Symposium Proceedings
- Reception         - Banquet
- Two Luncheons     - Coffee Breaks

ON-SITE REGISTRATION is available Wednesday evening at the reception, and 
Thursday morning at the Symposium.  

FOR MORE INFORMATION on registration contact Gloria Carrier 
by phone at (703)-883-4508 or via email to gcarrier@mitre.org.

==============================================================================

SYMPOSIUM REGISTRATION FORM

Name ______________________________________________________________________

Affiliation _______________________________________________________________

Name on Badge _____________________________________________________________

Special Requirements (e.g., dietary)? _____________________________________

Mailing Address ___________________________________________________________

___________________________________________________________________________

___________________________________________________________________________

Area Code/Phone # _________________________________________________________

Area Code/FAX # ___________________________________________________________

Email Address _____________________________________________________________

[ ]  Check here if you would prefer that your name NOT be included 
     in the list of attendees distributed at the symposium.

Make check (credit cards not accepted) payable to ISOC NDSS SYMPOSIUM.
(Registration is not effective until payment is received).  Mail
registration, no later than February 10, 1994, to:  ISOC Symposium, C/O
Gloria Carrier, The MITRE Corporation, 7525 Colshire Drive, M.S. Z605,
McLean, VA 22102-3481, USA.

==============================================================================

HOTEL REGISTRATION FORM


		   WELCOME ISOC SECURITY SYMPOSIUM
		        February 16-17, 1995
		
	       Single: $71.56		Double: $88.00
	       Triple: $103.00		Quad: $118.00

		        Extra Person $15.00

	        All rates subject to $10.50 room tax

	       Reservations required by: January 15, 1995

        Fax this form to the Catamaran Hotel at (619)-490-3328


Name ______________________________________________________________________

Street ____________________________________________________________________

City ___________________________________ State ___________ Zip ____________

Phone # ________________________________   Number in Party ________________

Arrival Date ___________________________   Departure Date _________________

Roommate(s) ____________________________   Special Needs __________________

Credit Card # __________________________   Expires ________________________

Name on Card ______________________________________________________________

Signature _________________________________________________________________

==============================================================================

-- 
David M Balenson

-----------[000162][next][prev][last][first]----------------------------------------------------
Date:      8 Dec 1994 00:25:43 -0500
From:      mcknight@rbdc.rbdc.com (Michael Mcknight)
To:        comp.protocols.tcp-ip
Subject:   Re: Multiple NICs on a server on same segment

Can't someone out there help me?

I'm all out of ideas!


-Michael McKnight
 mcknight@rbdc.rbdc.com

-----------[000163][next][prev][last][first]----------------------------------------------------
Date:      Wed, 7 Dec 1994 17:51:21 GMT
From:      peter@world.std.com (Peter Loshin)
To:        comp.protocols.tcp-ip
Subject:   IPNG

Sorry if I missed any discussion of this, but any infomration about
what's happening with IPNG would be appreciated.  There've been some
articles lately that don't shed much light on the topic, but indicate
that there has been some activity in this area.  Currently all I know
is this:

RFC 1454
SIP (Simple Internet Protocol)
PIP (???)
TUBA (TCP and UDP  with Bigger Addresses)
TP/IX (??? RFC 1475

Any news and developments would be appreciated; either by posting or
e-mail to me and I'll summarize if there is any demand.

THANKS in advance.

-Peter Loshin
  peter@world.std.com
   1 617 738 1125


-----------[000164][next][prev][last][first]----------------------------------------------------
Date:      7 Dec 1994 18:02:31 GMT
From:      skeung@mazatlan.eng.uci.edu (Stephen Sai-Lung Keung)
To:        comp.protocols.tcp-ip
Subject:   ACM SIGCOMM CCR (special issue on ATM networks)


Hi all, the following is a call for papers.  Please direct
your questions regarding this special issue to the persons below.  
Also, please direct any flames to me (skeung@ece.uci.edu), for I 
may intrude some irrelevant newsgroup(s).

-----------------------------------------------------------------------

                        Call for Papers
                         ACM SIGCOMM
                  Computer Communication Review
                   Special Issue on ATM Networks
 
 The April 1995 issue is planned to be a special issue on ATM networks.
 The objective is to make recent important developments in ATM
 accessible to the networking community at large.  The emphasis is on
 ``hot topics,'' work-in-progress, and practitioner papers related to
 all aspects of ATM networks. In particular, we are soliciting papers
 on the following or related topics:
 
                Traffic Management
                Network Management
                Signaling and Routing
                LAN Emulation
                Switch Architectures
                Broadband InterCarrier-Interface
                Service Aspects and Applications
                 Admission Control
                Multimedia over ATM
 
 
 The schedule is as follows:
 
                 Paper due: January 3, 1995
                 Author notified: February 14, 1995
                 Publication date: April 1995
 
 
 Potential contributors are encouraged to submit their papers in
 electronic forms, such as LaTex, WordPerfect, MS Word (PC or MAC), or
 Postscript. Manuscripts should be sent to one of the two
 co-editors: 
 
 Raj Jain                               Kai-Yeung (Sunny) Siu
 Department of CIS                      Department of ECE
 The Ohio State University              University of California, Irvine
 Columbus, OH 43210-1277                        Irvine, CA 92717
 Phone: +1 614 292 3989                 Phone: +1 714 824 3142
 Fax: +1 614 292 2911                   Fax: +1 714 824 3142
 Email: Jain@ACM.Org                    Email: Siu@ECE.UCI.edu
-----------------------------------------------------------------------


-----------[000165][next][prev][last][first]----------------------------------------------------
Date:      7 Dec 1994 18:17:40 GMT
From:      somani@green.cs.wisc.edu (Amit Somani)
To:        comp.protocols.tcp-ip
Subject:   Q: free buffer space for a UDP buffer ??

Hi, 

Is there some way to find out the amount of free buffer space 
associated with a UDP scoket? The socket has been initialized 
by a call to sectsockopt().

Thanks,
--Amit 

-----------[000166][next][prev][last][first]----------------------------------------------------
Date:      8 Dec 1994 10:15:52 -0800
From:      lstowell@pyrnova.mis.pyramid.com (Lon Stowell)
To:        comp.protocols.tcp-ip
Subject:   Re: SVR4.2 ping source addr

In article <3bobjo$t78@hustle.rahul.net> bsmith@wci.com writes:
>
>How does the SVR4.2 ping command choose the source IP address
>on a machine with multiple interfaces?  Is there any way to 
>control which interface is used as the source?

  The default is to use the local interface that is on the
  same subnet as the pingee.   

  If the pingee is not local, it will pick the interface which
  has a route to the pingee's network.  This could be a static
  route via route add or a routed if routed is enabled.

  A "quaint" feature is that if you have more than one interface
  on the same subnet, frames may come in from any of them but all
  outbound frames will be sent only from the primary interface. 
  Surprisingly enough most net.partners don't seem to care that
  they are sending frames to one mac address and receiving from
  another.



-----------[000167][next][prev][last][first]----------------------------------------------------
Date:      Wed, 7 Dec 1994 21:20:59 GMT
From:      wjw@pt.com (Wald Wojdak)
To:        comp.protocols.tcp-ip
Subject:   Wher can I find a TCP/IP source code

Hi,

I am looking for a TCP/IP source code that runs in the Berkeley socket
environment rather than streams environment. Does anybody know where I can 
find it.

-Thanks
-- 
__
Wald Wojdak, Performance Technologies Incorporated	wjw@pt.com
315 Science Parkway, Rochester, New York 14620            uupsi!ptsys1!wjw

-----------[000168][next][prev][last][first]----------------------------------------------------
Date:      07 Dec 1994 23:03:50 GMT
From:      croten@eosqa.gsfc.nasa.gov (Charles Roten)
To:        comp.unix.solaris,comp.sys.sun.admin,comp.protocols.tcp-ip
Subject:   Re: OK .. now, how to get the SPARCserver 1000 to see it's router ??

In article <CROTEN.94Dec5144747@eosqa.gsfc.nasa.gov> 
croten@eosqa.gsfc.nasa.gov (Charles Roten) writes:

>In article <CROTEN.94Dec2155624@eosqa.gsfc.nasa.gov> 
>croten@eosqa.gsfc.nasa.gov (Charles Roten) writes:
>

 [yell for help about netmask and broadcast address deleted]
 [format wierdness deleted]

>Now, 'ifconfig le0' shows the new interface to be righteous.  But the new 
>machine _cannot_ find it's router.  The twisted pair cable _is_ plugged in 
>to the proper places at both ends.  
>
>I know the gateway for this box ('spsosun') should be 128.183.117.1.  It's 
>_own_ IP address is 128.183.117.20.  But 'netstat -r' gives the following ...
>
>
>    Destination      Gateway            Flags    MTU    RTT RTTvar    Use Interface
>    localhost        localhost          UH         0      0      0   5695  lo0
>    128.183.117.0    spsosun            U          ........................le0
>
>
>'ping' from spsosun fails for both named hosts _and_ hosts specified by IP 
>number.  
>
>Can someone enlighten me about this?  What in chaos is going on?  The system's 
>IP number and name are in /etc/hosts.  /etc/nsswitch.conf is set up for DNS.  
>What have I overlooked?  Thanks.  

Thanks.  /etc/defaultrouter was indeed the target file, as folks let me 
know.  Even though it shouldn't really be neccessary to put the IP address 
of the default router in that file anyway, I did it.  

Still no soap.  It turns out the network guys had plugged the cable into 
the _wrong_ _network_.  Gaaah!  

Ah well.  Things are OK now.  

--
Charles D. Roten               | Hughes STX Inc.
croten@nyx.cs.du.edu           | NASA GSFC (Hurrah DAAC!)
croten@eosdata.gsfc.nasa.gov   | (301) 286-4413 (w), (301) 317-0872 (h)

-----------[000169][next][prev][last][first]----------------------------------------------------
Date:      8 Dec 1994 12:27:59 -0800
From:      guy@netapp.com (Guy Harris)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

Andrew Gabriel <Andrew@cucumber.demon.co.uk> wrote:
>NFS will probably never be as fast as TCP/IP based protocols (FTP, etc),

Will it ever be as fast as TCP-based protocols such as NFS? :-)

(Read Vernon Schryver's postings in this thread - yes, people *do* run
NFS over TCP....)

It may never be as fast over slow links as protocols such as FTP wherein
a single request unleashes a reply containing the entire file, and
wherein the request for the file contains the full pathname of the file,
rather than just an at-most-8K (or at-most-64K, for NFS V3, as I
remember) reply containing the specified chunk of the file, and wherein
component-by-component lookup requests followed by a separate read
requests are required.

-----------[000170][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 00:06:48 GMT
From:      mrpark@ynucc.yeungnam.ac.kr (MiRyong Park)
To:        comp.protocols.tcp-ip.domains,comp.protocols.tcp-ip
Subject:   Reverse mapping errors ?

Do you have any idea about reverse domain mapping ?

Any anonymous FTP site reject my connection.

Also says 'reverse mapping failure'.

How do i solve this problem ?

I have already setting up DNS, so nslookup resolve domain name to ip address...

Please send me any idea ....

to mrpark@ynucc.yeungnam.ac.kr

-----------[000171][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 00:26:29 GMT
From:      adam@comptech.demon.co.uk (Adam Goodfellow)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

In article <D0EI5u.B2r@cid.aes.doe.ca>,
                Bruce Macdonald <bruce@holmes.bed.ns.doe.ca> wrote:

>Okay - tell me I'm crazy - we're trying to use NFS over SLIP...
>
>It's slow, very slow, excruciatingly slow - it's much slower than
>say FTP, httpd or gopher for transferring files - should I expect
>otherwise?  What is it about the NFS protocol that would cause such
>degradation in performance?
>
>

Probably because it run over a remote procedure call system based upon
UDP (unless you have a very new TCP version, if they exist yet...). 

I think the RPC itself can carry rather large overheads on each packet,
and probably makes no attempt to implement any kind of sliding window
mechanism to speed it up over long delay links, so each RPC packet has
to wait for a responce, unlike applications sending data over a TCP
connection.

The end result - *very* slow on anything other than a quick turn-around
LAN.

-- 
Adam

=======================================================================
| Computech  Tel/Fax: 0181 673 7817  email: adam@comptech.demon.co.uk |
=======================================================================

-----------[000172][next][prev][last][first]----------------------------------------------------
Date:      8 Dec 1994 02:05:41 GMT
From:      raj@cup.hp.com (Rick Jones)
To:        comp.protocols.tcp-ip
Subject:   Re: Available buffer space in a UDP socket ??

Amit Somani (somani@sun2.cs.wisc.edu) wrote:
: Is there any way to find out the amount of available (free) space
: associated w/ a UDP socket ? The UDP socket has intitialized its 
: buffer space by a call to setsockopt().

On many (BSDish) systems, the UDP datagram spends so little time on
the SO_SNDBUF (outbound) that it is unlikely (I think) to be an issue.

On inbound, many systems provide an ioctl that says how many bytes are
available for reading on a file descriptor. The app knows what the
SO_RCVBUF size is (through explicitly setting it with setsockopt, or
reading it with getsockopt) and can do the subtraction.

rick jones

-----------[000173][next][prev][last][first]----------------------------------------------------
Date:      Thu,  8 Dec 1994 10:59:59 -0500
From:      Lyle_Seaman@transarc.com
To:        comp.protocols.nfs,comp.protocols.tcp-ip
Subject:   Re: NFS/SLIP

bruce@holmes.bed.ns.doe.ca (Bruce Macdonald) writes:
> Okay - tell me I'm crazy - we're trying to use NFS over SLIP...
> 
> It's slow, very slow, excruciatingly slow - it's much slower than

Andrew@cucumber.demon.co.uk (Andrew Gabriel) writes:
> NFS will probably never be as fast as TCP/IP based protocols (FTP, etc),
> but you should be able to get it near, providing you can get your SLIP
> link to operate without significant packet loss.

tli@cisco.com (Tony Li) writes:
> There are two big problems with this.  First, NFS normally fragments its
> packets.  If you drop one fragment, NFS will retransmit the ENTIRE packet.
> Yes, that's right, ALL fragments.
> 
> Second problem: many NFS implementations disable UDP checksums.  SLIP
> provides no data integrity check.  Your data may be getting corrupted
> too...  what little of it that gets there...

If using error-correcting modems, there should not be any
packet loss unless you are using antique hardware.  Nor should there
be data corruption (incidentally, I think all the workstation vendors
now enable UDP checksums by default. I don't know about things like
PC-NFS. What implementations are you referring to, Tony?).

I think Vernon most likely has the correct analysis:
vjs@calcite.rhyolite.com (Vernon Schryver) writes:
> which clobbers performance.  If you use a big MTU (at least 1006) then
> the cost of the IP headers on the fragments is not bad and the lack of
> header compression is insignificant.  I'd expect NFS to be about 60% as
> fast as FTP.

If NFS is not about 60% as fast as FTP, you should look at twiddling
the retransmission timeout parameters -- you may be retransmitting too
agressively and chewing up bandwidth.

-----------[000174][next][prev][last][first]----------------------------------------------------
Date:      8 Dec 1994 03:29:51 GMT
From:      bbradfor@glenayre.com (Bob Bradford [4525])
To:        comp.os.vxworks,comp.protocols.tcp-ip
Subject:   Simultaneous TCP/Socket Connections

The TCP spec allows for the case when two hosts connect to each
other at the same time (using the same port). The connection
handshake successfully resolves the two attempts into one
connection between the hosts. [RFC793, pg 32, Figure 8]

We would like to use this feature so our hosts can use and
offer a particular service to each other. We want to have
each host listen on the service port. When any host needs a
connection to another host, it uses the service port locally
and attempts to connect to the service port of the remote
system.

If the opposite end also does the same thing at the same time,
the TCP simultaneous connection mechanism resolves this into
one shared connection. On the other hand, if the opposite wants
a connection at a later time, it finds it already has one and
uses it.  (Duplicate connections are not suitable for the
application.)

My attempts using the socket interface have not succeeded. I
cannot bind two sockets to the same port number (one to listen
on and one to connect on).  I also cannot listen on a socket
and attempt a connect on the same socket.

Has anyone ever tried to do this using a socket interface and
had success?  Can anyone suggest how to get what we want?


---

Bob Bradford				Email: bbradfor@glenayre.com
Glenayre Electronics                    Phone: (604) 293-1611
1570 Kootenay Street                    Fax:   (604) 293-4317
Vancouver, BC, Canada  V5K 5B8



-----------[000175][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 10:29:23
From:      markw@vip.best.com (Mark Williams)
To:        comp.protocols.tcp-ip
Subject:   TCP-IP consultants

My company is urgently needing to set up a TCP/IP network. We have a very 
short time in which to make key decisions about the addressing scheme we will 
implement. Can anyone recommend a consultant in the Bay Area who can provide a 
day's worth of consultancy in making these decisions (we are reasonably well 
equipped to make these decisions but want to be absolutely sure we are not 
making the wrong moves).

Thanks, Mark


-----------[000176][next][prev][last][first]----------------------------------------------------
Date:      8 Dec 1994 05:51:34 GMT
From:      dave@youtools.com (Dave Van Allen)
To:        comp.protocols.tcp-ip.domains,comp.protocols.tcp-ip
Subject:   Re: Reverse mapping errors ?

MiRyong Park (mrpark@ynucc.yeungnam.ac.kr) wrote:
:>Do you have any idea about reverse domain mapping ?
 
:>Any anonymous FTP site reject my connection.


If you are running the DNS and are the authority for your IP and       
and domain, then check for incorrect in-addr.arpa records in named.boot
and failing that, ensure that the NIC has your nameserver listed
as the DNS for reverse lookup.

If you are getting your IP number from someone else, and they are the
reverse-lookup listed at the NIC, ask them what's wrong.

--

 *Dave Van Allen - You Tools/FASTNET - dave@youtools.COM - (610) 954-5910
        -=-=-=- FASTNET(tm) PA/NJ/DE Internet 800-967-2233 -=-=-=-

-----------[000177][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 06:20:36 +0000
From:      Andrew@cucumber.demon.co.uk (Andrew Gabriel)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Cc:        bruce@holmes.bed.ns.doe.ca
Subject:   Re: NFS/SLIP

In article <D0EI5u.B2r@cid.aes.doe.ca>
           bruce@holmes.bed.ns.doe.ca "Bruce Macdonald" writes:

> Okay - tell me I'm crazy - we're trying to use NFS over SLIP...
> 
> It's slow, very slow, excruciatingly slow - it's much slower than
> say FTP, httpd or gopher for transferring files - should I expect
> otherwise?  What is it about the NFS protocol that would cause such
> degradation in performance?

With FTP etc, TCP is responsible for timing out and retransmitting,
and most TCP implementations work out their timers dynamically based
on the observed delays of each connection. If an IP packet is lost,
TCP can resend it very quickly, say within a few times of the round
trip delay.

NFS operates over UDP/IP, not TCP/IP. UDP is a sort of "send and
forget" protocol - it doesn't know or care if the data is lost.
This means that NFS clients are responsible for retransmissions.
Most clients use fixed parameters specified (or defaulted) on their
mount commands. These are not dynamic, and the timeouts tend to be
very much longer than the equivalent TCP timeouts would be. The
timeouts are not just to cope with network delays, but must also
cope with NFS server delays, which can be of the order of seconds.

This should show why even a small rate of IP segment loss will have
a profound effect on NFS, but not on TCP (FTP, etc).

There is one more point to consider, fragmented IP datagrams.
If you send an IP packet which is bigger than the maximum transfer
unit size (MTU) of your link, it will be fragmented. I'm not too
familiar with SLIP, but I guess your MTU will be somewhere between
256 and 1500. NFS clients usually default to using much larger reads
and writes than this, say 4k or 8k bytes. These will be fragmented
into between 4 and 30+ IP fragments over your SLIP link. If *one*
of these fragmanets is lost, *all* must be retransmitted. You should
therefore ensure that your read and write size (usually specified on
the mount) is smaller than your MTU (to allow for RPC/UDP/IP headers).
Basically, experiment and see what's best for you - assuming you're
using a large default at the moment, try 1024, 512, and 256 for
starters.

NFS will probably never be as fast as TCP/IP based protocols (FTP, etc),
but you should be able to get it near, providing you can get your SLIP
link to operate without significant packet loss.

I hope this helps.
-- 
Andrew Gabriel                        Home: Andrew@cucumber.demon.co.uk
Consultant Software Engineer          Work: Andrew.Gabriel@gpt.co.uk

-----------[000178][next][prev][last][first]----------------------------------------------------
Date:      8 Dec 1994 06:29:45 GMT
From:      ant@cc.uq.oz.au
To:        comp.protocols.tcp-ip
Subject:   DOS PPP/SLIP server

Hi all,

A couple of weeks ago I noticed a DOS based PPP/SLIP server with
authentication etc.  At the time I thought I'd probably never need such a 
thing.  Guess what - I was wrong.  I'd dearly like to see this app know, 
but cannot find the reference for the life of me.  Can anyone help me out ?

Cheers
ant (ant@cc.uq.oz.au)

-----------[000179][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 11:42:02
From:      KTNAGEL@Materna.DE (Thomas Nagel)
To:        comp.protocols.tcp-ip
Subject:   Re: udp forwarding

In article <3bh18n$9td@mkas.mcs.com> martin@mkas.mcs.com (Matthew Martin) writes:
>From: martin@mkas.mcs.com (Matthew Martin)
>Subject: udp forwarding
>Date: 29 Nov 1994 23:06:31 -0600

Hello,

>Can anyone tell me if a novell 3.12 server with a plain setup, no add on's
>does udp forwarding, using direct or indirect mode?

The Novell NetWare Server OS v3.x and v4.x contains a TCPIP Router.
Q: What do you mean with "udp forwarding"? 
Do you need more than IP routing for it?

>I was told by one of our venders that his equipment world not work because
>novell uses udp indirect routing, and we use a novell router.

Q: What is udp indirect routing?

>Thank you for any info.
>-Matt.


Thomas Nagel


Thomas Nagel, Network Programmer     Dr. Materna GmbH, Dortmund, Germany
IN: Thomas.Nagel@Materna.De          CI$: 100143,3665

-----------[000180][next][prev][last][first]----------------------------------------------------
Date:      8 Dec 1994 06:49:00 GMT
From:      jsilva@cello.gina.calstate.edu (Jeff Silva)
To:        comp.protocols.tcp-ip
Subject:   Need help with Mosaic, PLEASE  -  mos [1/1]


I an running Mosaic at home on my PC. It is the latest version
and I am using the trumpet winsock 1.0. Things have been just fine
for about 3 weeks. Now, all of a sudden, for no reason I can think
of, there are only 2 url's that I can connect to. I've used the
hotlist, or typed them in but all I get is it trying to load
the http, folowed by SOCKET timed out. It tries for about 30 seconds,
but you can see that it's not going anywhere. I can't figure this out
since there are still two that I can load. I've realoaded Mosaic.exe,
the original mosaic.ini file, tcpman.exe, the dll, and win32. No
luck. Can it be related to my slip connection? Can it have something
to do with the size of the url that it's downloading and something
that got changed on my system.
Any help is much appreciated.


-----------[000181][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 09:22:57 GMT
From:      hjb@netcom.com (squeedy)
To:        comp.os.vxworks,comp.protocols.tcp-ip
Subject:   Re: Simultaneous TCP/Socket Connections

Bob Bradford [4525] (bbradfor@glenayre.com) wrote:
: The TCP spec allows for the case when two hosts connect to each
: other at the same time (using the same port). The connection
: handshake successfully resolves the two attempts into one
: connection between the hosts. [RFC793, pg 32, Figure 8]

this is true, but that case is intended to solve a collision problem
and keep the state-machine consistent when two active connects get initiated
simulatneously against each other -- not for what you seem to be trying to do.

: We would like to use this feature so our hosts can use and
: offer a particular service to each other. We want to have
: each host listen on the service port. When any host needs a
: connection to another host, it uses the service port locally
: and attempts to connect to the service port of the remote
: system.

this is not possible in 4.X BSD UNIX derived TCP/IP, as in
most UNIX systems as well as VxWorks.  see below...

: My attempts using the socket interface have not succeeded. I
: cannot bind two sockets to the same port number (one to listen
: on and one to connect on).  I also cannot listen on a socket
: and attempt a connect on the same socket.

1) you cannot bind two socket objects to a port; a TCP socket is an
	abstraction of a TCP port.

2) you cannot listen and then do a connect on the same socket because
	all 4.X BSD UNIX derived TCP/IP implementations specifically
	prohibit this at the top kernel socket level.  it is
	not a "host requirement" to provide this capability, and
	doing so in 4.X BSD would break its TCP state-machine 
	implmentation.  i know of no implementation of TCP that
	will allow this.  probably because, the original TCP RFC 
	was quite specific about the active vs. passive OPEN states.  
	however, see the next point...

3) you MAY, however, be able to do a listen/accept on a socket
	that is already blocked upon a connect() in progress.  this is
	part of a "host requirement" recommendation, intended
	primarily for busy server protocols such as SMTP (sendmail).
	note that this is the opposite order of what you tried.
	looking at the 4.X BSD code very quickly now, it seems
	that this is allowed.  i have not personally tried this.

-- 
Hwa-Jin Bae (hjb@netcom.com)
Peaceful Star, Oakland, CA		ftp.netcom.com:/pub/hj/hjb/README

-----------[000182][next][prev][last][first]----------------------------------------------------
Date:      8 Dec 1994 09:58:07 GMT
From:      tli@cisco.com (Tony Li)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

In article <D0EI5u.B2r@cid.aes.doe.ca> bruce@holmes.bed.ns.doe.ca (Bruce
Macdonald) writes: 

    Okay - tell me I'm crazy - we're trying to use NFS over SLIP...
    
Okay -- you're crazy.  ;-)

    It's slow, very slow, excruciatingly slow - it's much slower than
    say FTP, httpd or gopher for transferring files - should I expect
    otherwise?  What is it about the NFS protocol that would cause such
    degradation in performance?
    
There are two big problems with this.  First, NFS normally fragments its
packets.  If you drop one fragment, NFS will retransmit the ENTIRE packet.
Yes, that's right, ALL fragments.

Second problem: many NFS implementations disable UDP checksums.  SLIP
provides no data integrity check.  Your data may be getting corrupted
too...  what little of it that gets there...

Tony

-----------[000183][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 10:33:08 GMT
From:      jin@spdcc.com (Jerry Natowitz)
To:        comp.protocols.tcp-ip
Subject:   Re: dynamic routing & nfs

In article <3c5r7m$i30@tools.near.net>,
Barry Margolin <barmar@nic.near.net> wrote:
>In article <3c4fps$t0g@panix2.panix.com> rjm@panix.com (rjm) writes:
>>  My vendor is telling me that nfs keeps timing out across the routers
>>  (70 blocks) and that there is too much broadcast traffic across the
>>  network (non-ip stuff). When we do a netstat we get indications that
>>  the routers were are assigned to are not giving us a direct route between
>>  them. we have to cross about ten subnets to get to the router we want t
>>  to get to.
>
>If the dynamic routing protocol is choosing a long, slow route rather than
>the short one, there's a problem with the routing protocol.  Even a simple
>protocol like RIP won't choose 10 hops when there's a direct connection.
>What routing protocol are you using on your network?
>-- 
>
>Barry Margolin
>BBN Internet Services Corp.
>barmar@near.net

As a separate issue: AIX does not seem to do well with the default NFS
buffer size, which I believe is 8192.  Try adding ",wsize=1024,rsize=1024"
to the options stanza of each NFS entry in /etc/filesystems.  We got about
10-fold improvement in NFS performance due to fewer timeouts.
-- 
     Jerry Natowitz - jin@spdcc.com

-----------[000184][next][prev][last][first]----------------------------------------------------
Date:      08 Dec 1994 11:14:46 GMT
From:      pcj@merlin1.meche.psu.edu (Paul C. Janzen)
To:        comp.protocols.tcp-ip
Subject:   Local-only IP addresses?


We recently acquired a color printer with built-in TCP/IP & LPR support, 
as well as an Ethernet connection. 

We would like to prevent hosts from other subnets from being able to access
the printer. Can I choose the IP number so that only machines on our
segment can see the printer?

Thanks!!!

------------------------------------------------------------------------
Paul C. Janzen                Receiving a million dollars tax free will
pcj@sabine.acs.psu.edu        make you feel better than being flat broke
                              and having a stomach ache.   -- Dolph
                              Sharp, "I'm O.K., You're Not So Hot"
------------------------------------------------------------------------


-----------[000185][next][prev][last][first]----------------------------------------------------
Date:      8 Dec 1994 12:44:43 GMT
From:      Peter Aikins <p_aikins@postoffice.utas.edu.au>
To:        comp.protocols.tcp-ip
Subject:   Please Subscribe

please subscibe p_aikins@postoffice.utas.edu.au

-----------[000186][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 09:51:03 +0200
From:      Peter K <pko@kiko.paradigm.co.za>
To:        comp.protocols.tcp-ip
Subject:   Re: bootpclient for MS-DOS/Windows

Hi, Russell

>    I am looking for a bootp client running under MS-Dos and Windows, either one 
>    executable which can be used by both or seperate ones.
> 
> What do you plan to do with the data that the bootp client returns?
> It's not much good unless you're going to use it with a TCP/IP package...

We are already running tcp/ip. The only thing is to allocate IPs and set
gateways to individual PCs. Under Win NT AS one can run MS's equivalent
but then I have to assign a pool of IP addresses, which is something I
don't want to do. We have a 300+ box network (running Slo-es/2, SCO,
Linux, SunOS, AIX, MS-Dos, Win 3.11, WFW and Win NT) accross 4 class C
subnets and from a management perspective (under our circumstances) the
assignment of an IP-address pool is just not on. 

PETER KOOIMAN                 | Fax  : ++27-12-342-1754
Paradigm Systems Technology   |
Pretoria, South Africa        | DNS  : pko@paradigm.co.za
++27-12-342-1145              | UUCP : pko@silmaril.UUCP


-----------[000187][next][prev][last][first]----------------------------------------------------
Date:      8 Dec 1994 22:29:02 -0500
From:      mcknight@rbdc.rbdc.com (Michael Mcknight)
To:        comp.unix.solaris,comp.protocols.tcp-ip
Subject:   Re: TCP/IP and IPX/SPX routing

If you have a novell server, let it do the routing.  Novell version 3.12 and
later (maybe even 3.11) have a NLM called TCPIP.NLM (and some other utils)
that will allow the novell server to pass IP and IPX.

You could always get a dedicated router, maybe Cicso or WellFleet... there
are many others out there as well that can do IP and IPX.

Good luck

-Michael McKnight
 mcknight@rbdc.rbdc.com

-----------[000188][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 14:43:38 GMT
From:      prasert@vast.unsw.edu.au (Prasert Kanthamanon)
To:        comp.protocols.tcp-ip
Subject:   HELP: Subnet routing.


Dear Netters,

We have a problem to configure a routing system to handle subnets.  Our
network topology is shown below.


			+-----+
			| G3  |
			+-----+
			   |IP: 202.44.8.3
			   |Netmask: 255.255.255.00
			   |				   Backbone
---------------------------------------------------------- Network: 202.44.8.0
	|					|
	|IP: 202.44.8.5				|IP: 202.44.8.7
	|Netmask: 255.255.255.00		|Netmask: 255.255.255.00
     +-----+				     +-----+
     | G1  |				     | G2  |
     +-----+				     +-----+
	|IP: 202.44.9.33			|IP: 202.44.9.65
	|Netmask: 255.255.255.224		|Netmask: 255.255.255.224
	|					|
	|				------------- 
     -----------			Subnet: 202.44.9.64
     Subnet: 202.44.9.32

G1 is a PC running Netware.
G2 and G3 are SUN SPARC workstations running SunOS 4.1.3.

We tried to use static routing on G2 and G3 to manage the routing task.
However, there is a problem in the routing system.  The problem can be
explained like this.
Note: In every case, we have a default entry to a main gateway.

From G3:
========
1) If there are two subnets that are from the same class C network in the
   routing table of the machine, a request to an IP address in that class
   C will go through the gateway of the first subnet of that class C number

For example:
If we configure the routing table on the gateway G3 like below:
Routing table
Destination          	Gateway 
......
202.44.9.32		202.44.8.5
202.44.9.64		202.44.8.7
.....
In this case, a request to address 202.44.9.xx will go through
the gateway 202.44.8.5.

If the routing table is
Destination             Gateway 
......
202.44.9.64             202.44.8.7
202.44.9.32             202.44.8.5
.....
a request to address 202.44.9.xx will go through the gateway 202.44.8.7.

2) When there is only one subnet routing entry in the table, a request to
   every address in that class C number will go through the gateway of the
   subnet.
For example:
Routing table:
Destination             Gateway 
......
202.44.9.64             202.44.8.7
A request to address 202.44.9.xx will go through the gateway 202.44.8.7.

From G2:
========
The routing works correctly for a subnet that is added in its routing table
if it is a subnet from the same class C number as the one that has a direct
connection to it.  In addition, the subnets must have the same netmask.

For example:
If the routing table of G2 is
Destination             Gateway		Flags
......
202.44.9.32             202.44.8.5	UG
202.44.9.64             202.44.9.65	U
.....
all requests for IP address of subnet 202.44.9.32 and 202.44.9.64 will be
delivered to the right gateway. 
-------------------------------------------------------------------

My questions are:
=================
1. Is this network configuration valid?
2. What cause the problem?
3. How can I configure the routing system to handle these subnets?
4. Can OSPF based routing solve the problem?

Any help would be appreciated.

Thank you in advance.

Prasert.
----------------------------------------------------------------------------
Prasert Kanthamanon            | Computer and Systems Technology Laboratory,
                               | School of Computer Science & Engineering,
                               | The University of New South Wales
prasert@vast.unsw.edu.au       | Sydney 2052, Australia.
----------------------------------------------------------------------------

-----------[000189][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 15:26:10 GMT
From:      vjs@calcite.rhyolite.com (Vernon Schryver)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

In article <19941208.002629.38@comptech.demon.co.uk> adam@comptech.demon.co.uk writes:
>In article <D0EI5u.B2r@cid.aes.doe.ca>,
>                Bruce Macdonald <bruce@holmes.bed.ns.doe.ca> wrote:
>
>>Okay - tell me I'm crazy - we're trying to use NFS over SLIP...
>>
>>It's slow, very slow, excruciatingly slow - it's much slower than
>>say FTP, httpd or gopher for transferring files - should I expect
>>otherwise?  What is it about the NFS protocol that would cause such
>>degradation in performance?
>
>Probably because it run over a remote procedure call system based upon
>UDP (unless you have a very new TCP version, if they exist yet...). 

TCP implementations of NFS have been common for a few years.  The free
source implementation of NFS from U.Guelph that is present in the free
BSD UNIX implementations as well as BSDI's product supports NFS/TCP/IP.
I seem to recall something about OSF supporting NFS/TCP/IP, but I could
easily be wrong about that.

However, NFS/TCP/IP is not significantly faster than NFS/UDP/IP.  If
you want a proof by authority, dig up Rick Macklem's words to that effect
here.  (He is the author of the U.Guelph code).  If you want a rational
explanation, remember than NFS is a request-response protocol regardless
of what is below, and that its only windowing or other kinds of double
buffering comes from the read-ahead and write-behind behavior, if any,
of NFS client file system.  TCP helps NFS on wide area links by congestion
losses, and by doing retransmissions of 1 MTU at a time instead of a full
request-response (ie. no IP fragmentation).


>I think the RPC itself can carry rather large overheads on each packet,

The per-packet overhead of the rpc/xdr use by NFS is about 100 bytes.
(100 is an inconvenient value for NFS on FDDI rings between hosts with
a 4096 page size because it is not divisible by 8.  10 bonus points to
anyone who know why I care about that fact).

If you use the common default rsize and wsize of 8K and the common SLIP
MTU Of 1006, for each big datagram you will have 100 bytes of rpc/xdr,
12 bytes of UDP, and 9 IP headers each 20 bytes long, for a total overhead
of 292, which amounts to a total overhead of 3.5%.  Compare that to the
2.7% of FTP/TCP/IP without VJ header compression and 0.3% with header
compression.  In other words, for the data NFS packets, the headers are
irrelevant.

The headers on the requests, attributes, and status answers do matter,
but only because they increase the latency of the requests, which is
not good for a request-response protocol.


>and probably makes no attempt to implement any kind of sliding window
>mechanism to speed it up over long delay links, so each RPC packet has
>to wait for a responce, unlike applications sending data over a TCP
>connection.
> ...

Yes, the trouble is that NFS is a request-response protocol.


Vernon Schryver    vjs@rhyolite.com

-----------[000190][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 15:35:46 GMT
From:      vjs@calcite.rhyolite.com (Vernon Schryver)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

In article <3c6lbf$jnu@cronkite.cisco.com> tli@cisco.com (Tony Li) writes:

> ...
>Second problem: many NFS implementations disable UDP checksums.  SLIP
>provides no data integrity check.  Your data may be getting corrupted
>too...  what little of it that gets there...

Please name a common, current system with UDP checksums off by default.

Once upon a time, UDP checksums were off by default in a very common
UNIX TCP/IP implementation.  If either the NFS client or server is one
of those old systems, it would be wise to turn on UDP checksums before
using NFS over SLIP.  Of course, unless the client or server is as slow
as a system from the mid-1980's, it would be wise to turn on UDP checksums
even if both are on the same Ethernet.

It is a gross overstatement to say SLIP has no data integrity check.
The most common data error on SLIP links today is lost bytes.  Most
modern modems use v.42 or NMP, which eliminates twiddled bits (i.e.
corrupted bytes) almost as well as the Ethernet CRC does on an Ethernet.
The remaining errors are lost bytes between the modems and computers.
Lost bytes are reliably detected by the header and data lengths in the
IP header, and the IP header is protected by the IP checksum.


Vernon Schryver    vjs@rhyolite.com

-----------[000191][next][prev][last][first]----------------------------------------------------
Date:      8 Dec 1994 15:37:04 GMT
From:      smeets@cci.de (Vincent Smeets)
To:        comp.unix.solaris,comp.protocols.tcp-ip
Subject:   TCP/IP and IPX/SPX routing

Hallo,

Is it possible to do TCP/IP and IPX/SPX routing with one host?

Our problem is: 
We have now one physical network with one class C IP network (for Unix
and PC's) and a Novell server (for the PC's).  As we run out of IP
numbers, we orderd an other class C IP network number.
Now, we want to split out physical network in two parts and assign
every part an IP number.
Then we want to connect these two network with a router. The only
problem is that this router must pass the IPX/SPX packages too.

We were thinging about a Sun with Solaris, but as far as I know,
doesn't is foward the IPX/SPX packages. Is there some software for the
Sun to make it an IPX/SPX forwarder, or must I use some other hardware
(which)?

Please give me some hints
Vincent


---
--      Vincent Smeets                  Competence Center Informatik GmbH
--      Tel. : +49-5931-805461          Postfach 1225
--      EMail: smeets@cci.de            49702 Meppen, Germany



-----------[000192][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 00:22:00 -0500
From:      msbeebe@mtu.edu (Matt Beebe)
To:        comp.protocols.tcp-ip
Subject:   FTPD Problem

I'm running in.ftpd via inetd...  my question is, how do
I enable anonymous access?

				-Matt


-----------[000193][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 16:47:23 GMT
From:      rick@CERCA.UMontreal.CA (Richard Lefebvre)
To:        comp.os.ms-windows.apps.comm,comp.os.ms-windows.networking.tcp-ip,comp.protocols.tcp-ip,alt.dcom.slip-emulators
Subject:   Re: Twinsock 1.2 on IRIX 5.2

dries@elvis.rowan.edu (PowerSlave) writes:

>dries@elvis.rowan.edu (PowerSlave) writes:
 
>>Hello All.
 
>>I've been trying to get a successful TSHOST going since TwinSock v1.0,
>>to no avail.
 
>>I've been able to get tshost v1.1 and 1.2 to compile fine (1.1 after a
>>bit of editing) on a SunOS machine (SPARC), however, my local host is an
>>SGI Indy running IRIX 5.2. I've been able to compile tshost 1.1 and 1.2
>>find on the Indy, with just a couple of edits (on 1.1, 1.2 compiles
>>fine) ... The problem I have is after the TwinSock gets connected,
>>running applications:

I didn't need to edit anything, I just added -cckr to the cc line to
get it complied. But it still doesn't work.

>>I've tried the following under OS/2 v2.10 WinOS2, OS/2 v2.11 WinOS2, and
>>Windows for Workgroups v3.11, I get the same result in all
>>circumstances:
 
>>Twinsock starts `normally', without any errors or incident. I fire up
>>Netscape 0.94, and on the SunOS built tshost, everything works fine.
>>Under the IRIX 5.2 built tshost, I always get a `Twinsock caused a GPF
>>in WINSOCK.DLL at address 0001:14AC' .. the address almost never
>>changes, although I have gotten a 0002:xxxx once.

I get the same error. I tried IRIX 4.0.5 and had a few problems
I still didn't finish trying that avenue. 

>>Unless the SGI Indy is 64bit... then perhaps that is my problem. Anyone
>>working on a 64bit port?
 
>To follow up on my own post, no Scarlet, the SGI Indy is not 64bit... It
>outputs the same size ints as the SPARC running SunOS, so that isn't the
>reason. There goes that theory. 
 
>So, can anyone else suggest a reason that Twinsock likes SunOS and not
>IRIX?

I second that question.

Rick
-- 
Richard Lefebvre, Co-sys-admin, CERCA                                    "42"
http://www.CERCA.UMontreal.CA/~rick/                                -- THGTTG

-----------[000194][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 16:50:10 GMT
From:      landin@cherokee.nsuok.edu (Mark C. Landin)
To:        comp.protocols.tcp-ip,comp.dcom.lans.fddi,comp.dcom.lans.ethernet
Subject:   Max. # of routers on a network?

Hi. We are about to bridge together our Ethernet network which has 1 router
to our FDDI backbone, which has about 30 routers (which are actually Novell
File Servers each connecting a Token-Ring network to the backbone) on it.
TCP/IP is the protocol in question. 

Is having 30+ routers on one network kosher? Even if it is, are there
"gotchas" to it? Thanks for any replies.

-- 
*----------------------------------------------------------------------------*
* Mark C. Landin                  "Those who wish to be must put aside the   *
* Systems Manager                  alienation; get on with the fascination,  *
* Northeastern State Univ.         the real relation, the underlying theme"  *

-----------[000195][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 17:41:31 GMT
From:      evansmp@mb51913.aston.ac.uk (Mark Evans)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

Adam Goodfellow (adam@comptech.demon.co.uk) wrote:
: In article <D0EI5u.B2r@cid.aes.doe.ca>,
:                 Bruce Macdonald <bruce@holmes.bed.ns.doe.ca> wrote:
 
: >Okay - tell me I'm crazy - we're trying to use NFS over SLIP...
: >
: >It's slow, very slow, excruciatingly slow - it's much slower than
: >say FTP, httpd or gopher for transferring files - should I expect
: >otherwise?  What is it about the NFS protocol that would cause such
: >degradation in performance?
: >
: >
 
: Probably because it run over a remote procedure call system based upon
: UDP (unless you have a very new TCP version, if they exist yet...). 
 
: I think the RPC itself can carry rather large overheads on each packet,
: and probably makes no attempt to implement any kind of sliding window
: mechanism to speed it up over long delay links, so each RPC packet has
: to wait for a responce, unlike applications sending data over a TCP
: connection.

You should be able to tune the block size. e.g. so you avoid IP fragmentation.

-----------[000196][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 17:46:27 GMT
From:      evansmp@mb51913.aston.ac.uk (Mark Evans)
To:        comp.os.vxworks,comp.protocols.tcp-ip
Subject:   Re: Simultaneous TCP/Socket Connections

squeedy (hjb@netcom.com) wrote:

: 2) you cannot listen and then do a connect on the same socket because
: 	all 4.X BSD UNIX derived TCP/IP implementations specifically
: 	prohibit this at the top kernel socket level.  it is
: 	not a "host requirement" to provide this capability, and

None of the RFC's appear clear on this. 
One posibility is to treat this as the simaltanious open.

: 	doing so in 4.X BSD would break its TCP state-machine 
: 	implmentation.  i know of no implementation of TCP that
: 	will allow this.  probably because, the original TCP RFC 
: 	was quite specific about the active vs. passive OPEN states.  
: 	however, see the next point...

-----------[000197][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 17:56:03 GMT
From:      cist00n@t4d.carrier.utc.com (Mark Marsula)
To:        comp.protocols.tcp-ip
Subject:   Pinging QMS printers

I'm pinging two QMS printers from an OSF/1 system.  I am getting messages that
say there are duplicate addresses when I ping the printers.  Our Sniffer does
not detect any duplicate addresses on that subnet.  One printer is working
fine, but the other is occasionally losing prints.

Can anyone tell me what triggers the duplicate address message?  Is it caused 
by multiple echo replies?  Is it timing?  It is checking the MAC address on
the reply packet?

Thanks.

-----------[000198][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 18:12:51 GMT
From:      wjw@pt.com (Wald Wojdak)
To:        comp.protocols.tcp-ip
Subject:   Q: TCP/IP source code

Hi,

Can anyone point me to an ftp location where I could find a TCP/IP source code.

-Thanks
-- 
__
Wald Wojdak, Performance Technologies Incorporated	wjw@pt.com
315 Science Parkway, Rochester, New York 14620            uupsi!ptsys1!wjw

-----------[000199][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 19:29:35 GMT
From:      doug@eng.auburn.edu (Doug Hughes)
To:        comp.protocols.tcp-ip
Subject:   Re: POP server for SunOS?

In article <clyde.786507279@trojan>, clyde@hitech.com.au (Clyde Smith-Stubbs) writes:
> Does anyone know of a POP server that will run under SunOs?
> 
> TIA.
> --
>  Clyde Smith-Stubbs       | HI-TECH Software,       | Voice: +61 7 300 5011
>  clyde@hitech.com.au      | P.O. Box 103, Alderley, | Fax:   +61 7 300 5246
>  ...!nwnexus!hitech!clyde | QLD, 4051, AUSTRALIA.   | BBS:   +61 7 300 5235
>                     HI-TECH C: Compiling the real world...

We have been very happy with

ftp://lilac.berkeley.edu/pub/pop/popper/popper-1.831beta.tar.Z

(despite the "beta" in the name the code is solid and has been for many months)
-- 
____________________________________________________________________________
Doug Hughes					Engineering Network Services
System/Net Admin  				Auburn University
			doug@eng.auburn.edu
"The Light at the end of the tunnel is the headlamp of an oncoming train"

-----------[000200][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 20:15:33 GMT
From:      NADEL@litc.lockheed.com (Ron Nadel)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

In article <D0F2BL.ID8@calcite.rhyolite.com> vjs@calcite.rhyolite.com (Vernon Schryver) writes:

>In article <D0EI5u.B2r@cid.aes.doe.ca> bruce@holmes.bed.ns.doe.ca (Bruce Macdonald) writes:
>>Okay - tell me I'm crazy - we're trying to use NFS over SLIP...
>>
>>It's slow, very slow, excruciatingly slow - it's much slower than
>>say FTP, httpd or gopher for transferring files - should I expect
>>otherwise?  What is it about the NFS protocol that would cause such
>>degradation in performance?
 
>  If you use a big MTU (at least 1006) then
>the cost of the IP headers on the fragments is not bad and the lack of
>header compression is insignificant.  I'd expect NFS to be about 60% as
>fast as FTP.

This is true, Vernon, but a large MTU over SLIP is not recommended in the 
event of retransmits.  It is better to retransmit a small packet or two then a 
couple huge ones.  

Bruce, if you can tune your MTU, try it at 512 or less even (depends on the 
line speed and error rates.

Ron

-----------[000201][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 20:20:42 GMT
From:      jmorriso@bogomips.ee.ubc.ca (John Paul Morrison)
To:        comp.os.linux.development,comp.os.linux.admin,comp.os.linux.misc,comp.protocols.tcp-ip,comp.protocols.tcp-ip.domains
Subject:   Re: Multi-homed (multiple IP addresses) Linux hosts?

In article <3c586q$jvr@giant.dswi.com>, Pete Kruckenberg <pete@dswi.com> wrote:
>I saw a reference in another newsgroup to the capabilities of some new
>*nix OS's that allow a host to have multiple IP addresses on a single
>network interface (using an "ifconfig alias" command). 
>
>Does Linux have this capability, or anything similar to it? What other
>OS's (free or commercial) support this capability? What do I need to
>get/have to have this capability?

Linux already has this: ifconfig the "dummy" interface (make sure you've
compiled it in). Then update your routing and if necessary, your arp
tables (proxy arp for that IP address). 

You can compile in more dummy interfaces if you need them. 

It would be nice to have this rolled into ifconfig so you only need
to use one command.

There's one little problem, but it probably affects other operating
systems too, is that people can connect to your second IP address, but
if you initiate a connection to them, the source IP address (on your
outgoing packet) will be the IP address of the eth0 device, not of the
dummy device. Normally this is good, but some hosts might require that
you connect using the second IP address as your source address, and
there's no way to change this behaviour without rewriting each network
app.

But if you need to use the dummy interface, it's probably because
something is really evil on your network, like braindead routers/hosts
that don't support classless routing or really bizarre routing setups
that you are powerless to change, so you are forced to kludge it.


>
>Thanks.
>Pete Kruckenberg
>pete@dswi.com
>


-- 
---------------------------------------------------------------------------
BogoMIPS Research Labs  --  bogosity research & simulation  --  VE7JPM  -- 
jmorriso@bogomips.ee.ubc.ca ve7jpm@ve7jpm.ampr.org jmorriso@ve7ubc.ampr.org
---------------------------------------------------------------------------

-----------[000202][next][prev][last][first]----------------------------------------------------
Date:      8 Dec 1994 20:36:36 GMT
From:      bbradfor@glenayre.com (Bob Bradford [4525])
To:        comp.os.vxworks,comp.protocols.tcp-ip
Subject:   Re: Simultaneous TCP/Socket Connections

In article Fyz@netcom.com, hjb@netcom.com (squeedy) writes:
>Bob Bradford [4525] (bbradfor@glenayre.com) wrote:
>: The TCP spec allows for the case when two hosts connect to each
>: other at the same time (using the same port). The connection
>: handshake successfully resolves the two attempts into one
>: connection between the hosts. [RFC793, pg 32, Figure 8]
>
>this is true, but that case is intended to solve a collision problem
>and keep the state-machine consistent when two active connects get initiated
>simulatneously against each other -- not for what you seem to be trying to do.
>

The collision problem is what I want to handle. I want to
resolve colliding connect attempts to resolve into a single
shared connection.  But this only works if both sides are using
the same port number for each end of the connection (e.g.  port
5000 connecting to port 5000). If this is not true, then you
get 2 distinct connections.

If the connect attempts don't collide, then one connection will
be established before a second would be attempted, and both
sides will then use this single connection for as long as it is
needed (a long time in this application).

>: We would like to use this feature so our hosts can use and
>: offer a particular service to each other.   ...
 
>this is not possible in 4.X BSD UNIX derived TCP/IP, as in
>most UNIX systems as well as VxWorks.  see below...
>

That's the answer I was afraid of, but I am not surprised to hear it.

>: My attempts using the socket interface have not succeeded. I
>: cannot bind two sockets to the same port number (one to listen
>: on and one to connect on).  I also cannot listen on a socket
>: and attempt a connect on the same socket.
>
>1) you cannot bind two socket objects to a port; a TCP socket is an
>	abstraction of a TCP port.
>
>2) you cannot listen and then do a connect on the same socket because
>	all 4.X BSD UNIX derived TCP/IP implementations specifically
>	prohibit this at the top kernel socket level.  ...

The TCP spec shows how colliding connect attempts are handled
when both ends start in the CLOSED state, not the LISTEN state.
What would happen if the collision did  not occur and only one
side attempts to connect to the other?  It seems to me that the
connecting side is trying to connect to something that is not
LISTENing, and thus the connection attempt would fail.

This is how BSD-based implementations behave in this case. I
don't know if BSD-based implementations can handle colliding
connect attempts (even though the protocol does) because each
side wants to connect to something that is listening on the
port at the same time that it might be attempting a connect on
the same port. As noted above, this is is not permitted.

>3) you MAY, however, be able to do a listen/accept on a socket
>	that is already blocked upon a connect() in progress.  this is
>	part of a "host requirement" recommendation, intended
>	primarily for busy server protocols such as SMTP (sendmail).
>	note that this is the opposite order of what you tried.
>	looking at the 4.X BSD code very quickly now, it seems
>	that this is allowed.  i have not personally tried this.
>
>-- 
>Hwa-Jin Bae (hjb@netcom.com)
>Peaceful Star, Oakland, CA		ftp.netcom.com:/pub/hj/hjb/README

I don't think this will work for us. We want to be able to
accept connections on the service port from any system at any
time. When we need service from another system, we will use any
existing connection or create a new connection.  We don't want
to shut down our LISTEN on the service port when we set up a
new connection.

It looks like we will have to set up connections using unique
port numbers and write some code to detect duplicate
connections and tear one down.

---

Bob Bradford				Email: bbradfor@glenayre.com
Glenayre Electronics                    Phone: (604) 293-1611
1570 Kootenay Street                    Fax:   (604) 293-4317
Vancouver, BC, Canada  V5K 5B8



-----------[000203][next][prev][last][first]----------------------------------------------------
Date:      08 Dec 1994 20:38:00 GMT
From:      Steinar.Haug@runit.sintef.no (Steinar Haug)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

> Please name a common, current system with UDP checksums off by default.

SunOS 4.1.3 (and 4.1.3_U1, and probably also 4.1.4. I'm going to check
tomorrow...)

Steinar Haug, SINTEF RUNIT, University of Trondheim, NORWAY
Email: Steinar.Haug@runit.sintef.no

-----------[000204][next][prev][last][first]----------------------------------------------------
Date:      Thu, 08 Dec 1994 21:13:00 CET
From:      marco@dmrt-2.dmrt.nl
To:        comp.protocols.tcp-ip
Subject:   Re: PC as nameserver?

aboba@netcom.com (Bernard Aboba) writes:

>The DIS version of KA9Q (Large model) can act as a nameserver. However, 
>be aware that it will only answer queries for locally stored records. It
>will not recurse if requested to do so; rather than querying the root servers
>for a record it doesn't have, it will return an error. 

Nor will it do zone transfers. Those use TCP instead of UDP, and it doesn't want                                                                                                                                                                                                                                                          
to know about that. It listens on the UDP port only.

>This is assuming that you don't have even  a 386/16 with 4 MB RAM; KA9Q will
>run in 1 MB, even on a 286. If you have a 386/16 wtih 4 MB, you can run
>named under Linux or FreeBSD instead. 

Yup. That's what we ended up doing.
                                   
				Marco.
--

-----------[000205][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 22:30:47 GMT
From:      cleelacj@agedwards.com (Chris Cleeland)
To:        comp.protocols.tcp-ip
Subject:   Re: DHCP Info ???

In article <Pine.ULT.3.91.941202202814.2391A-100000@chinook.halcyon.com>,
Robin Callender  <robinc@halcyon.com> wrote:
>I'm pretty sure (vis. know) that Win/NTAS has a DHCP server with it.  

Yes.

>Also I believe that HP has a DHCP server for their unix system.

News to me.  Haven't seen hide nor hair of that yet.

-cj

-- 
==============================================================================
Chris Cleeland 	       	       	|  NeXTMail:  chris@milo.st-louis.mo.us
BOS Dev. Team                   |  MIMEMail:  cleeland@agedwards.com
                                |  BellNet:    (314) 289-5372

-----------[000206][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 22:36:59 GMT
From:      cleelacj@agedwards.com (Chris Cleeland)
To:        comp.object,comp.protocols.tcp-ip,comp.unix.solaris,comp.lang.c++,comp.client-server
Subject:   Re: ACE version 2.15.5 now available

In article <1994Dec5.113000.25036@unix.brighton.ac.uk>,
Dom De Vitto <ddv@unix.brighton.ac.uk> wrote:
>: : It is not clear to me why you would 'naturally' not be responsible for
>: : any problems caused by using this library.  What if a problem within this
>: : library caused all of the Iridium satellites to become useless?  I'm not
>: : saying you have full responsibility by any means, but to claim a desire
>: : of a comprehensive and robust library without responsibility for any
>: : lack of those characteristics seems like a conflict.
 
>: Don't look a gift horse in the mouth!
 
>: If you want someone to take responsibility fork out the $$$$. The above 
>: is simply a standard disclaimer that anyone in their right mind would put 
>: with free software.
>
>Look at a comercial software package sometime - they say *exactly* the same
>thing...:(

Of course.  And you don't even get the source code so that you *might*
have a chance to fix the problems yourself.

First of all, about stuff in general: CAVEAT EMPTOR.
Secondly, with respect to free software: You Get What You Pay For.

You are always free to find another vendor or roll your own and take
on the responsibility yourself.

'nuff...
-cj


-- 
==============================================================================
Chris Cleeland 	       	       	|  NeXTMail:  chris@milo.st-louis.mo.us
BOS Dev. Team                   |  MIMEMail:  cleeland@agedwards.com
                                |  BellNet:    (314) 289-5372

-----------[000207][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 22:56:52 GMT
From:      vjs@calcite.rhyolite.com (Vernon Schryver)
To:        comp.protocols.nfs,comp.protocols.tcp-ip
Subject:   Re: NFS/SLIP

In article <UitmpzSSMUgERdWr49@transarc.com> Lyle_Seaman@transarc.com writes:

> ...
>If NFS is not about 60% as fast as FTP, you should look at twiddling
>the retransmission timeout parameters -- you may be retransmitting too
>agressively and chewing up bandwidth.

That is very a good point.  If you use the common default of 0.7 seconds
with the default packet size of 8K on a typical v.32bis+v.42+v.42bis
SLIP or PPP link, one packet will take 5 or more seconds, by which time
you will have already asked for a few retransmissions.  You can expect
the server to oblidgingly answer each retransmitted request, filling
the link for a block of 5 seconds for each one.  Well, the v.42bis
compressor will be well trained after the first couple and so the later
ones will take fewer than 5 seconds--oh joy.

I use automounter with NFS over a link that includes one PPP link, which
contributes about 140 ms to the total 300-3000 ms RTT.  The following
are the parameters I use:
    timeo=105,acregmin=30,acregmax=300,acdirmax=300,rsize=4096
As you can see, I use a long retransmission timeout.

Notice that I effectively turn off attribute cache aging.  That drastically
reduces the getattr traffic that otherwise precedes most operations.
However, if something else changes a file on a server, then I usually
get a the old version, and so I often run `find / -name foo` for 5 or
10 seconds to flush the cache.  I use a 4K packet size, having found
that to be the best compromise for me among lost IP packets and
retransmissions, rpc/xdr overhead, and queuing latency on the long link.

Again, I recommend using FTP or rcp over WAN links except when other
considers are compelling.  For example, NFS is an easy quick-and-dirty,
if sub-optimal way to shadow WWW or other archives.


Vernon Schryver    vjs@rhyolite.com

-----------[000208][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 23:06:58 GMT
From:      vjs@calcite.rhyolite.com (Vernon Schryver)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

In article <NADEL.321.008A4F50@litc.lockheed.com> NADEL@litc.lockheed.com (Ron Nadel) writes:

>This is true, Vernon, but a large MTU over SLIP is not recommended in the 
>event of retransmits.  It is better to retransmit a small packet or two then a 
>couple huge ones.  
>
>Bruce, if you can tune your MTU, try it at 512 or less even (depends on the 
>line speed and error rates.

I strongly disagree.  Reducing the SLIP MTU does not affect the costs
of retranmissions.  Any bit errors in any of the IP datagrams comprising
an NFS UDP/IP packet requires the retransmission of the entire packet.
(Of course, assuming NFS/UDP instead of NFS/TCP.)  Thus, reducing the
MTU does not help NFS.  On the other hand, it hurts NFS because it
doubles the number of 20 byte IP headers.  Reducing the rsize/wsize to
512 would reduce the cost of a retransmission, but it would increase
the cost of the xdr/rcp/udp/IP headers to more than 20%.

If your SLIP link has a high enough error rate to cause a noticable
retransmission rate, then you really do not want to be using NFS over
it.  For that matter, you really do not want to doing any file transfers
using common protocols over it.

As with FTP, rcp or any other common file transfer scheme, you want the
largest practical MTU.  The reasons smaller MTU's are recommend for SLIP
and PPP links concern latencies for interactive traffic sharing the link
with file transfers.


Vernon Schryver    vjs@rhyolite.com

-----------[000209][next][prev][last][first]----------------------------------------------------
Date:      Thu, 8 Dec 1994 23:09:04 GMT
From:      vjs@calcite.rhyolite.com (Vernon Schryver)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

In article <STEINAR.HAUG.94Dec8213801@runit.sintef.no> Steinar.Haug@runit.sintef.no (Steinar Haug) writes:
>> Please name a common, current system with UDP checksums off by default.
>
>SunOS 4.1.3 (and 4.1.3_U1, and probably also 4.1.4. I'm going to check
>tomorrow...)

Those are common, but please note I also wrote "current."  
Isn't some flavor of Solaris the current Sun Microsystems system?

To paraphrase what someone to me not long ago, everyone's old products
are pigs.


Vernon Schryver    vjs@rhyolite.com

-----------[000210][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 00:02:39 GMT
From:      briant@gateway.bsis.com (Brian Toole)
To:        comp.infosystems.www.users,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.infosystems.www.providers,comp.os.ms-windows.networking.tcp-ip,comp.os.ms-windows.networking.misc
Subject:   Re: WWW Server/Client without INTERNET??

You have to have a "NULL" WinSOCK DLL for this to work.
Read the faq that is included with the Windows version
of Mosaic and it'll tell you where to find a copy on the
net.

 
-------------------------------------------------------------------
Brian Toole                            Broadway & Seymour, Inc.
Vice President                         5000 Aerial Center Parkway
919.319.6500                           Morrisville, NC 27511
CI$ 73321,1120                         briant@gateway.bsis.com
-------------------------------------------------------------------

-----------[000211][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 14:01:40 -0800
From:      shunt@eis.calstate.edu (Steve J. Hunt)
To:        comp.protocols.tcp-ip
Subject:   Re: What is CSLIP?

chiejin@nfx.com (Chie-Jin Cheng) writes:
> Hi there,
> 
> What is CSLIP?  What does it differ from SLIP?
> 

CSLIP is compressed SLIP.  Much better throughput.  PPP includes 
compression by default, for SLIP it was a feature added later to many 
implementations.



-----------[000212][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 14:17:45 -0800
From:      dag@ossi.com (Darren Alex Griffiths)
To:        comp.protocols.tcp-ip,comp.unix.solaris,comp.unix.sys5.r4
Subject:   Re: TLI t_snd flow restriction

whalenm@ulabsgi (Matthew Whalen) writes:

>I'm writing an application that listens and sends data on a TLI connection
>endpoint simultanously.  I have one (solaris) thread that listens
>on the endpoint for data to arrive using t_rcv, while another
>thread wants to send data on the endpoint using t_snd.  It seems
>that the t_snd gets blocked by a "flow control restriction" on the
>endpoint, and so both frequently sit there blocked.  Short of making
>the endpoint asynchronous, is there a way get around this flow control
>restriction?  It seems silly to resort to asynchronous communication
>on a TLI endpoint when I have a threads library.  Any ideas?  Thanks.


Instead of having one thread block on the t_rcv you should have it block
on a poll().  The poll() will not block the t_snd so the sending thread will
work fine.  When the poll exits you should check the revents flag to see if
there is any data waiting, if there is then you can do a t_rcv without 
blocking.  You should also check revents to see if the connection was closed,
if it is still up then you can go back and do the poll again.

  Cheers,
   --dag

_____________________________________
Alex Griffiths 
dag@ossi.com
Senior Software Engineer
Fujitsu Open Systems Solutions, Inc.
408-456-7815
_____________________________________


-----------[000213][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 09:27:41 -0500
From:      dmr@dsinc.com (David M. Ramsey)
To:        comp.protocols.tcp-ip
Subject:   only 1 machine on LAN on Internet?

Given only one valid Internet IP number, how can things be configured
such that only one machine on a LAN is "on" (SLIP/PPP) the Internet?

I'd like to arrange things such that other machines on the LAN can
telnet to the machine connected to the Internet, and go 'out' from there.

I seem to recall reading somewhere (where?) that there is a special
block of IP numbers reserved for isolated machines that are not connected
to the internet.  If so, could I simply use those addresses for the other
machines and turn off IP forwarding on the one connected machine?

Thanks for any help/advice/pointers/info.

David
-----------------------------------------------------
David M. Ramsey           Office Voice: (704)847-8904
Decision Support Inc.     Office Fax  : (704)847-4875
Internet: dmr@dsinc.com   Home Voice  : (704)521-8265
-----------------------------------------------------

-----------[000214][next][prev][last][first]----------------------------------------------------
Date:      Fri, 9 Dec 1994 03:01:51 GMT
From:      earth@nanguo.cstpl.com.au (Bob Chalmers)
To:        comp.protocols.tcp-ip
Subject:   LANTASTIC 6 + email to Internet


Does any one out there have a solution for getting Lantastic 6 and
Lantastic TCP/IP, to exchange email with the Internet through a
local Unix Internet provider gateway.?

I can't believe that it's not possible.!!!  

and Novell if you know how.

cheers,
Bob
-- 
...........................  ................  .............................
|  Robert Chalmers.         ,-._|\*<-Mackay    robert@nanguo.cstpl.com.au  |
|   cstpl.com.au.          /      \ FAX: +61 79 524032  PH: +61 79 524395  |
| International Access.    \_.-\__/  *Info,	        info@cstpl.com.au  |
| AARNet Value Added Reseller.	 v   *Sales,  	        sales@cstpl.com.au |
|---MIME CAPABLE------* ao zhong you yi wan sui *--------------------------|

-----------[000215][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 03:21:59 GMT
From:      fmedina@bu.edu (Fernando Medina)
To:        comp.protocols.tcp-ip
Subject:   AS/400 with TCP/IP

We are experimenting with TCP/IP on an AS/400, as APPN has been very unstable in our WAN.  The problem we are facing is with Printer sessions.  Anyone have any experience using Printer sessions over TCP/IP on an AS/400??  Wall Data's RUMBA won't support printer sessions with TCP/IP and we're not sure if it can be done with Chameleon 4.0, any help would be appreciated..

thanks,

Fernando Medina
MIS
Dole Fresh Fruit Int'l, Ltd.


-----------[000216][next][prev][last][first]----------------------------------------------------
Date:      Fri, 09 Dec 94 12:08:05 PST
From:      cat@olddaisy.ee.und.ac.za (Abdul Rehman Gani)
To:        comp.protocols.tcp-ip
Subject:   TCP/IP info - where...

Where can I get info on TCP/IP and protocols used (PPP/SLIP/FTP etc)?

Mailed replies appreciated (cat@olddaisy.ee.und.ac.za), but posting here 
okay

Thanks

Abdul Rehman Gani


-----------[000217][next][prev][last][first]----------------------------------------------------
Date:      Fri, 09 Dec 94 11:25:00 PDT
From:      MIKEC@BIMUNCIE.BALL.COM
To:        comp.protocols.tcp-ip
Subject:   TCP books


Is anyone familiar with the books re: TCP/IP authored by Douglas Comer?  I am 
looking for more info - ie: publisher, opinions, etc.

-----------[000218][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 04:58:45 GMT
From:      dries@elvis.rowan.edu (PowerSlave)
To:        comp.os.ms-windows.apps.comm,comp.os.ms-windows.networking.tcp-ip,comp.protocols.tcp-ip,alt.dcom.slip-emulators
Subject:   Re: Twinsock 1.2 on IRIX 5.2

dries@elvis.rowan.edu (PowerSlave) writes:
>dries@elvis.rowan.edu (PowerSlave) writes:
>>I've tried the following under OS/2 v2.10 WinOS2, OS/2 v2.11 WinOS2, and
>>Windows for Workgroups v3.11, I get the same result in all
>>circumstances:
 
>>Twinsock starts `normally', without any errors or incident. I fire up
>>Netscape 0.94, and on the SunOS built tshost, everything works fine.
>>Under the IRIX 5.2 built tshost, I always get a `Twinsock caused a GPF
>>in WINSOCK.DLL at address 0001:14AC' .. the address almost never
>>changes, although I have gotten a 0002:xxxx once.

Using WSATest I have narrowed the WinSock API calls down to
gethostbyname and gethostbyaddr. These two calls will cause a GPF in the
calling program (in this case WSATest)... All the other API calls I was
able to test worked fine.

These two calls returned the proper values running against tshost
compiled under SunOS. The tshost compiled under IRIX 5.2 causes a GPF in
WSATest, so I cannot get a log of what was returned, if anything.

If someone knows where I should go from here, please feel free to
contact me so I can continue to debug this, I'd love to get tshost
working under IRIX. 

I did compare the man pages between SunOS and IRIX, and the functions
are described similarly, so the invocation in `commands.c' shouldn't be
causing any fuss.. but something is.

-----------[000219][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 14:11:22 -0500
From:      donham@iii1.iii.net (Perry Donham)
To:        comp.protocols.tcp-ip
Subject:   Re: only 1 machine on LAN on Internet?

RFC 1597 describes private network numbers.

One block of addresses reserved for private networks is 192.168.0.0.
What you can do is assign addresses from this block to the machines 
on the LAN (including the SLIP machine), and *not* set up a route
through the SLIP interface..

Perry

David M. Ramsey (dmr@dsinc.com) wrote:
: Given only one valid Internet IP number, how can things be configured
: such that only one machine on a LAN is "on" (SLIP/PPP) the Internet?
 
: I'd like to arrange things such that other machines on the LAN can
: telnet to the machine connected to the Internet, and go 'out' from there.
 
: I seem to recall reading somewhere (where?) that there is a special
: block of IP numbers reserved for isolated machines that are not connected
: to the internet.  If so, could I simply use those addresses for the other
: machines and turn off IP forwarding on the one connected machine?
 
: Thanks for any help/advice/pointers/info.
 
: David
: -----------------------------------------------------
: David M. Ramsey           Office Voice: (704)847-8904
: Decision Support Inc.     Office Fax  : (704)847-4875
: Internet: dmr@dsinc.com   Home Voice  : (704)521-8265
: -----------------------------------------------------

-----------[000220][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 06:30:54 GMT
From:      jedi@susu.hnc.re.kr (Joohyee Lee)
To:        comp.unix.solaris,comp.protocols.tcp-ip
Subject:   Re: TCP/IP and IPX/SPX routing

Vincent Smeets (smeets@cci.de) wrote:
: Hallo,
 
: Is it possible to do TCP/IP and IPX/SPX routing with one host?
 
: Our problem is: 
: We have now one physical network with one class C IP network (for Unix
: and PC's) and a Novell server (for the PC's).  As we run out of IP
: numbers, we orderd an other class C IP network number.
: Now, we want to split out physical network in two parts and assign
: every part an IP number.
: Then we want to connect these two network with a router. The only
: problem is that this router must pass the IPX/SPX packages too.
 
: We were thinging about a Sun with Solaris, but as far as I know,
: doesn't is foward the IPX/SPX packages. Is there some software for the
: Sun to make it an IPX/SPX forwarder, or must I use some other hardware
: (which)?
 
: Please give me some hints
: Vincent


: ---
: --      Vincent Smeets                  Competence Center Informatik GmbH
: --      Tel. : +49-5931-805461          Postfach 1225
: --      EMail: smeets@cci.de            49702 Meppen, Germany

Hi, Vincent.
Basically, Novell NetWare can route IP packet very well.
First, Do you want to add another Novell NetWare server to your network?
Second, Do you want to construct a WAN? for example, your second network
will be far from first net?.
If nothing applies to you, Its very simple.
1. Add another LAN card to your Novell server and assign new IP address.
2. load TCPIP.NLM with forward=yes option.
3. Bind IP to that board with assigned IP address and proper Netmask.
4. Insert assinged IP adress to your UNIX machine's routing table as a 
   gateway to your new network.
5. Enjoy your new network.

If you have a problem, please mail me again. 

	jedi@hnc.re.kr
	Joohyee Lee.


-----------[000221][next][prev][last][first]----------------------------------------------------
Date:      Fri, 9 Dec 1994 07:25:48 GMT
From:      ericma@netcom.com (Eric S. Ma)
To:        comp.protocols.tcp-ip,comp.os.linux.help,comp.sys.sun.admin
Subject:   Linux Slip 0.7.5 and SunOS4.1.3's cslip-2.7 compatibility

Does anyone know if they are compatible?  I dialed up successfully from a
Linux box to a Sun sparc (through a terminal server) but the SLIP connection
is bad.  Pinging from SUN to Linux always results in the error "ICMP: Bad
checksum" on both machines (though Linux apparently gets an ICMP packet with
a bad checksum) and then a timeout on Sun.  I don't see response on Sun by
pinging from Linux to SUN.  However, the connection seems to allow a few
session-beginning telnet or ftp packets to go through occasionally --
followed by resending of telnet or ftp packets, as reviewed by tcpdump
(netstat shows active connections; I sometimes saw "Connected to <hostname>"
with ftp/telnet from Sun to Linux -- but the ftp or telnet session times out
after a while and never went past the initial "Connected to <host>" string.
I believe the MTUs are set to the same values on both machines.  Would
anyone please share his/her experience with setting up (C)SLIP between a
Linux box and a sparc running SunOS 4.1.3_U1?  If so, what software (public
toolsdomain, if possible) is used?  What tools are available for testing
initial SLIP setup?  Is there a HOWTO on SLIP that describes frequently
encountered problems during initial setup?  Thanks in advance for help.

Eric
ericma@netcom.com

-----------[000222][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 08:22:28 GMT
From:      trall@trall.almaden.ibm.com (Tony Rall)
To:        comp.protocols.tcp-ip
Subject:   Re: Multiple NICs on a server on same segment

There are several techniques to get the clients to send to a
distribution of addresses on the server, even though they are all on
the same subnet.  I know nothing about Suns, but normal IP routing is
likely to choose just one of its own addresses for sending the
responses.  Probably not what you want.

You might be able to fake the Sun's routing decisions with a phony
subnet mask.  Let's say your real subnet mask is 255.255.255.0.
You define each of the interfaces on the Sun server with a mask
of 255.255.255.224.  Then make an effort to get the client host
addresses distributed evenly:

25% in the range .33 - .62
25% in the range .65 - .94
25% in the range .97 - .126
25% in the range .129 - .158

(This assumes you're using a class C network.  If it's class A or B
you can get about 62 hosts in each range instead of 30.)  The Sun
would have to use one address in each range. 

-- 
Tony Rall    trall@almaden.ibm.com

-----------[000223][next][prev][last][first]----------------------------------------------------
Date:      Fri, 9 Dec 1994 08:32:38 GMT
From:      rogersm@BBS-CE.UAB.ES (i Font)
To:        comp.protocols.tcp-ip
Subject:   Zmodem

 Hi all!!

 How is Zmodem transmitted over a TCP/IP connexion. We want to do it, so
our telephone connected users could download their data over the phone line.
What port is used? Where can I find the info?

Regards,
Roger - rogersm@bbs-ce.uab.es

PS: Please, reply by mail, thanks!!


-----------[000224][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 09:10:37 GMT
From:      chou@pds.nchu.edu.tw (Chou Chaw Yi)
To:        comp.protocols.tcp-ip
Subject:   Q:Install MULTICAST Kernel IN Ultrix4.4

Dear Netows:
   I was in Ultrix 4.4(Rev.69) compiler MULTICAST KERNEL appeared the
follwing message.:


Too many command lines for `igmp.o'
Too many command lines for `ip_mroute.o'
/bin/rm -f vmunix
/bin/sh ../../conf/mips/newvers.sh
awk: newline in string near line 9
awk: syntax error near line 10
awk: illegal statement near line 10
?
cc -EL -I. -c -G 8 -O2 -g3 -I. -I.. -DNEWS -DDS5000 -DUWS -DSYS_TPATH -DAUDIT -DPACKETFILTER -DCDFS -DLAT -DNETMAN -DUFS -DDLI -DRPC -DNFS -DMROUTING -DMULTICAST -DINET -DQUOTA -DMIPS -DKERNEL  -c vers.c
loading vmunix
/bin/ld:
igmp.o: init_igmp: multiply defined
igmp.o: accept_igmp: multiply defined
igmp.o: send_igmp: multiply defined
ip_mroute.o: ip_mrouter: multiply defined
ip_mroute.o: ip_mrtproto: multiply defined
ip_mroute.o: mrtdebug: multiply defined
ip_mroute.o: ip_mrouter_cmd: multiply defined
ip_mroute.o: ip_mrouter_done: multiply defined
ip_mroute.o: log: multiply defined
ip_mroute.o: ip_mforward: multiply defined
Undefined:
rtfree
uvifs
numvifs
s1
s2
k_update_route
k_del_route
ntohl
malloc
inet_fmt
inet_fmts
inet_valid_subnet
k_add_route
find_vif
update_neighbor
fprintf
errno
k_hdr_include
k_set_rcvbuf
k_set_ttl
k_set_loop
k_init_dvmrp
htonl
accept_group_report
inet_cksum
k_set_if
check_vif_state
rtinit
igmp_joingroup
igmp_leavegroup
rtalloc
rtrequest
igmp_input
igmp_init
igmp_fasttimo
rtredirect
rtioctl
mayfairsw
ffoxsw
pmaxsw
ds5400sw
ds5000sw
ds3minsw
dsMaxinesw
dskn03sw
ln_softc
lninfo
nLNNRCV
nLNNXMT
nLNMULTI
*** Error code 1

Stop.

-----------[000225][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 19:00:17 -0500
From:      barmar@nic.near.net (Barry Margolin)
To:        comp.protocols.tcp-ip
Subject:   Re: Fkey characters getting split on frame relay

In article <3c55ej$44l@wsrcc.com> wolfgang@wsrcc.com (Wolfgang Rupprecht) writes:
>Why not have your app parse the key sequence the same way no matter if
>it arrived in 50ms or 50 seconds?

This makes it difficult to assign a separate function to the ESC key, or
even generate an error message if the user inadvertently hits it.  Instead,
the application will seem to hang, as it awaits the remainder of the escape
sequence that a function key would send.
-- 

Barry Margolin
BBN Internet Services Corp.
barmar@near.net

-----------[000226][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 13:14:31 GMT
From:      dtheese@motown.ge.com (David C. Theese)
To:        comp.protocols.tcp-ip
Subject:   Need UDP protocol help for Solaris 2.3


Hello everyone,

   I am trying to use the UDP protocol under Solaris 2.3 on Suns.  The problem I am
having is that I cannot get the IO to be interrupt based - the program always hangs
on the recv call and eats CPU cycles while waiting for a message to come in.  I have
successfully written interrupt driven code under 4.1.3 by making the following two
fctl calls:

RVAL:=FCNTL(SOCKET_NUM, F_SETOWN, GETPID);
RVAL:=FCNTL(SOCKET_NUM, F_SETFL, O_SYNC);

Please ignore the strange syntax - I am developing in Ada and I have written
interfaces to the system calls.

The problem I am having is with the first call (F_SETOWN) - it returns a -1.
Answerbook says that F_SETOWN is no longer supported, then in another section,
it gives sample code that uses this option!

Does anybody out there know how to properly write interrupt driven code under
Solaris 2.3?  I would greatly appreciate any info. in this area as my development
has been held up for weeks trying to get by this.  Thanks!

Dave Theese


-----------[000227][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 22:34:25 -0500
From:      lin@cs.purdue.edu (John Chueng-Hsien Lin)
To:        comp.protocols.tcp-ip
Subject:   To find a USENIX paper about delivering NetNews using multicasting

I am interested in reading a 1994 Winter USENIX paper about using
IP multicasting to deliver Net News.  (Sorry, I don't know the
paper's title and name(s) of the author(s)) 
Is the paper on-line?

Thanks,
John Lin

-----------[000228][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 14:17:32 GMT
From:      jake@venus.record.unipd.it (Luca Polo)
To:        comp.unix.solaris,comp.sys.sun.admin,comp.protocols.tcp-ip
Subject:   Re: Under SunOS 5.3, where to set broadcast address and netmask ??

In article <CROTEN.94Dec2155624@eosqa.gsfc.nasa.gov>,
Charles Roten <croten@eosqa.gsfc.nasa.gov> wrote:
[....]
>But Stunned MicrobeSystems doesn't _use_ straight Sys V R4.  So _where_, 
>for Pete's sake, do I set the netmask ?  And _HOW_ ??  I've fgrepped the 

/etc/netmasks, which is a link to ./inet/netmasks.
The format is:
<network address>	<netmask>

E.g.:
147.162.0.0     255.255.255.0  # my netmask entry

Oh, I forgot to mention "netmasks" NIS/NIS+ table, too... 

Regards,
Luca Polo.
-- 
+-----------------------------------------++---------------------------+
| Luca Polo   : jake@gest.unipd.it        || System administrator,     |
| (finger jake@berica.gest.unipd.it for   || Ist. di Ing. Gestionale,  |
| address and phone numbers)              || Universita` di Padova.    |

-----------[000229][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 14:20:00 -0000
From:      cmattern@mindspring.com (Chuck Mattern)
To:        comp.protocols.tcp-ip
Subject:   Thin ethernet cabling

I realize this is a protocols group but do not know where else to
turn...  

I am wiring my house with c-net and would like to place the tap behind
a wall plate and then run a short drop to the NIC rather than
connecting the tap directly to the nic.  This would allow me to have
only one cable coming from the wall and lessen the clutter.  Does
anyone know if such a topology is legal/functional?  Or where else I
might inquire?


-- 
----------------------------------------------------------------------
|Chuck Mattern              | "Not failure, but low aim, is crime."  |
|cmattern@mindspring.com    | -James Russell Lowell-                 |
----------------------------------------------------------------------

-----------[000230][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 14:21:06 GMT
From:      Gordon McAndrew <gordon@aec.env.gov.ab.ca>
To:        comp.infosystems.www.users,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.infosystems.www.providers,comp.os.ms-windows.networking.tcp-ip,comp.os.ms-windows.networking.misc
Subject:   Re: WWW Server/Client without INTERNET??

briant@gateway.bsis.com (Brian Toole) wrote:
>
> You have to have a "NULL" WinSOCK DLL for this to work.
> Read the faq that is included with the Windows version
> of Mosaic and it'll tell you where to find a copy on the
> net.
 
> Brian Toole                            Broadway & Seymour, Inc.

Hi Brian.  I believe you missed some of the messages on this thread.

The original poster wanted to supply WWW services to the other PCs on the 
same LAN.

Therefore he needs a complete setup with a WWW server and WWW clients.

Gordon McAndrew
Vegreville, Alberta
gordon@aec.env.gov.ab.ca


-----------[000231][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 14:37:14 GMT
From:      jake@venus.record.unipd.it (Luca Polo)
To:        comp.unix.solaris,comp.sys.sun.admin,comp.protocols.tcp-ip
Subject:   Re: OK .. now, how to get the SPARCserver 1000 to see it's router ??

In article <CROTEN.94Dec5144747@eosqa.gsfc.nasa.gov>,
Charles Roten <croten@eosqa.gsfc.nasa.gov> wrote:
>In article <CROTEN.94Dec2155624@eosqa.gsfc.nasa.gov> croten@eosqa.gsfc.nasa.gov (Charles Roten) writes:
>
>Check this out ... from another SunOS 5.3 host ...
 ...
>    128.183.0.0     255.255.255.0
>    ^^^^^^^^^^^
>Then .. 
>    
>    cddev1.gsfc.nasa.gov% /sbin/ifconfig ie0
>    ie0: flags=863<UP,BROADCAST,NOTRAILERS,RUNNING,MULTICAST> mtu 1500
>            inet 128.183.117.37 netmask ffffff00 broadcast 128.183.117.255
>                                                           ^^^^^^^^^^^^^^^
>Wierd ...  
>
Why ? In /etc/netmasks you basically say that your class B network is
subnetted to a Net.Subnet.Host form: "broadcast" address refers only to the
Host part, hence the address above: if you want a different broadcast address
for some special reason, then you must edit your rc files and do it by hand;
but in 99% of cases, the standard approach works well (unless one needs
broadcasts to spawn over the whole network...).

[ story of missing router deleted ]
>Can someone enlighten me about this?  What in chaos is going on?  The system's 
>IP number and name are in /etc/hosts.  /etc/nsswitch.conf is set up for DNS.  
>What have I overlooked?  Thanks.  

Maybe /etc/defaultrouter does not exists or does not contains a valid entry.
Put router's address or name in it (and in /etc/hosts).
Also note that when routing informations are set up during boot process
(/etc/rc2.d/S69inet), in.named is not yet running (this is important if
your host acts as DNS server); that's why I prefer to put "files" before
"dns" in /etc/nsswitch.conf (I also have some trouble with remote license
server if I don't use this setup. Finally, it's faster during startup).

Regards,
Luca Polo.
-- 
+-----------------------------------------++---------------------------+
| Luca Polo   : jake@gest.unipd.it        || System administrator,     |
| (finger jake@berica.gest.unipd.it for   || Ist. di Ing. Gestionale,  |
| address and phone numbers)              || Universita` di Padova.    |

-----------[000232][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 16:47:45 GMT
From:      pete@louie.timeplex.com (Pete Flugstad)
To:        comp.protocols.tcp-ip
Subject:   Re: bytes received from ping

In article 91o@news.xs4all.nl, kaat@xs1.xs4all.nl (kaat) writes:
>Noah_Davids@vos.stratus.com wrote:
>
>: I've noticed that the local copy of ping reports X+8 bytes received where
>: X is the number of bytes that I requested it send. As far as I can tell the
>: extra 8 bytes represent the ICMP header. I've looked at several systems
>: and Unix Network Programming by R Stevens and they all report the same +8
>: bytes.
 
>: Is there a reason I am missing for including the ICMP header length in the
>: bytes received count? If it includes the ICMP header why not the IP header?
>
>I'll really like to know this too!
>Currently I'm doing a project for school conserning TCP/IP.
>One of the major questions I had was indeed why the extra 8 bytes are
>there?
>
>Is there anyone who can help us here?
>
>thanks in advance,
>
>Rob.

I suspect that what is happening is that the destination machine receiving
the ping (ICMP ECHO REQUEST) is simply stuffing the entire received data
part of the message into the ECHO REPLY.  Technically, I think it's only 
supposed to pass the optional data part of the ECHO REQUEST.

To find out for sure, take a look at the sources for Ping, and the ICMP
implementaion of a unix that shows this behavior.  There are _many_ free
versions of unix around, virtually all of which have IP stacks, and have
ping utilities. 

Pete

---
Peter J. Flugstad              Ascom Timeplex
pete@louie.timeplex.com        1807 Park 270 Drive, Suite 350
+1 314 579 6520                St. Louis, MO 63146  USA
#include <std/disclaimer>      #include <funny/saying>



-----------[000233][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 17:01:00 GMT
From:      pete@louie.timeplex.com (Pete Flugstad)
To:        comp.protocols.tcp-ip
Subject:   Re: IP- forwarding

In article 1024@zh014.ubs.ubs.ch, zhebu@svusenet.ubs.ch writes:
>: 	bar> telnet foo 2000 	
>: 
>:         and foo:2000 forwards the TCP packet to an given server e.g. at
>:         myserver.de port 23
>
>It is definitely not a good idea, but:
>
>Yes, it is called telnet ;-). If you are in Unix, put an entry "myforward" 
>in /etc/services and then put a an entry telnet telnet myserver.de 23 in 
>your /etc/inetd.conf file:

I don't think this will work in the general case.  Doesn't Telnet implement
a protocol on top of TCP for Terminal stuff?  Wouldn't this possibly mess
with binary data that's going through?

I wrote a program a while ago, that _any_ user can start on any TCP port 
(specified on the command line), and it will relay any data on any received 
connection to another machine/port (also specified on the command line).  I
use it when I go to conferences because our company doesn't allow incoming
TCP sessions to low ports  So I start this program on some high port, and tell 
it to connect to the local host on port 21 (telnet).  So I can telnet to the
high port (which the firewall router allows), and it'll connect me to telnet.

Contact me if you want the source.

Pete
---
Peter J. Flugstad              Ascom Timeplex
pete@louie.timeplex.com        1807 Park 270 Drive, Suite 350
+1 314 579 6520                St. Louis, MO 63146  USA
#include <std/disclaimer>      #include <funny/saying>



-----------[000234][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 17:03:15 GMT
From:      pete@louie.timeplex.com (Pete Flugstad)
To:        comp.protocols.tcp-ip
Subject:   Re: local address reuse

In article uq@nyheter.chalmers.se, thed@mtek.chalmers.se (Thed Lyngmo) writes:
>Robert Smith (robert@banana.dis.fedex.com) wrote:
>
>: I have a problem with a simple client-server set of programs that are
>: being used to test the behaviour of the system when the server side	
>: is killed and then needs to be restarted. The problem in general is	
>: that the local address stays in a TIME-WAIT condition when the server	
>: side is killed. The client side detects connection is broken and exits	
>: normally. When I try to restart the server, I get error: bind()
>: (125Address already in use). I am unable to restart the server for	
>: 4 minutes because of this condition. I have tried using
>: setsockopt(SO_LINGER with l_linger=0) and setsockopt(SO_REUSEADDR) in
>: both the server and the client as well as no setsockopt calls.

I know this sounds stupid, but make sure you are setting SO_REUSEADDR
on the socket _before_ you bind it to the port.  Doing so after the
bind does not work.  I was bitten by this for a long time until 
someone suggested it to me.  Solved the problem in a snap.

Pete
---
Peter J. Flugstad              Ascom Timeplex
pete@louie.timeplex.com        1807 Park 270 Drive, Suite 350
+1 314 579 6520                St. Louis, MO 63146  USA
#include <std/disclaimer>      #include <funny/saying>



-----------[000235][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 17:11:32 GMT
From:      pete@louie.timeplex.com (Pete Flugstad)
To:        comp.protocols.tcp-ip
Subject:   Where is ARP implemented?

I'm not sure if this is the right newsgroup for this or not.

I'm wondering where the ARP function is implemented most of this
time.  Is it implemented as part of the IP layer (on the bottom
side), or is it part of the Network Interface Layer (and hence
can use network specific methods to resolution)?  I'm inclined to
think it's part of the NI layer?

Pete
---
Peter J. Flugstad              Ascom Timeplex
pete@louie.timeplex.com        1807 Park 270 Drive, Suite 350
+1 314 579 6520                St. Louis, MO 63146  USA
#include <std/disclaimer>      #include <funny/saying>



-----------[000236][next][prev][last][first]----------------------------------------------------
Date:      Fri, 9 Dec 1994 18:03:05 GMT
From:      jim@reptiles.org (Jim Mercer)
To:        comp.protocols.tcp-ip
Subject:   Re: AS/400 with TCP/IP

In article <3c8ign$n9s@news.bu.edu>, Fernando Medina <fmedina@bu.edu> wrote:
>We are experimenting with TCP/IP on an AS/400, as APPN has been very
>unstable in our WAN.  The problem we are facing is with Printer sessions.
>Anyone have any experience using Printer sessions over TCP/IP on an AS/400??
>Wall Data's RUMBA won't support printer sessions with TCP/IP and we're not
>sure if it can be done with Chameleon 4.0, any help would be appreciated..

you might want to look at the LPD/LPR support of the AS/400 TCP/IP.

this allows the AS/400 to advertise it's print queues to the TCP/IP
network.  your TCP/IP for DOS/Windows should support printing to LPD
printers.

also, you can set up printers on your network, and put LPD servers in 
front of them, then the AS/400 can print on them.

some printers (HP4si) have LPD servers built in.  you can also get
DOS/Windows programs that will LPD advertise locally attached printers.

you can also get LPD/LPR support for novell servers which makes
almost all of your printers available to anyone on your network.

there should be suitable security facilities to limit who can actually
print where.



-- 
[ Jim Mercer                 jim@reptiles.org              +1 416 506-0654 ]
[          Reptilian Research -- Longer Life through Colder Blood          ]
[   Never, ever forget to replace the toiletseat after use!!! A wet        ]
[   chinchilla is a very funny and pathetic sight.   -- alt.chinchilla     ]

-----------[000237][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 1994 18:06:04 GMT
From:      anewman@ernie.eecs.uic.edu (Aaron Newman)
To:        comp.unix.solaris,comp.protocols.tcp-ip
Subject:   Help! binding socket to address with bind dies with already..

Hello.

I am developing a  small TCP-IP C/S app and when I try to bind()
before listen()ing on a port, it works fine the first time.  But after I
exit the app and restart, Bind dies with 'address in use' error code.
This goes on for ~10 minutes, and then it works again.  I am closing
the socket down after the app exit with 'close()'.  Is this the wrong
way to close it?  The man pages are no help. (surprise)

This is on sparc Solaris 2.3

What stupid thing am I (not) doing?

Thanks for saving me (again),
Aaron

--
--------------------------------------------------------------------
|  anewman@eecs.uic.edu             | It's better to fade away than |
|  http://reagan.eecs.uic.edu/      | to shoot yourself in the      |
|  http://www.eecs.uic.edu/         | head if you ask me.           |
|        ~anewman/Welcome.html      | -me                           |
|                                   |                               |
--------------------------------------------------------------------

-----------[000238][next][prev][last][first]----------------------------------------------------
Date:      Fri, 9 Dec 1994 18:31:36 GMT
From:      nagendra@bu.edu
To:        comp.protocols.tcp-ip
Subject:   need help decoding tcp info

I'm running os2 with an app which consists of three executables running on the 
same machine.  Each app sends a ping type packet on all of it's active tcp 
connections.  After a while, they don't seem to be able to communicate with each
other because the ping packets don't get across!!!

So I ran netstat to gather the tcp information.. Now I need help deciphering
the output.  I ran netstat every few minutes and did a diff to see what has changed.
The following is a listing of the changes.. I don't really know what they all stand
for and so I'm not sure where to go to next.. any help appreciated..

the options for netstat are:

[C:\TMP]netstat
Usage: netstat [ -? ] | [ -mtuisprcna ]

Where:
m - mbufs
t - tcp
u - udp
i - ip
s - sockets
r - routes
c - icmp
n - interfaces
a - address
p - arp
? - help

this is the diff in output to netstat -t: 


7,9c7,9
< segs where we tried to get rtt    3631
< times we succeeded                3621
< delayed acks sent                 1516
---
> segs where we tried to get rtt    3841
> times we succeeded                3831
> delayed acks sent                 1718
16,18c16,18
< total packets sent                6424
< data packets sent                 2024
< data bytes sent                   97577
---
> total packets sent                6836
> data packets sent                 2234
> data bytes sent                   102407
21c21
< ack-only packets sent             3162
---
> ack-only packets sent             3364
27,29c27,29
< total packets received            6433
< packets received in sequence      2431
< bytes received in sequence        97549
---
> total packets received            6845
> packets received in sequence      2641
> bytes received in sequence        102379
36c36
< dup. bytes in part-dup. packets   5740
---
> dup. bytes in part-dup. packets   6143
45,46c45,46
< rcvd ack packets                  3206
< bytes acked by rcvd acks          98780
---
> rcvd ack packets                  3416
> bytes acked by rcvd acks          103610



-----------[000239][next][prev][last][first]----------------------------------------------------
Date:      Fri, 9 Dec 1994 19:22:48 GMT
From:      "Steve Bult" <bultx001@maroon.tc.umn.edu>
To:        comp.protocols.tcp-ip, mcknight@rbdc.rbdc.com
Subject:   Re: Multiple NICs on a server on same segment

I don't have any reference material for you, only an opinion. I think this 
will work but it might not do what you want. You will probably have to 
hardcode each user to a specific destination address ie you will not be 
able to create a single giant pipe that gets dynamically allocated. You 
will get four discrete connections that need to be divied up among the 
users. If you want a single large pipe, you might want to consider ATM but 
that will raise a different set of issues such as lan emulation...Steve Bult



On 3 Dec 1994 17:57:13 -0500, 
Michael Mcknight  <mcknight@rbdc.rbdc.com> wrote:

>
>We have the need to connect a Sun server to a very busy network.  The idea
>so far has been to connect the users and servers, etc. to an etherswitch
>to manage the traffic better.  Here's the idea.
>
>We want to have a wide pipeline going from the switch to the server... say
>4 30MB pipes.  We want the server to logically be on 1 (one, uno) segment.
>I wont go into all of the reasons why, but my question is this.. can we?
>
>For example:
>
>      |--------|               
>      |        |---------------|==========|     (no need to draw users)
>      | SPARC  |---------------| EtherNet |-------------^
>      | SERVER |---------------|  Switch  |-----------------^
>      |        |---------------|==========|
>      |--------|      ^
>                      |-- Cat-5 UTP 10-Base-T Ethernet Cables
>
>Lets say the SPARC has 4 10BT ethernet cards numbered and named as follows:
>	le0		sun1a	155.125.10.1
>	le1		sun1b	155.125.10.2
>	le2		sun1c	155.125.10.3
>	le3		sun1d	155.125.10.4
>
>All NICs would be on the same logical & physical network.
>
>Is this a problem?  We invision the clients simply connecting to a different
>card depending on their use/location.  This is different than segmentation 
>in that we are not really segmenting the network.
>
>It seems like the sun would simply see each NIC as a different card and
>not really care about what network it is on.  The users would simply conect
>to a specified card without having to really care where it is.
>
>I've been told this can be done and that it cannot be done.  I can see it
>as a problem if we were trying to route between cards, but in this case,
>we are not routing between them.
>
>If it cannot work, why?  Is it a limit of TCP/IP or the SPARC system or the
>SunOS?
>
>As much detail as possible (or a refence document) would be greatly
>appreciated.
>
>Please Email all replies.
>Thanks to all in advance.
> 
>-Michael McKnight
> mcknight@rbdc.rbdc.com

-----------[000240][next][prev][last][first]----------------------------------------------------
Date:      Fri, 9 Dec 1994 19:38:50 GMT
From:      evansmp@mb51910.aston.ac.uk (Mark Evans)
To:        comp.os.linux.development,comp.os.linux.admin,comp.os.linux.misc,comp.protocols.tcp-ip,comp.protocols.tcp-ip.domains
Subject:   Re: Multi-homed (multiple IP addresses) Linux hosts?

John Paul Morrison (jmorriso@bogomips.ee.ubc.ca) wrote:
: In article <3c586q$jvr@giant.dswi.com>, Pete Kruckenberg <pete@dswi.com> wrote:
: >I saw a reference in another newsgroup to the capabilities of some new
: >*nix OS's that allow a host to have multiple IP addresses on a single
: >network interface (using an "ifconfig alias" command). 
: >
: >Does Linux have this capability, or anything similar to it? What other
: >OS's (free or commercial) support this capability? What do I need to
: >get/have to have this capability?
 
: Linux already has this: ifconfig the "dummy" interface (make sure you've
: compiled it in). Then update your routing and if necessary, your arp
: tables (proxy arp for that IP address). 
 
: You can compile in more dummy interfaces if you need them. 
 
: It would be nice to have this rolled into ifconfig so you only need
: to use one command.
 
: There's one little problem, but it probably affects other operating
: systems too, is that people can connect to your second IP address, but
: if you initiate a connection to them, the source IP address (on your
: outgoing packet) will be the IP address of the eth0 device, not of the
: dummy device. Normally this is good, but some hosts might require that
: you connect using the second IP address as your source address, and
: there's no way to change this behaviour without rewriting each network
: app.
 
: But if you need to use the dummy interface, it's probably because
: something is really evil on your network, like braindead routers/hosts
: that don't support classless routing or really bizarre routing setups
: that you are powerless to change, so you are forced to kludge it.

To fix this should not be to difficult, just requires some thinking.

-----------[000241][next][prev][last][first]----------------------------------------------------
Date:      Fri, 09 Dec 94 18:57:30 CET
From:      olla@ssgrr.it (Antonello Olla)
To:        comp.protocols.tcp-ip
Subject:   WinNT & Windows TCP/IP connection

I would know how can I connect a Windows 3.1 (or WfW 3.1 or WfW3.11) 
PC with an WinNT ( 3.1 or 3.5 ) PC using TCP/IP protocol.
I know there are two methods: PPP and SLIP (or CSLIP), but I don't know
which is supported by WinNT as server.
Can someone tell me how can I realize this connection and using what 
programs?
Thank you in advance.
Antonello



-- 
Antonello Olla
Scuola Superiore G. Reiss Romoli
S.P. per Coppito Km 0.300
67010  L'AQUILA
ITALY
Tlx 600870 SSGRR I
Tel +39-862-336-384
Fax +39-862-336-363 (o 491)
Email olla@ssgrr.it


-----------[000242][next][prev][last][first]----------------------------------------------------
Date:      9 Dec 94 19:13:14 EET
From:      jaakola@cc.helsinki.fi
To:        comp.protocols.tcp-ip,comp.protocols.ibm
Subject:   Re: AS/400 with TCP/IP

In article <3c8ign$n9s@news.bu.edu>, fmedina@bu.edu (Fernando Medina) writes:
> We are experimenting with TCP/IP on an AS/400, as APPN has
>been very unstable in our WAN. The problem we are facing is with Printer
>sessions. Anyone have any experience using Printer sessions over TCP/IP
>on an AS/400?? Wall Data's RUMBA won't support printer sessions with
>TCP/IP and we're not sure if it can be done with Chameleon 4.0, any help
>would be appreciated..

I have asked about the same question many times, and it just seems to be
the case that IBM has not defined any way to transfer "print data
streams" (whatever they are) via TCP/IP.

The only way I know is to use SNDTCPSPLF, which is a LPR client in
AS/400. It sends a print job via the LPR/LPD protocol on top of TCP/IP
to a LPD (LPR/LPD server), which prints the job. Chameleon includes such
an LPD.

The problem, however, is how to fetch print jobs from AS/400 spool
queues and call SNDTCPSPLF on them automatically. I have seen a C
program which does this, but I don't have a C compiler on AS/400. Anyone
know whether it's possible to do it in command language?
--
Juhani Jaakola, jaakola@cc.helsinki.fi

-----------[000243][next][prev][last][first]----------------------------------------------------
Date:      Fri, 9 Dec 1994 21:30:38 GMT
From:      wjw@pt.com (Wald Wojdak)
To:        comp.protocols.tcp-ip
Subject:   Q: looking for a TCP/IP source code

Hi,
 
Can anyone point me to an ftp location where I could find a TCP/IP source code.

-Thanks
-- 
__
Wald Wojdak, Performance Technologies Incorporated	wjw@pt.com
315 Science Parkway, Rochester, New York 14620            uupsi!ptsys1!wjw

-----------[000244][next][prev][last][first]----------------------------------------------------
Date:      09 Dec 1994 23:05:47 GMT
From:      gritton@abraham.et.byu.edu (Jamie Gritton)
To:        comp.protocols.tcp-ip,comp.sys.hp.hpux,comp.sys.sun.admin
Subject:   SunOS POSIX non-blocking connect doesn't work (HP-UX does)

   When I use POSIX-style nonblocking I/O (fcntl O_NONBLOCK) on a
connecting socket, HP-UX (9.05) correctly returns EINPROGRESS
immediately. SunOS (4.1.3) doesn't - it blocks until the connect call
completes, even through an agonizing wait for a timeout.
   The fcntl(5) man page on SunOS doesn't mention O_NONBLOCK, but
<sys/fcntlcom.h> (referred to in <fcntl.h>) defines it and even
mentions its POSIXness in the comment; connect(2) lists EINPROGRESS in
the ERRORS section. Using O_NDELAY works as expected, but its
semantics have some deficiencies.
   So who's wrong? I would guess Sun. And how do I get around this? As
a last (or at least later) resort, I could use O_NDELAY for the
connect, and switch to O_NONBLOCK for read/write, under the assumption
that connect is the only problem point (is it?). Or is there something
cleaner, something I'm missing? Here's a sample bit of code that works
correctly on HP but not Sun:

#include	<sys/types.h>
#include	<sys/signal.h>
#include	<sys/socket.h>
#include	<netinet/in.h>
#include	<syslog.h>
#include	<fcntl.h>
#include	<netdb.h>
#include	<stdio.h>
#include	<time.h>

extern	int	errno;

main()
{
    int			fd, len, r, w, e;
    struct sockaddr_in	sin, sin2;

    fd = socket( AF_INET, SOCK_STREAM, 0);
    fcntl( fd, F_SETFL, O_NONBLOCK);
    sin.sin_family = AF_INET;
    sin.sin_port = htons( 47000);
    sin.sin_addr.s_addr = htonl( 0x80bb030a);
    if (connect( fd, &sin, sizeof sin) < 0)
	perror( "connect 1");
    if (connect( fd, &sin, sizeof sin) < 0)
	perror( "connect 2");
    r = e = 0;
    w = 1L << fd;
    select( fd + 1, &r, &w, &e, NULL);
    if (connect( fd, &sin2, sizeof sin) < 0)
    {
	perror( "connect 3");
	len = sizeof errno;
	getsockopt( fd, SOL_SOCKET, SO_ERROR, &errno, &len);
	perror( "connect 3");
    }
 return 0;
}
--
James Gritton - gritton@byu.edu - I disclaim - "

-----------[000245][next][prev][last][first]----------------------------------------------------
Date:      10 Dec 1994 00:18:23 GMT
From:      yky@mystery.Stanford.EDU (YKY)
To:        comp.protocols.tcp-ip
Subject:   Looking for OSPF introdusction ...


HI,

I am lookinf for OSPF introduction materials.
Can someone please point me to a book, article, spec,
or anything that can teach me how OSPF works and
the protocol details?

Thank you very much.

Yours,
	YKY


-----------[000246][next][prev][last][first]----------------------------------------------------
Date:      Sat, 10 Dec 1994 00:42:00 GMT
From:      pulrich@muni.cz (Petr Ulrich)
To:        comp.protocols.tcp-ip
Subject:   gated.conf for OSPF over PPP?

Hi,

please could anybody help me with a /etc/gated.conf for OSPF over PPP?
I'm new to gated configs and OSPF.

I've got gated-3.0.3, SunOS 4.1.1

network topology:
             _______		    _____________
            |       |               |            |
  Internet  | cisco |--------PPP----|SunOS 4.1.1 |
	    |_______|               | gated      |
                                    --------------
                                          |
					TCP/IP LAN					
                                          |

In the diagram above cisco is the PPP server that connects us to the
Internet. It's routing protocol setup is only OSPF.
On our LAN gatewayed to the internet via SunOS and PPP link we don't have
any dynamic routing sice SunOS is our only gateway ouy.
We connect to cisco only sometimes and each time to a different port with a
different IP address.
We need to propagate route to our network only when we are connected + ttl.
So, therefore we need to run OSPF on our SunOS and get routes from cisco
and propagate our route to the rest of the world.

I tried to write a /etc/gated.conf on my own, but no luck :-(
here comes my /etc/gated.conf :

ospf yes {
	area 0.0.0.2 {
		interface ppp0 cost 10
			{
			enable;
			priority 10;
		};
	};
};

import proto ospfase { };

export proto ospfase type 1 {
	proto static { default metric 1; };
	proto ospfase {
		ALL
		metric 1; };
};


Please could you help me out? any pointer will be greatly appriciated.

	thank you

		petr ulrich

Please CC to pulrich@muni.cz

-----------[000247][next][prev][last][first]----------------------------------------------------
Date:      10 Dec 94 00:51:52 GMT
From:      werme@alingo.zk3.dec.com (Eric Werme)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

vjs@calcite.rhyolite.com (Vernon Schryver) writes:

>TCP implementations of NFS have been common for a few years.  The free
>source implementation of NFS from U.Guelph that is present in the free
>BSD UNIX implementations as well as BSDI's product supports NFS/TCP/IP.
>I seem to recall something about OSF supporting NFS/TCP/IP, but I could
>easily be wrong about that.

OSF/1 from OSF comes with the Univ. of Guelph code and supports NFS over
TCP.  DEC OSF/1 replaced the Guelph code with Sun ONC plus the various
optimizations made here over the years.  Hence we don't do NFS over
TCP.  But we will someday.

>However, NFS/TCP/IP is not significantly faster than NFS/UDP/IP.

Which is one of the reasons why we did NFS V3 first.  I expect NFS over
TCP to be quite a bit faster on lossy links, especially at large
block sizes.  I tried mounting a file system in Hong Kong on my
workstation in New Hampshire.  It's a good thing I had a tcpdump
running to keep me entertained!

>>I think the RPC itself can carry rather large overheads on each packet,
 
>The per-packet overhead of the rpc/xdr use by NFS is about 100 bytes.

Note that for AUTH_UNIX, the rpc header includes variable length things
like group membership and hostname, so if you simply adjust your
group membership a bit....

>Yes, the trouble is that NFS is a request-response protocol.

Note that V3 tries to address some of that by returning attributes on
most operations and the new READDIRPLUS operation that returns filenames
and the files attributes (ls -l on that system in Hong Kong was pretty awful
with V2.  I wonder if they're running NFS V3 yet.)
-- 
Eric (Ric) Werme         |  werme@zk3.dec.com
Digital Equipment Corp.  |  This space intentionally left blank.

-----------[000248][next][prev][last][first]----------------------------------------------------
Date:      10 Dec 1994 02:35:57 GMT
From:      Larry Himes <lehimes@access.digex.net>
To:        comp.protocols.tcp-ip
Subject:   Re: Q: TCP/IP source code

wjw@pt.com (Wald Wojdak) wrote:
>
> Hi,
> 
> Can anyone point me to an ftp location where I could find a TCP/IP source code.
> 
> -Thanks
> -- 
> __

Try anon ftp to dorm.rutgers.edu. 

-----------[000249][next][prev][last][first]----------------------------------------------------
Date:      10 Dec 1994 11:32:18 -0500
From:      donham@iii1.iii.net (Perry Donham)
To:        comp.protocols.tcp-ip
Subject:   Re: Thin ethernet cabling

For thinwire (10Base2), the tee connector must be attached
directly to the computer...you can't have a 'stub' running from
the wall. The problem is that 10Base2 is very picky about
reflections on the wire; any stub more than an inch or two
will cause pulses to be reflected, resulting in a severely
distorted signal.

Of course, this only aplies to machines in the middle of
the cable... :^)

If you really must have only one wire coming out of the wall,
you need to use 10BaseT (twisted pair) in a star topology.

I have 10Base2 in my house, what I did was put two connectors
at each wall plate. A short jumper allows me to have unused
net drops in the middle of the cable.

Regards,

Perry

Chuck Mattern (cmattern@mindspring.com) wrote:
: I realize this is a protocols group but do not know where else to
: turn...  
 
: I am wiring my house with c-net and would like to place the tap behind
: a wall plate and then run a short drop to the NIC rather than
: connecting the tap directly to the nic.  This would allow me to have
: only one cable coming from the wall and lessen the clutter.  Does
: anyone know if such a topology is legal/functional?  Or where else I
: might inquire?


: -- 
: ----------------------------------------------------------------------
: |Chuck Mattern              | "Not failure, but low aim, is crime."  |
: |cmattern@mindspring.com    | -James Russell Lowell-                 |
: ----------------------------------------------------------------------

-----------[000250][next][prev][last][first]----------------------------------------------------
Date:      Fri, 09 Dec 1994 09:24:20 +1100
From:      splint@mail.newcastle.edu.au (Simon Plint)
To:        comp.protocols.appletalk,comp.protocols.tcp-ip
Subject:   Re: NewsWatcher & ARA

In article <gerard-2911941253190001@jdgmac.ciit.org>, gerard@ciit.org (Jim
Gerard) wrote:

> In article <splint-2911941124360001@tbs-simon.newcastle.edu.au>,
> splint@newcastle.edu.au (Simon Plint) wrote:
> 
> > Is there anyone out there using NewsWatcher via an appletalk remote
> > access(v1.0) connection. I would really appreciate any feedback.

[snip]

> The "no permission to speak" message usually means that the news server
> is unable to look up your name from a name server.  Your machine that is
> directly connected to the net probably has a fixed IP address with a fixed
> name associated with it, so News works.  When you connect remotely via ARA,
> however, your IP address is usually assigned by the ARA server, so it may
> change from one session to another and many different remote users may be
> assigned the same IP address (but never at the same time).  For these reasons,
> some network managers do not give names to the IP addresses that are 
> assigned dynamically by servers.  The way around the problem is for the net
> manager to give the dynamic addresses generic names such as:
> ara1.newcastle.edu.au, ara2.newcastle.edu.au, etc.  You will not always have
> the same name when you connect remotely, and other users will share the
> name when they connect, but the news server will be satisfied that your IP
> address has a name.

Jim,
this is exactly what our net manager has since said to me. However, when I
check my IP address on the remote mac it is not one of the 18 that have
been set aside for the multigate that connects our appltalk net to the uni
ethernet.

-- 
Simon Plint
TBS University of Newcastle
splint@mail.newcastle.edu.au

-----------[000251][next][prev][last][first]----------------------------------------------------
Date:      10 Dec 1994 08:13:35 GMT
From:      trall@trall.almaden.ibm.com (Tony Rall)
To:        comp.protocols.tcp-ip
Subject:   Re: Local-only IP addresses?

In article <PCJ.94Dec08111448@merlin1.meche.psu.edu>,
Paul C. Janzen <pcj@sabine.acs.psu.edu> wrote:
>
>We would like to prevent hosts from other subnets from being able to access
>the printer. Can I choose the IP number so that only machines on our
>segment can see the printer?

If you have any control over the IP configuration on the printer
you can achieve your wishes without using special addresses.  Just
make sure the box doesn't have any routes other than to its own
subnet.  On a normal IP implementation you would do this by:

1. configuring your interface normally

2. not running any routing demons (such as routed or gated)

3. not setting any additional routes at all; in particular, not
   setting a default route

I suspect the printer does 2 already, but needs you to set a default
route - simply don't do it.

-- 
Tony Rall    trall@almaden.ibm.com

-----------[000252][next][prev][last][first]----------------------------------------------------
Date:      10 Dec 1994 08:52:46 GMT
From:      trall@trall.almaden.ibm.com (Tony Rall)
To:        comp.protocols.tcp-ip
Subject:   Re: HELP: Subnet routing.

In article <1994Dec8.144338.11400@usage.csd.unsw.oz.au>,
Prasert Kanthamanon <prasert@vast.unsw.edu.au> wrote:
>			+-----+
>			| G3  |
>			+-----+
>			   |IP: 202.44.8.3
>			   |Netmask: 255.255.255.00
>			   |				   Backbone
>---------------------------------------------------------- Network: 202.44.8.0
>	|					|
>	|IP: 202.44.8.5				|IP: 202.44.8.7
>	|Netmask: 255.255.255.00		|Netmask: 255.255.255.00
>     +-----+				     +-----+
>     | G1  |				     | G2  |
>     +-----+				     +-----+
>	|IP: 202.44.9.33			|IP: 202.44.9.65
>	|Netmask: 255.255.255.224		|Netmask: 255.255.255.224
>	|					|
>	|				------------- 
>     -----------			Subnet: 202.44.9.64
>     Subnet: 202.44.9.32
>
>G1 is a PC running Netware.
>G2 and G3 are SUN SPARC workstations running SunOS 4.1.3.
>
>We tried to use static routing on G2 and G3 to manage the routing task.
>However, there is a problem in the routing system.  The problem can be
>explained like this.
>Note: In every case, we have a default entry to a main gateway.
>
>My questions are:
>=================
>1. Is this network configuration valid?

Yes.

>2. What cause the problem?

It wasn't clear to me what you perceived the problem to be.  It looks
like it will work, although some packets may go through an extra
router hop before hitting the right one.  And yes, this extra hop
means the packet appears twice on the 202.44.8 net.

>3. How can I configure the routing system to handle these subnets?

This is what subnet masks are for.  If your IP software doesn't allow
you to set a mask with a route, I would call it deficient.
Nevertheless, you can still let the actual routers make all the
decisions - again, packets may take extra hops when you do it this
way.

>4. Can OSPF based routing solve the problem?

Probably, if your hosts support it.  But usually they don't (I'm
unfamiliar with your particular systems).  Serious routers usually
do have this capability, so once you get the packet to one of these
routers, the right things should happen to it.  (And the routers
you show can make it work with static routes if they don't handle
the fancy routing protocols.)

-- 
Tony Rall    trall@almaden.ibm.com

-----------[000253][next][prev][last][first]----------------------------------------------------
Date:      Sat, 10 Dec 1994 10:28:11 GMT
From:      bill@twg.bc.ca (Bill Irwin)
To:        comp.protocols.tcp-ip
Subject:   Re: Fkey characters getting split on frame relay

wolfgang@wsrcc.com (Wolfgang Rupprecht) writes:

|Delivery timing is not guarenteed.  It would be incorrect for your app
|to assume function key escape sequences will arrive in one packet.
|Why not have your app parse the key sequence the same way no matter if
|it arrived in 50ms or 50 seconds?

As it turns out, the MF/Cobol application vendor suddenly said
that there is a "new environment variable" that will control how
long the runtime waits to see if a sequence is a function key.
Setting this at 10 has resolved the problem.  However, I still
think that the best solution is at the OS level so one fix/tuning
could be done, rather than one for each application.

I like the suggestions that I have received about setting the
stty speed lower so that the timeout values are raised, thinking
that the user is on a slow serial line.  I haven't tested this
yet, but if it works you have one place to tune for network
performance.  A simple check in "/etc/profile" for a ttyp* and an
"stty 2400" is all that might be required.
-- 
Bill Irwin     -      The Westrheim Group     -    Vancouver, BC, Canada
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
uunet!twg!bill              604-431-9600 (voice) |     Your Computer  
bill@twg.bc.ca              604-430-4329 (fax)   |    Systems Partner

-----------[000254][next][prev][last][first]----------------------------------------------------
Date:      Sat, 10 Dec 1994 11:11:28 GMT
From:      SKALA@olimp.irb.hr (Karolj SKALA)
To:        comp.protocols.tcp-ip
Subject:   M/H Systems - MIPRO 95, Call for Papers



   ANNOUNCEMENT AND CALL FOR PAPERS
****************************************

http://srcm1.zems.etf.hr/mipro.html   
----------------------------------------

XVIII. International Symposium MIPRO `95
May   22.-26.  1995.
Opatija, CROATIA
Congress Center Grand Hotel Adriatic


The Croatian Society MIPRO invites you to participate in one 
or more activities at the MIPRO `95: Symposia/Conference with 
invited and contributed papers, tutorials, workshops, exhibitions 
and commercial presentations.


1. Microcomputers in Technical Systems / MTS
2. Microcomputers in Intelligent Information Systems / MIS
3. Microcomputers in Telecommunications /MTE
4. Multimedia and Hypermedia Systems /MHS
5. Microelectronics and Electronics Technologies /MEET                        


CONFERENCE ON :

Multimedia and Hypermedia Systems 

Steering Committee:

 Chair:     Karolj Skala, Institute "Rudjer Boskovic" Zagreb, Croatia
 members:   Jan Gecsei, Universite de Montreal, Canada
	    Sasa Divjak, University of Ljubljana, Slovenia
	    Kalman Fazekas, Technical University of Budapest, Hungary          
	    Fabio Vitale, University of  Bolognia, Italy
	    Pavle Dragojlovic, University of Rijeka, Croatia

This conference is devoted to presenting and exploring scientifical and 
technological advancement and innovations in field of  Multimedia and 
Hypermedia.

Major Topics

Papers describing original work and a successful applications to the 
following topics are requested:

	- Multimedia/Hypermedia Technologies and Systems
	- Multimedia/Hypermedia Applications
	- Evaluation of Multi-Hypermedia Systems
	- Computer Mediated Integral Communication
	- Computer Supported Cooperative Work
	- Virtual Reality


Submissions

Authors are invited to submit abstracts in English of 100 words 
maximum not later than  3. January, 1995.  Authors will be notified 
of acceptance, when they will send papers for submission. Submission
should include a  cover page and the full paper (3-6 pages). The cover 
page includes the title of the paper with the name of authors, complete 
addresses, phone/fax numbers, E-mail address and AV equipment needs.  
Papers may be submitted in hard-copy (send 3 copies produced by WinWord 
or WordPerfect) and in electronic form (diskettes or Internet mail/ftp) 
by  15. February 1995.

The submit please send to address:

			dr. K. Skala
		Institute "Rudjer Boskovic"
		P.O.Box 1016, 41000 Zagreb, Croatia
		Tel. +385 1 42 57 53
		Fax  +385 1 42 48 69
		e-mail:  mipro@olimp.irb.hr


Paper review

The Proceedings papers will be reviewed by the conference chairs for 
technical merit and content.


Proceeding of the Conference

Accepted papers will be published by MIPRO with ISBN number under the title:

       Conference  Proceeding:   MULTIMEDIA AND HYPERMEDIA SYSTEMS
 

Registration fee

The  registration fee of 150 DM per paper is to be paid also not later 
than 15. February, 1995 to account No. 33800-620-16-25731-3999424 of 
HD MIPRO at Rijecka banka, d.d, Rijeka, Croatia, with the participant's 
name and the designation "for MIPRO95".
One person may appear once  as the author and once as a co-author.


TUTORIALS:

Chair:    K. Skala, Institute "Rudjer Boskovic" Zagreb, Croatia
	  J. Gecsei, Universtiy of Montreal, Canada
	  P. Dragojlovic, University of Rijeka, Croatia
	  B. Zovko Cihlar, University of Zagreb
	
A.  Multimedia Technologies - Hypermedia in Education

Multimedia is an emerging and exciting field that requires integration 
of telecommunication, audio/video technologies, computing hardware and 
software and information services. This seminar discusses the concepts 
of multimedia, the reasons why multimedia is becoming a key technology, 
what are the key applications, and the need for various multimedia 
component-technologies like compression, storage, hardware/software etc. 
The detailed description of Hypermedia Education technology will be done. 
An overview of architectural requirements for supporting Multimedia and 
Hypermedia will be discussed. 


 B. High-Speed Network for Multimeda

     prof.dr. Jan Gecsei, Universtiy of Montreal, Canada

1. High-speed networking in general
- Technological evolution, from twisted pairs to fiber optics, common 
  carrier offerings
- Where we are in LAN, WAN
	systems: ethernet+variations, FDDI-II, Token Ring, DQDB, BISDN
	H.s. protocols principles, Sonet, STM, ATM protocol reference model
- H.s. switching technology,  principles, topologies, implementations 
- H.s. network requirements by application areas 
- Internet and electronic highways, national h.s. infrastructures 
	what it is for whom?
	the last kilometer problem
		
2. H.s. networks for multimedia
- Gigabits: who needs them, what can you do with them?
- Media transmission characteristics, rate requirements, isochrony, 
  multipoint topology
- Main issues
	supporting media integration and synchronisation
	providing high-level services for developers
	handling Quality of service (QoS)
- Standardization
	protocols with QoS: HeiTP, XTPX, ST-II
	protocol stacks
	support for multipoint communication     
- Examples of MM networks/systems: 
	US National Gigabit Network Testbed 
	Berkom, Pandora

3. Networked MM systems
- Central issue: handling distributed MM objects
- QoS centered design, issues, examples
- Application areas
	generic services: conversational, messaging, retrieval, broadcast 
	example applications: cooperative design, distance teaching/learning
	distributed interactive real-time visualization, distributed VR
- Existing systems, products, projects
	Broadband Services Project of NSRC in Canada  
	WWW is becoming MM
- Related standardization processes
	OSF DCE (Open Software Foundation's Distributed Computing Environment)
	ODP (Open Distributed Processing) 
	CORBA (Common Object Request Broker Architecture) 
	MMSS (MM System Services Architecture) proposal by IBM, HP, Sun

4. Conclusions, questions, doubts
- Where we go?
- Is this what we want?



SEMINAR TOPICS:

	 - MULTIMEDIA /HYPERMEDIA   OVERWIEV
	 - MULTIMEDIA COLLABORATION, GROUPWORK
	 - HIGH-SPEED NETWORKS
	 - MULTIMEDIA MAIL, SOFTWARE AND DATABASES
	 - COMMUNICATION  AND COMPRESSION TECHNOLOGIES
	 - MULTIMEDIA/HYPERMEDIA DEVELOPMENT TOOLS
	 - HYPERMEDIA IN EDUCATIONAL TECHNOLOGIES
	 - MULTIMEDIA/HYPERMEDIA  MARKET TRENDS AND ITS FUTURE


WORKSHOPS AND EXHIBITIONS:

MIPRO`95  includes the Exhibition of Information Technologies and 
Electronics with eighteen year old tradition. World`s leading manufacturers, 
as well as  small  perspective  companies are  frequent  guests at   the 
exhibition. Professional and commercial presentations are a part of  the 
exhibition, which makes MIPRO `95 both a professional experience and  an 
opportunity for business contact. You are kindly invited to express your 
interest for this part of Europe by taking part at this exhibition.
Exhibitors and presentations have the status of CO-organizers of MIPRO`95, 
and as such are referred to in the MIPRO`95  Final Notice, in the symposium 
and tutorial proceedings, in all public information media, as well as at the 
exhibition and conference rooms.
 

WELCOME  TO OPATIJA

Opatija`s present tradition as the place of meeting dates back to the time 
of Franz Joseph I and Wilhelm II. , the Queen of tourism on the Adriatic and  
one of the most popular convention venues in Croatia.
We are proud of our home  town Opatija. A town where Middle Europe and the 
Mediterranean meet. An exclusive climate, rich vegetation and fascinating 
architecture. Where the first impressions are unique elegance during from 
the times when amidst camellia and palm-trees Europe`s elite and 
Austro-Hungarian aristocracy met and rested here in romantic villas.
Situated in the beautiful surroundings of the Opatija Riviera, ADRIATIC 
guarantees pleasant accommodation, rich gastronomic offer, excellent 
entertainment, shopping and many other amenities.


CONGRESS CENTER

Convention Center GRAND HOTEL ADRIATIC is much more than just a center 
offering technical and banqueting services. A highly trained staff, the 
newest technical equipment and cooperation with national management 
companies have to date ensured the sissies of conventions and incentive 
programs held in ADRIATIC: FIS, International Council of Social Welfare, 
EFNMS, IPCA and many others.
Designed as an amphitheater the convention hall seats over 600 participants 
in comfortable armchairs with tables of folding desks. The convention center 
possesses first class audio-visual technical equipment: personal system with 
several microphones, slide-projectors, overhead-projectors, videorecorders 
with monitors and video-projectors and MPC computers.
Grand Hotel Adriatic is part of the total Opatija convention and 
accommodation offer, which consists of 15 halls for over 2300 participants 
and 6000 beds. It is supported by many world renowned first class 
restaurants, several casinos and night clubs plus lots of entertainment's 
and amenities.


-----------[000255][next][prev][last][first]----------------------------------------------------
Date:      11 Dec 1994 00:31:01 -0500
From:      barmar@nic.near.net (Barry Margolin)
To:        comp.protocols.tcp-ip,comp.dcom.lans.fddi,comp.dcom.lans.ethernet
Subject:   Re: Max. # of routers on a network?

In article <D0I43n.50@cherokee.nsuok.edu> landin@cherokee.nsuok.edu (Mark C. Landin) writes:
>Is having 30+ routers on one network kosher? Even if it is, are there
>"gotchas" to it? Thanks for any replies.

There's no specific limit on the number of routers on a network.  The only
gotcha I can think of is the large amount of routing protocol traffic.
-- 

Barry Margolin
BBN Internet Services Corp.
barmar@near.net

-----------[000256][next][prev][last][first]----------------------------------------------------
Date:      11 Dec 1994 00:44:39 -0500
From:      gwright@connix.com (Gary Wright)
To:        comp.protocols.tcp-ip
Subject:   Re: Available buffer space in a UDP socket ??

In article <3c5pll$gj1@hpindda.cup.hp.com>, Rick Jones <raj@cup.hp.com> wrote:
>Amit Somani (somani@sun2.cs.wisc.edu) wrote:
>: Is there any way to find out the amount of available (free) space
>: associated w/ a UDP socket ? The UDP socket has intitialized its 
>: buffer space by a call to setsockopt().
>
>On many (BSDish) systems, the UDP datagram spends so little time on
>the SO_SNDBUF (outbound) that it is unlikely (I think) to be an issue.

Actually it spends no time on the send buffer,  UDP datagrams are
passed directly to the output queue of the outgoing interface when
the kernel processes the write/sendto/sendmsg system call.

For datagram protocols, SO_SNDBUF is the maximum amount of data
that can be sent in a single datagram. For BSD 4.4-lite, this
defaults to 9*1024 = 9216.
-- 
Gary Wright

-----------[000257][next][prev][last][first]----------------------------------------------------
Date:      11 Dec 1994 00:47:06 -0500
From:      barmar@nic.near.net (Barry Margolin)
To:        comp.protocols.tcp-ip
Subject:   Re: Where is ARP implemented?

In article <3ca344$820@bigguy.timeplex.com> pete@louie.timeplex.com writes:
>I'm wondering where the ARP function is implemented most of this
>time.  Is it implemented as part of the IP layer (on the bottom
>side), or is it part of the Network Interface Layer (and hence
>can use network specific methods to resolution)?  I'm inclined to
>think it's part of the NI layer?

ARP should be implemented in the link layer, since it's not specific to IP,
and IP can use links that don't make use of ARP.  In general, the interface
between the network layer (such as IP or Appletalk) and the link layer
(IEEE 802.2, Localtalk, X.25, or PPP) needs a way to ask "what is the MAC
address for address A of protocol P".  The network layer doesn't care how
the link layer resolves this; on an IEEE 802.2 network it will use ARP.
-- 

Barry Margolin
BBN Internet Services Corp.
barmar@near.net

-----------[000258][next][prev][last][first]----------------------------------------------------
Date:      11 Dec 1994 00:51:27 -0500
From:      barmar@nic.near.net (Barry Margolin)
To:        comp.protocols.tcp-ip
Subject:   Re: "Default" Hosts

In article <3cdo7v$k48@aristotle.algonet.se> mhugo@algonet.se (Mikael Hugo) writes:
>I have one host I would like the telnet to go into. ie telnet mycorp.dom 
>would connect to bbs.mycorp.dom. Same for gopher, ftp, fsp, etc. (Spread 
>on several different machines in a C-net). 

Just specify the same address in the DNS for mycorp.dom and
bbs.mycorp.dom.  Or make bbs.mycorp.dom a CNAME for mycorp.dom.
You can use either method for the other equivalences.
-- 

Barry Margolin
BBN Internet Services Corp.
barmar@near.net

-----------[000259][next][prev][last][first]----------------------------------------------------
Date:      11 Dec 1994 01:10:30 -0500
From:      barmar@nic.near.net (Barry Margolin)
To:        comp.protocols.tcp-ip
Subject:   Re: ftp control connection protocol

In article <RV.94Dec10223833@tahoe.cs.brown.edu> rv@tahoe.cs.brown.edu (rodrigo vanegas) writes:
>1) What "existing telnet module"?  The size of the BSD telnet client
>sources exceed those of the BSD ftp client!  What am i supposed to be
>extracting?

The comment in the RFC assumes that the system has a "telnet module" as a
callable library.  The expectation was probably that many application
protocols would be defined that use the telnet protocol as a "presentation"
layer.  But "modern" systems such as Unix aren't constructed using such
modularity; there's a telnet I/O module in Multics and Genera, but I don't
think there's an easy way to extract it from most Unix telnet
implementations.

>2) What does "very little of the Telnet Protocol" mean?  Am i supposed
>to already known which little part?

An FTP implementation doesn't need to support most of the options; it just
needs enough of the option protocol so that it can recognize the
negotiations ignore them.  And most of the other Telnet IAC sequences are
only useful in interactive sessions (e.g. Abort Output), so it just needs
to be able to recognize them and ignore them.
-- 

Barry Margolin
BBN Internet Services Corp.
barmar@near.net

-----------[000260][next][prev][last][first]----------------------------------------------------
Date:      11 Dec 1994 01:17:40 -0500
From:      barmar@nic.near.net (Barry Margolin)
To:        comp.protocols.tcp-ip
Subject:   Re: rlogin vs. telnet

In article <RV.94Dec10222250@tahoe.cs.brown.edu> rv@tahoe.cs.brown.edu (rodrigo vanegas) writes:
>Can someone please explain the major differences between the two?  I
>know that telnet can connect to any port other than the default, but i
>am confused about the purpose of rlogin if telnet does the same thing
>(save automatic login).

Rlogin was designed specifically for Unix-to-unix logins (although it has
since been implemented on other systems).  It automatically passes the
login name, and makes use of the Unix-specific "privileged port" mechanism
to indicate that the client login name is correct.  It uses the Unix
character set on the wire.  In general, it's a much simpler protocol than
Telnet.

Telnet is a much more general protocol, designed for a wide variety of OSes
and environments.  It has a very flexible option negotiation mechanism, to
support many combinations of client and server systems.  As a result of the
full generality, it's a more complex protocol.
-- 

Barry Margolin
BBN Internet Services Corp.
barmar@near.net

-----------[000261][next][prev][last][first]----------------------------------------------------
Date:      10 Dec 1994 19:28:02 GMT
From:      er@sai.msu.su (Evgeny Rodichev)
To:        comp.unix.solaris,comp.sys.sun.admin,comp.protocols.tcp-ip
Subject:   Re: OK .. now, how to get the SPARCserver 1000 to see it's router ??

The following is an example from our SPARCserver 1000 with two interfaces:

/etc:14>ifconfig -a
lo0: flags=849<UP,LOOPBACK,RUNNING,MULTICAST> mtu 8232
        inet 127.0.0.1 netmask ff000000 
le0: flags=863<UP,BROADCAST,NOTRAILERS,RUNNING,MULTICAST> mtu 1500
        inet 158.250.30.1 netmask ffffff80 broadcast 158.250.30.127
le1: flags=863<UP,BROADCAST,NOTRAILERS,RUNNING,MULTICAST> mtu 1500
        inet 158.250.30.129 netmask ffffff80 broadcast 158.250.30.255

/etc:15>netstat -nr

Routing Table:
  Destination           Gateway           Flags  Ref   Use   Interface
-------------------- -------------------- ----- ----- ------ ---------
127.0.0.1            127.0.0.1             UH       0    681  lo0
158.250.30.0         158.250.30.1          U        3  16910  le0
158.250.30.128       158.250.30.129        U        2   2084  le1
224.0.0.0            158.250.30.1          U        3      0  le0
default              158.250.30.2          UG       0   2383

cat /etc/defaultrouter
158.250.30.2

As a result the route through gateway 158.250.30.2 works well.

Regards.
Evgeny Rodichev

Network administrator of sai.msu.su
  

-----------[000262][next][prev][last][first]----------------------------------------------------
Date:      10 Dec 1994 23:34:58 GMT
From:      bmah@propaganda.CS.Berkeley.EDU (Bruce A. Mah)
To:        comp.protocols.tcp-ip
Subject:   Re: To find a USENIX paper about delivering NetNews using multicasting

John Chueng-Hsien Lin writes:
> I am interested in reading a 1994 Winter USENIX paper about using
> IP multicasting to deliver Net News.  (Sorry, I don't know the
> paper's title and name(s) of the author(s)) 
> Is the paper on-line?

K. Lidl, J. Osborne, and J. Malcolm.  "Drinking from the Firehose:
Multicast USENET News", Proceedings of the USENIX Winter 1994
Technical Conference, San Francisco, CA, January 1994.

The authors can be reached by email at lidl@uunet.uu.net,
stripes@uunet.uu.net, and jmalcolm@uunet.uu.net (so says the paper).

Bruce.
--
Bruce A. Mah		   Graduate Student	          bmah@CS.Berkeley.EDU
		Tenet Group, Computer Science Division
		 University of California at Berkeley

-----------[000263][next][prev][last][first]----------------------------------------------------
Date:      Sun, 11 Dec 1994 00:43:30 GMT
From:      bpeoples@iglou.iglou.com (Ben Peoples)
To:        comp.protocols.tcp-ip
Subject:   Thin/ThickNet

I've seen this term in a few places, and had considered it just a generic term,
but then I recently saw it used specifically, so what is Thinnet and Thicknet,
I know its not the name of a network, but a type of network...

				Thanks,
					 Ben

-- 
Ben Peoples					       	bpeoples@iglou.com
Unless otherwise stated, the above opinions are ***WORDS***
"As I said, with this net I will catch them, I bet.  With this net I will catch
those things, yet!"  --- _The Cat in The Hat_

-----------[000264][next][prev][last][first]----------------------------------------------------
Date:      Sun, 11 Dec 1994 01:20:07 GMT
From:      hjb@netcom.com (squeedy)
To:        comp.protocols.tcp-ip
Subject:   Re: SLIP pins for RS232

Chris Schefler (css@netcom.com) wrote:


: Hi, 
 
: I'm setting up a hardwired SLIP connection from my UNIX server to a PC
: using RS232 connectors with modular adapters (RJ25).  

one some machines this type of connection is sufficient as long as
you have a minimal null-modem type connection (e.g. vxworks slip/tty
drivers will work fine with this type of serial line).

on most unix machines you will need to have a fully handshaked null-modem
line:  	1-1, 2-3/3-2, 4-5/5-4, 6-(8&20)/(8&20)-6, 7-7

-- 
Hwa-Jin Bae (hjb@netcom.com)
Peaceful Star, Oakland, CA		ftp.netcom.com:/pub/hj/hjb/README

-----------[000265][next][prev][last][first]----------------------------------------------------
Date:      Sun, 11 Dec 1994 02:52:38 GMT
From:      cap@isac.hces.com (Simon Casady)
To:        comp.protocols.tcp-ip
Subject:   Re: Fkey characters getting split on frame relay

Bill Irwin (bill@twg.bc.ca) wrote:
: wolfgang@wsrcc.com (Wolfgang Rupprecht) writes:
 
: |Delivery timing is not guarenteed.  It would be incorrect for your app
: |to assume function key escape sequences will arrive in one packet.
: |Why not have your app parse the key sequence the same way no matter if
: |it arrived in 50ms or 50 seconds?


: I like the suggestions that I have received about setting the
: stty speed lower so that the timeout values are raised, thinking
: that the user is on a slow serial line.  I haven't tested this
: yet, but if it works you have one place to tune for network
: performance.  A simple check in "/etc/profile" for a ttyp* and an
: "stty 2400" is all that might be required.
: -- 

I have seen this question over and over it must be in a FAQ somewhere.
but I don't know where.  Anyway, stty has for many years
if not since the beginning had a configurable timer when in raw mode.
You can tell it how long to wait for another char in 0.1 second increments.
see the min and time "control-characters" in the man pages for stty and
termio.
Unfortunately not all applications let you get to stty to change the
settings and some that do reset them.  However if you can use this
feature it does work.

-- 
Simon Casady                                      Health Care Expert Systems
cap@hces.com                                      voice       (515)-222-1717
casady@acm.org                                    fax         (515)-222-1716

-----------[000266][next][prev][last][first]----------------------------------------------------
Date:      Sun, 11 Dec 1994 03:22:49 GMT
From:      rv@tahoe.cs.brown.edu (rodrigo vanegas)
To:        comp.protocols.tcp-ip
Subject:   rlogin vs. telnet

Can someone please explain the major differences between the two?  I
know that telnet can connect to any port other than the default, but i
am confused about the purpose of rlogin if telnet does the same thing
(save automatic login).

-- 
rodrigo vanegas
rv@cs.brown.edu

-----------[000267][next][prev][last][first]----------------------------------------------------
Date:      Sun, 11 Dec 1994 03:38:33 GMT
From:      rv@tahoe.cs.brown.edu (rodrigo vanegas)
To:        comp.protocols.tcp-ip
Subject:   ftp control connection protocol

The FTP RFC says that the control connection of an FTP session should
comply to the telnet protocol.  As rationale, it goes on to suggest
that one "may make use of the existing Telnet module in the system".
Alternatively, it adds, one can simply write it oneself since "in
practice, FTP relies on very little of the Telnet Protocol, so [this]
approach does not necessarily involve a large amount of code."

I have two questions.  

1) What "existing telnet module"?  The size of the BSD telnet client
sources exceed those of the BSD ftp client!  What am i supposed to be
extracting?

2) What does "very little of the Telnet Protocol" mean?  Am i supposed
to already known which little part?

-- 
rodrigo vanegas
rv@cs.brown.edu

-----------[000268][next][prev][last][first]----------------------------------------------------
Date:      11 Dec 1994 03:30:23 +0100
From:      mhugo@algonet.se (Mikael Hugo)
To:        comp.protocols.tcp-ip
Subject:   "Default" Hosts

How do I go about making default host for a domain, is this default routed?

I have one host I would like the telnet to go into. ie telnet mycorp.dom 
would connect to bbs.mycorp.dom. Same for gopher, ftp, fsp, etc. (Spread 
on several different machines in a C-net). 

Please give me some advice on where to look.

--

  -------------------------------------------------------------------------
  |  Mikael Hugo                     |  Running Novell, OS/2, DOS, WIN    |
  |  DataPhone Online Systems        |  Multinode - 7.4 Gb Files - E-mail |
  |  Box 209, 161 15 Bromma, Sweden  |  BBS: +46-8-7043121 HST DS         |
  -------------------------------------------------------------------------

-----------[000269][next][prev][last][first]----------------------------------------------------
Date:      11 Dec 1994 21:59:20 -0500
From:      barmar@nic.near.net (Barry Margolin)
To:        comp.protocols.tcp-ip,comp.dcom.lans.fddi,comp.dcom.lans.ethernet
Subject:   Re: Max. # of routers on a network?

In article <D0nKp6.60G@calcite.rhyolite.com> vjs@calcite.rhyolite.com (Vernon Schryver) writes:
>In article <3ce2ql$kf1@tools.near.net> barmar@nic.near.net (Barry Margolin) writes:
>>There's no specific limit on the number of routers on a network.  The only
>>gotcha I can think of is the large amount of routing protocol traffic.
 
>What large amount of traffic is that?  30 routers, each with an incredibly
>large 750 routes and needing 30 packets is only about 30 packets/second
>of routing protocol traffic, which is less than 2% of the available
>1MB/s bandwidth of an Ethernet.

I wasn't thinking specifically of the network bandwidth (although I admit
that I hadn't calculated it, so didn't realize it was so small).  If RIP is
used, the routing packets will all be broadcasts, so every workstation on
the network will be interrupted by every routing packet.  And if they're
running routed, the packets will cause process switches.

Admittedly, one process switch a second isn't really that much of a burden
on modern systems.  I never said this would be a *big* problem, but it's
something that should be anticipated.  For instance, you might noticed
routed in top's output more than before.
-- 

Barry Margolin
BBN Internet Services Corp.
barmar@near.net

-----------[000270][next][prev][last][first]----------------------------------------------------
Date:      Sun, 11 Dec 1994 15:36:42 GMT
From:      vjs@calcite.rhyolite.com (Vernon Schryver)
To:        comp.protocols.tcp-ip,comp.dcom.lans.fddi,comp.dcom.lans.ethernet
Subject:   Re: Max. # of routers on a network?

In article <3ce2ql$kf1@tools.near.net> barmar@nic.near.net (Barry Margolin) writes:
>In article <D0I43n.50@cherokee.nsuok.edu> landin@cherokee.nsuok.edu (Mark C. Landin) writes:
>>Is having 30+ routers on one network kosher? Even if it is, are there
>>"gotchas" to it? Thanks for any replies.
>
>There's no specific limit on the number of routers on a network.  The only
>gotcha I can think of is the large amount of routing protocol traffic.

What large amount of traffic is that?  30 routers, each with an incredibly
large 750 routes and needing 30 packets is only about 30 packets/second
of routing protocol traffic, which is less than 2% of the available
1MB/s bandwidth of an Ethernet.  A more plausible number of 20
routes/router in even a large network is more like 1 or 2 packets/router,
which is less than 0.1% of the bandwidth available on an Ethernet.

(Of course, 750 routes is a very large number only for private internets.
On the Internet it is less than noise.)


Vernon Schryver    vjs@rhyolite.com

-----------[000271][next][prev][last][first]----------------------------------------------------
Date:      12 Dec 94 03:55:24 -0500
From:      scottpf@rcwusr.bp.com
To:        comp.protocols.tcp-ip
Subject:   MacTCP and Static addressing

Static addressing with MacTCP

Does anybody use MacTCP with static addressing ?. We have ca. 400 macintoshes
that I would like to switch to using static addressing, however I am unable to
say if this would work. 

I am worried that if we switch on static addressing, that the router (a cisco
ags+) will not be able to function with proxy-rarp. Does anybody else use this
? with 4 segments and 400 macintoshes.


Any help/information would be usfull.

Thanks,

Peter Scott

-----------[000272][next][prev][last][first]----------------------------------------------------
Date:      Sun, 11 Dec 1994 23:37:15 GMT
From:      paul@atlas.abccomp.oz.au (Paul Brooks)
To:        comp.protocols.tcp-ip
Subject:   Re: Fkey characters getting split on frame relay

In article <D0BCMr.FG3@twg.bc.ca> bill@twg.bc.ca (Bill Irwin) writes:
|I have a situation where the characters being sent by a
|terminal's function keys are being split into separate packets as
|they travel through the network on their way to the host
|application.  This usually results in an error "beep", with one
|or more straggling characters being displayed at the cursor
|position.

This is a fairly common problem - function keys tend to send three or 
more bytes at once, and these may be split over several network packets,
depending on the sending software. After they are split, there's nothing
that an intermediate router/bridge can do - the problem must be solved
at one of the end points.

It can't be solved in the OS / TCP/IP of the receiving host in any
good way - the OS, TCP/IP or Telnet code doesn't know the data 
corresponds to a single logical data block. Only the application, or
some termcap/terminfo settings, might be altered to recognise the first
byte of a multi-byte sequence, and delay a little waiting for any
subsequent bytes to "catch up". We see that in applications running here -
the first byte of a vt220 function-key sequence is an ESC char, and this can
also occur by itself to exit from menus. When ESC is hit by itself, and
nothing else is sent, the remote host 'hesitates' for approx. half a second,
then exits the menu. If another ESC or other character is sent, the menu
exits instantly - the application recognises it must have been an isolated
ESC, not a multi-byte sequence.

|Does anyone know of a way to ensure that Fkeys get recognized
|reliably by applications running on the host/server?

There should be a tunable timeout setting for the first byte of a multi-byte
sequence, typically the ESC character, in the application.
Good luck!

-- 
Paul Brooks              |paul@abccomp.oz.au       |Emerging Standard:
TurboSoft Pty Ltd        |pwb@newt.phys.unsw.edu.au|  one that has not yet
579 Harris St., Ultimo   |                         |  been superseded.
Sydney Australia 2007    |ph: +61 2 281 3155       |  

-----------[000273][next][prev][last][first]----------------------------------------------------
Date:      Mon, 12 Dec 1994 00:26:57 GMT
From:      adam@comptech.demon.co.uk (Adam Goodfellow)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

In article <D0I07n.903@calcite.rhyolite.com>,
                Vernon Schryver <vjs@calcite.rhyolite.com> wrote:


<Sorry - removed all your interesting comments re NFS>

>Yes, the trouble is that NFS is a request-response protocol.
>

Perhaps it is time for someone to re-examine the RPC used by NFS as I
should think it could very usefully gain from an extension to deal with
"large" messages (ie files, or parts of) by means of a sliding window or
similar extension that is transparently used as appropriate.

-- 
Adam

=======================================================================
| Computech  Tel/Fax: 0181 673 7817  email: adam@comptech.demon.co.uk |
=======================================================================

-----------[000274][next][prev][last][first]----------------------------------------------------
Date:      Mon, 12 Dec 94 00:40:33 GMT
From:      alun@fc.net (Alun Jones)
To:        comp.protocols.tcp-ip
Subject:   Re: ftp control connection protocol

rv@tahoe.cs.brown.edu (rodrigo vanegas) wrote:
>The FTP RFC says that the control connection of an FTP session should
>comply to the telnet protocol.  As rationale, it goes on to suggest
>that one "may make use of the existing Telnet module in the system".
>Alternatively, it adds, one can simply write it oneself since "in
>practice, FTP relies on very little of the Telnet Protocol, so [this]
>approach does not necessarily involve a large amount of code."

Have a read of RFC 1123 - it goes on a little further on this topic, to
mention that the telnet stuff that an FTP server needs to deal with is the
DO/WON'T communication.  An FTP client doesn't really need
to do anything like telnet - if that's what you're writing, you should be
able to survive with just connecting, and sending commands or data -
the telnet protocol information really doesn't need to be sent.

Alun.
====


-----------[000275][next][prev][last][first]----------------------------------------------------
Date:      Mon, 12 Dec 94 09:19:01 PDT
From:      mcrump@BALL.COM
To:        comp.protocols.tcp-ip,comp.protocols.ibm
Subject:   Re: AS/400 with TCP/IP


In article <1994Dec9.191314.1@cc.helsinki.fi>, <jaakola@cc.helsinki.fi> writes:
> In article <3c8ign$n9s@news.bu.edu>, fmedina@bu.edu (Fernando Medina) writes:
> > We are experimenting with TCP/IP on an AS/400, as APPN has
> >been very unstable in our WAN. The problem we are facing is with Printer
> >sessions. Anyone have any experience using Printer sessions over TCP/IP
> >on an AS/400?? Wall Data's RUMBA won't support printer sessions with
> >TCP/IP and we're not sure if it can be done with Chameleon 4.0, any help
> >would be appreciated..
> 
> I have asked about the same question many times, and it just seems to be
> the case that IBM has not defined any way to transfer "print data
> streams" (whatever they are) via TCP/IP.
> 
> The only way I know is to use SNDTCPSPLF, which is a LPR client in
> AS/400. It sends a print job via the LPR/LPD protocol on top of TCP/IP
> to a LPD (LPR/LPD server), which prints the job. Chameleon includes such
> an LPD.
> 
> The problem, however, is how to fetch print jobs from AS/400 spool
> queues and call SNDTCPSPLF on them automatically. I have seen a C
> program which does this, but I don't have a C compiler on AS/400. Anyone
> know whether it's possible to do it in command language?
> --
> Juhani Jaakola, jaakola@cc.helsinki.fi
Yes.  There is a program in QUSRTOOL called SNDNETOUTQ which sends all the 
spooled files in the named OUTQ (except those in open status) to a user in a 
network.  The program was designed to front end the SNDNETSPLF command which is 
SNADS distribution.  Since source is provided it is easy to replace the 
SNDNETSPLF with SNDTCPSPLF.  Now, to get this to function I have a shell 
program that runs automatically and checks the queue on a regular basis (5 
minutes).  We just started using this program this week.  I am using this as a 
stop gap to satisfy current demand while I work on a program that uses the API 
QRCVDTAQ - or should I say investigate.  I'm not sure as to the benefits of one 
approach over the other.  I do know that we have two products from different 
vendors that use the API approach to processing OUTQ's.  I plan on talking to 
them as to why they use this approach.  

In the meantime, I think you will find the modified QUSRTOOL approach to be 
sufficient and easy to implement.  It took me 20 minutes from start to finish - 
that is once I got the LPR to LPD issues out of the way.  As I said before, we 
just started using it and so far so good.

One other note - we are currently only using this for one OUTQ so......

For AS/400 to AS/400 print transfer I am using a product called Remote 
SpoolPrint from Broderick.  We are extremely pleased with the capabilities and 
the 'industrial strength'.  It is not too expensive $300 - $4,000 
approximately.  We have it installed on a E70 and a 310-2043 at $2,400 a pop.  
If you are interested I can get you some more info on Broderick.  It's around 
here somewhere.

-----------[000276][next][prev][last][first]----------------------------------------------------
Date:      12 Dec 1994 11:18:04 -0500
From:      donham@iii1.iii.net (Perry Donham)
To:        comp.protocols.tcp-ip
Subject:   Need help with simple gated.conf

I've been banging my head on gated.conf for two days now. I've read
the crab book, looked at the sample files on the WWW, and still can't
seem to get what should be a simple configuration running.

Here's what I want to do:

I have a local subnet that I want to announce via a SLIP interface
to my service provider's router using RIP.

Subnet: 199.232.43.32 mask 255.255.255.240
Gateway: SLIP is 199.232.40.133, local ethernet is 199.232.43.34
Provider: Router (and my default) is 199.232.40.9

I'm receiving RIP updates from 199.232.40.9 fine, but no updates
are being sent from my gateway (I'm using tcpdump to watch the packets).

I'm using gated r3_5alpha_7 on a Linux box. The gated.conf I've been
mucking with is:

rip yes;  
static
	{
        default gateway 199.232.40.9; 
        199.232.43.32 mask 255.255.255.240 gateway 199.232.43.34;
	};
export proto rip interface 199.232.40.133
{
        proto direct
	{
                199.232.43.32 mask 255.255.255.240 metric 1; 
	};
};

Any help would be much appreciated.

Regards,

Perry


-----------[000277][next][prev][last][first]----------------------------------------------------
Date:      Mon, 12 Dec 1994 03:14:11 GMT
From:      mhunt@epx.cis.umn.edu (Matt Hunt)
To:        comp.protocols.tcp-ip
Subject:   SLIP via a telnet session?

  Can someone set me straight?  I thought I read somewhere that one could use a dialup connection to a server, run a SLIP host application, and use that dialup connection as SLIP connection.  Allowing me to run Winsock applications after
doing some setup work of course.  Is this possible?  If so could you point me toa site were the SLIP host software is avaible?  Is it shareware or a commercial
application?

Thanks in advance,

Matt

mhunt@csom.umn.edu

-----------[000278][next][prev][last][first]----------------------------------------------------
Date:      Mon, 12 Dec 1994 07:11:39 GMT
From:      adam@comptech.demon.co.uk (Adam Goodfellow)
To:        comp.protocols.tcp-ip
Subject:   Re: simple_Client_Serv_HELP

In article <D0FC2z.K6z@pts.mot.com>,
                Jason Kronz x8645 p5412 <jkronz@pts.mot.com> wrote:

>Well, I have been trying to get this to work for many days...
>could someone please download this and try to get these 
>two c programs to work?  They are a simple INET Client-Server pair.
>the Server sends a Welcome and the client replys PRINT JOB:.
>
>That's it... I know the client works because I have connected to the
>DAYTIME port and it does seem to read the time off of that port 
>just fine...
>

I havnt actually tried running either of these, but a quick look shows
that the struct sockaddr pointer passed to bind() in your server isnt
actually set to anything.

ServerSockAddrPtr = (struct sockaddr *) &ServerINETAddress;

before the bind() call would probably do wonders...

-- 
Adam

=======================================================================
| Computech  Tel/Fax: 0181 673 7817  email: adam@comptech.demon.co.uk |
=======================================================================

-----------[000279][next][prev][last][first]----------------------------------------------------
Date:      Mon, 12 Dec 1994 12:25:51
From:      dvlch@hol.ath.forthnet.gr (Dimitris P. Vlachogiannis)
To:        comp.protocols.tcp-ip
Subject:   Unisys 6000/65 (SVR4) with PC (WFW311) with slip

I try to connect my pc (wfw311 - trumpet winsock 2.0b) with the my Unisys 
6000/65 running Unisys SVR4 unix but after the slattach command on the port 
and when i do ping to my pc i get an IP header checksum error or an ICMP 
checksum error.

I would be thankfull for any assistance.

DVlch@hol.ath.forthnet.gr


-----------[000280][next][prev][last][first]----------------------------------------------------
Date:      12 Dec 1994 16:10:29 -0500
From:      barmar@nic.near.net (Barry Margolin)
To:        comp.protocols.tcp-ip,comp.dcom.lans
Subject:   Re: packets sent to own address

In article <1994Dec12.094346.2937@zh014.ubs.ubs.ch> urs.eberle@zhflur.ubs.ubs.ch writes:
>What happens exactly to a packet which is sent to it's own address --> 

It probably depends on the operating system.

>1. The packet is sent from an ip-driver to it's own address (not the ip-
>   loopback address). Does the ip-layer construct a packet for the lan
>   (i.e. the ethernet)?

In most Unix implementations I believe the IP layer detects that the
destination address is one of the local addresses, and simply moves the
packet from the send queue to the receive queue.  I don't think it ever
sends it to the interface driver.

>I need to know, if somebody with (hardware)access to a LAN
>is able to listen to traffic created by a tcp-Client-Server connection on
>the same computer, not using the ip-loopback address.

I just tried to use tcpdump to capture packets between a SunOS 4.1.3 system
and itself, and didn't see any.

I suggest that if you're really concerned, you do something similar.  Ping
or telnet to the address of the machine and look at the traffic from
another system.
-- 

Barry Margolin
BBN Internet Services Corp.
barmar@near.net

-----------[000281][next][prev][last][first]----------------------------------------------------
Date:      12 Dec 1994 16:37:35 -0500
From:      barmar@nic.near.net (Barry Margolin)
To:        comp.protocols.tcp-ip
Subject:   Re: TCP RFC 793 to BERKLEY SOCKETS correlation  -  post.fil [1/1]

In article <3chl6b$jsk@getty.tekelec.com> simha@tekelec.com (Ajay Simha) writes:
>Question:  When using the Berkley Socket Interface to access the TCP protocol,
>           where (when) does one specify the "foreign socket" that the RFC
>	    talks about (I mean which call)?

The sockets API doesn't provide a way to specify the foreign socket in a
passive OPEN.  It can only be specified in an active OPEN, as a parameter
to the connect() call.

>2. RFC 793, Page 12, first paragraph
>	"If there are several pending passive OPENs (recorded in TCBs) with the 
>same local socket, an foreign active OPEN will be matched to a TCB with the
>specific foreign socket in the foreign active OPEN, if such a TCB exists,
>before selecting a TCB with an unspecified foreign socket."
 
>Question: Assuming the "socket" the RFC is talking about means a logical
>	   end point in a TCP connection, which really translates to a pair
>	   of Hostaddress+Portnumber combination, how is it possible to open and
>	   bind using the same local hostaddress+portnumber combination.  Will
>	   we not get a "Address already in use" error?

Since sockets doesn't allow you to specify the foreign information in a
passive open, there isn't much point in allowing two TCBs to listen to the
same port.  Unfortunately, Unix still gets this wrong; it doesn't allow you
to bind to a local address/port combo even if all the existing TCBs have
foreign info specified (i.e. they are for existing connections).  This is
why you have to use the SO_REUSEADDR option when restarting a server.

>3. Question: How does one specify the PUSH flag on most TCPs?
>	 	 My understanding is, to use the TCP_NODELAY flag through
>	 	 the setsockoption type of call.  But, does this PUSH every
>	 	 time I do a send() or is just for that one time?

It causes each send() to have the PUSH flag.
-- 

Barry Margolin
BBN Internet Services Corp.
barmar@near.net

-----------[000282][next][prev][last][first]----------------------------------------------------
Date:      Mon, 12 Dec 1994 09:43:46 GMT
From:      zhebu@svusenet.ubs.ch
To:        comp.protocols.tcp-ip,comp.dcom.lans
Subject:   packets sent to own address

Hello

What happens exactly to a packet which is sent to it's own address --> 

1. The packet is sent from an ip-driver to it's own address (not the ip-
   loopback address). Does the ip-layer construct a packet for the lan
   (i.e. the ethernet)?

2. If ethernet-driver has to send a packet with it's own ethernet address,
   does it send the packet to the (physical) LAN?

3. How do other LAN-Interfaces handle this situation?

I need to know, if somebody with (hardware)access to a LAN
is able to listen to traffic created by a tcp-Client-Server connection on
the same computer, not using the ip-loopback address.

Thanks for enlighten me.

--
NAME   Urs Eberle
EMAIL  urs.eberle@zhflur.ubs.ubs.ch
PHONE  ++41-1-236-58-08

-----------[000283][next][prev][last][first]----------------------------------------------------
Date:      Mon, 12 Dec 1994 12:30:03 +0000
From:      peter@bhcnt.demon.co.uk (Peter Gross)
To:        comp.protocols.tcp-ip
Subject:   IP address translation ?

This is basically a similar question to one I asked a few months ago, I would
be grateful for more specfic details now please.

We have a large internal TCP/IP network using unregistered IP addresses.

I have set-up my PC running Linux to talk to my local network, and when I
issue a particluar command I swap to my official Internet address and attach
to the provider <DEMON>.

	Script does this:
	-	copy current network files to safe place
	-	copy Internet network files to /etc
	-	reboot

I have a similar one to put me back onto the local net.

I now want to take this further, and understand that it is possible to
configure the system to translate local addresses to valid IP numbers, and 
vice-versa.

The connection to the Ineternet is via SLIP over a comm port, and my local 
connection is via the Ethernet card.

BTW, I will discuss my agreement with Demon BEFORE trying. :-)

Peter Gross
<peter@bhcnt.demon.co.uk>
-- 
Peter Gross
Unix Systems Manager
Brighton Health Care NHS Trust
#include <std.disclaimer>

-----------[000284][next][prev][last][first]----------------------------------------------------
Date:      12 Dec 94 17:30:27
From:      billw@glare.cisco.com (William )
To:        comp.protocols.tcp-ip
Subject:   Re: Interpretation of Nagle algorithm


    I'm looking for clarification of the Nagle algorithm as described in
    RFC 896 and RFC 1122.  It concerns behaviour when a buffer is
    queued to TCP for transmission which exceeds the segment size (say 1.5
    x segment size), when there are no unacknowledged segments.

Strictly speaking, I think if your data shows up in chunks of 1.5MSS, then
the nagle algorithm doesn't apply, and you get left with the slow start
algoritrhms instead.  Nagle's algorithm applys when your data is trickling
in a byte or two at a time...

One of the elegant things about the slow start algorithms is that they
nicely devolve to the Nagle algorithm if you assume that the congestion
window gets reset whenever you run out of data.

If it makes you feel better, you can always say that the first MSS gets
sent because it is a full packet, and the second chunk is the first piece
which the nagle algorithm applies to.

Finally, both Nagle and slow start are network optimization algorithms,
so it doesn't really matter much if you are off by one in the number of
packets you send before they start limiting output.  The "2 packets" from
slow start is chosen pretty arbitrarilly, and the Nagel algorithm is still
good for your network even you allow more than one packet on the retransmit
queue to prevent additional packets...

BillW
cisco


-----------[000285][next][prev][last][first]----------------------------------------------------
Date:      12 Dec 1994 13:03:32 GMT
From:      stefani@yesway.lkg.dec.com (Larry Stefani)
To:        comp.protocols.tcp-ip,comp.dcom.lans
Subject:   Re: packets sent to own address


In article <1994Dec12.094346.2937@zh014.ubs.ubs.ch>, zhebu@svusenet.ubs.ch writes:
|>What happens exactly to a packet which is sent to it's own address --> 
|>
|>1. The packet is sent from an ip-driver to it's own address (not the ip-
|>   loopback address). Does the ip-layer construct a packet for the lan
|>   (i.e. the ethernet)?

Typically, the stack loops it back internally.  That is, the driver never sees
the request to transmit a packet destined to itself.

|>2. If ethernet-driver has to send a packet with it's own ethernet address,
|>   does it send the packet to the (physical) LAN?

Depends.  Some drivers may need to filter those packets out because of hardware
limitations.  Others just send it out.  I suspect that most drivers just send
it out.

|>3. How do other LAN-Interfaces handle this situation?

On FDDI, I just send it out.  Unless the "copy-own" bit is set, I won't copy
that packet up, but I will strip it.

+---------------------------------------------------------------------------+
| Larry Stefani                               stefani@yesway.lkg.dec.com    |
| Networks Engineering                        Digital Equipment Corporation |
|                       Comments are mine, of course...                     |
+---------------------------------------------------------------------------+

-----------[000286][next][prev][last][first]----------------------------------------------------
Date:      12 Dec 1994 13:38:02 GMT
From:      Ralf Hack <hack@pluto.rz.fh-reutlingen.de>
To:        comp.protocols.tcp-ip
Subject:   Linux-Router cannot cope with Sun-tcpip of our local stations

Hello world,

we setup a linux-router for our local network. With the local suns,
we cannot route files (via ftp) above roughly 1k. However, we can
receive them with our router. The same for xsessions. Routing sessions
with netscape/mosaic .. does not work. However, routing an xterm-inal
does work.

We use a 486/33 PC with Linux 1.1.23. Our subnet consists of SGI's and
PC's. The suns are running SunOS 5.3 Generc sun4m sparc.

Any hint to solve this problem is very much appreciated. 

Thanks in advance.

mfg
Ralf Hack

-----------[000287][next][prev][last][first]----------------------------------------------------
Date:      12 Dec 1994 14:02:51 GMT
From:      simha@tekelec.com (Ajay Simha)
To:        comp.protocols.tcp-ip
Subject:   TCP RFC 793 to BERKLEY SOCKETS correlation  -  post.fil [1/1]


Hi Netters,

I have the following issues/questions that I would like to put out to you:

1. RFC 793, Page 11, last paragraph
   "There are two principal cases for matching the sockets in the local passive
OPENs and an foreign active OPENs.  In the first case, the local passive OPENs 
has fully specified the foreign socket.  In this case, the match must be exact.
In the second case, the local passive OPENs has left the foreign socket 
unspecified.  In this case, any foreign socket is acceptable as long as the 
local sockets match.  Other possiblities include pareially restricted matches."

Question:  When using the Berkley Socket Interface to access the TCP protocol,
		   where (when) does one specify the "foreign socket" that the RFC
		   talks about (I mean which call)?

2. RFC 793, Page 12, first paragraph
	"If there are several pending passive OPENs (recorded in TCBs) with the 
same local socket, an foreign active OPEN will be matched to a TCB with the
specific foreign socket in the foreign active OPEN, if such a TCB exists,
before selecting a TCB with an unspecified foreign socket."

Question: Assuming the "socket" the RFC is talking about means a logical
		  end point in a TCP connection, which really translates to a pair
		  of Hostaddress+Portnumber combination, how is it possible to open and
		  bind using the same local hostaddress+portnumber combination.  Will
		  we not get a "Address already in use" error?

3. Question: How does one specify the PUSH flag on most TCPs?
			 My understanding is, to use the TCP_NODELAY flag through
			 the setsockoption type of call.  But, does this PUSH every
			 time I do a send() or is just for that one time?


All responses will be appreciated.

-ajay






-----------[000288][next][prev][last][first]----------------------------------------------------
Date:      12 Dec 94 15:10:28 GMT
From:      werme@alingo.zk3.dec.com (Eric Werme)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

adam@comptech.demon.co.uk (Adam Goodfellow) writes:

>In article <D0I07n.903@calcite.rhyolite.com>,
>                Vernon Schryver <vjs@calcite.rhyolite.com> wrote:
 
>>Yes, the trouble is that NFS is a request-response protocol.
>>
 
>Perhaps it is time for someone to re-examine the RPC used by NFS as I
>should think it could very usefully gain from an extension to deal with
>"large" messages (ie files, or parts of) by means of a sliding window or
>similar extension that is transparently used as appropriate.

This is more a transport problem than a request-response problem.
NFS over TCP can send very long packets - both the NFS V2 and V3 protocols
have a 32 bit field for the data length!  Even the UDP 8800 byte limit
is more arbitrary than mandated by UDP.

Also, keep in mind that Vernon was responding to grumbles about FTP being
faster than NFS.  One thing he should have pointed out is that FTP was
designed for accessing whole files.  FTP just won't work for random
access to a database without hacking the protocol.  NFS was designed
for implementing a file system.  OSes tend not to have system calls that
say "read the whole file", or at least the system calls are not used
that way.  OSes are asked to read or write part of a file.  That's a
request/response sort of transaction, and that what NFS was designed to
match.

For distributing files across the Internet, even FTPing files isn't that
great.  That's why people keep compressed tar archives. Three FTP commands
and you can get 1,000 files with no server overhead to compress the network
traffic.  NFS and FTP have as many similarities as as apples and oranges,
but they are not interchangable!
-- 
Eric (Ric) Werme         |  werme@zk3.dec.com
Digital Equipment Corp.  |  This space intentionally left blank.

-----------[000289][next][prev][last][first]----------------------------------------------------
Date:      Mon, 12 Dec 1994 15:37:48 GMT
From:      vjs@calcite.rhyolite.com (Vernon Schryver)
To:        comp.protocols.tcp-ip,comp.dcom.lans.fddi,comp.dcom.lans.ethernet
Subject:   Re: Max. # of routers on a network?

In article <3cgea8$l8d@tools.near.net> barmar@nic.near.net (Barry Margolin) writes:

> ...
>I wasn't thinking specifically of the network bandwidth (although I admit
>that I hadn't calculated it, so didn't realize it was so small).  If RIP is
>used, the routing packets will all be broadcasts, so every workstation on
>the network will be interrupted by every routing packet.  And if they're
>running routed, the packets will cause process switches.
>
>Admittedly, one process switch a second isn't really that much of a burden
>on modern systems.  I never said this would be a *big* problem, but it's
>something that should be anticipated.  For instance, you might noticed
>routed in top's output more than before.
> ...

Some of us have hacked commercially distributed versions of routed to
not wake up on every received RIP packet when they do not need to.
As I recall, it's just a matter of fiddling with select() masks and
timeouts.  It seemed to help the number of CPU cycles consumed by 
non-forwarders.


Vernon Schryver    vjs@rhyolite.com

-----------[000290][next][prev][last][first]----------------------------------------------------
Date:      Mon, 12 Dec 1994 16:00:15 GMT
From:      vjs@calcite.rhyolite.com (Vernon Schryver)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

In article <19941212.002657.67@comptech.demon.co.uk> adam@comptech.demon.co.uk writes:
>
>Perhaps it is time for someone to re-examine the RPC used by NFS as I
>should think it could very usefully gain from an extension to deal with
>"large" messages (ie files, or parts of) by means of a sliding window or
>similar extension that is transparently used as appropriate.

NFS v3 supports bigger blocks, but that still not the same.

I do not think such an idea is good.  The trouble is that sliding windows
is a particular (or more general) way of doing what was once called
double buffering.  Such things are well and good when you know you are
going to be reading the source of data (e.g. 7-track tape) for the next
several minutes.  If you know you want the next several minutes of a
disk file, then you could use a network file system with sliding window,
timers, or various other techniques.  On the hand, moving several minutes
of a file sounds to me more like a "file transfer" than a "file access,"
and so why not use a "file transfer protocol"?

File accesses tend to jump all over the file system.  As far as the
server and the network is concerned, network file system requests are
rarely [transfer file #1, transfer file #2, ...].  Instead they are more
like [transfer middle 8K of file #99, transfer last 4K of file #2, ...]
A file system that decides to transfer all 10MB of file #99 before
delivering those last 4K of file #2 will not perform well.

To say it a third way, it is very difficult to get the right information
from the application on the network file system client to the network
file system code on the server and client to anticipate what the
application will need and when.  It's hard to connect something like
TCP's window advertisments to the application through the file system
code.


Vernon Schryver    vjs@rhyolite.com

-----------[000291][next][prev][last][first]----------------------------------------------------
Date:      Mon, 12 Dec 1994 16:12:56 GMT
From:      skov@ilx.com (John Skovron)
To:        comp.protocols.tcp-ip
Subject:   Re: bootpclient for MS-DOS/Windows

In article <Pine.LNX.3.91.941208094421.94A-100000@kiko.paradigm.co.za>,
Peter K  <pko@kiko.paradigm.co.za> wrote:
>Hi, Russell
>
>>    I am looking for a bootp client running under MS-Dos and Windows, either one 
>>    executable which can be used by both or seperate ones.
>> 
>> What do you plan to do with the data that the bootp client returns?
>> It's not much good unless you're going to use it with a TCP/IP package...
>
>We are already running tcp/ip. The only thing is to allocate IPs and set
>gateways to individual PCs. Under Win NT AS one can run MS's equivalent
>but then I have to assign a pool of IP addresses, which is something I
>don't want to do. We have a 300+ box network (running Slo-es/2, SCO,
>Linux, SunOS, AIX, MS-Dos, Win 3.11, WFW and Win NT) accross 4 class C
>subnets and from a management perspective (under our circumstances) the
>assignment of an IP-address pool is just not on. 

Many of the PC TCP/IP protocol stacks can be configured to use BOOTP.
Which stack are you running?

For instance, I believe Novell's TCP/IP stack (which might be Beame &
Whiteside's), will use BOOTP if you specify "ip_address 0.0.0.0" in
the net.cfg file.

OTOH, I don't know if the 32-bit "Wolverine" stack from Msft will
support ordinary BOOTP, as opposed to DHCP.

A BOOTP client is useless unless it is closely integrated with the
protocol stack.

-- 
-----------------------------------------------------------------------------
John Skovron
ILX Systems Inc.
skov@ilx.com

-----------[000292][next][prev][last][first]----------------------------------------------------
Date:      Mon, 12 Dec 1994 16:19:50 GMT
From:      skov@ilx.com (John Skovron)
To:        comp.protocols.tcp-ip
Subject:   Re: DHCP Info ???

In article <D0IJvC.FA7@agedwards.com>,
Chris Cleeland <cleelacj@agedwards.com> wrote:
>In article <Pine.ULT.3.91.941202202814.2391A-100000@chinook.halcyon.com>,
>Robin Callender  <robinc@halcyon.com> wrote:
>>I'm pretty sure (vis. know) that Win/NTAS has a DHCP server with it.  
>
>Yes.
>
>>Also I believe that HP has a DHCP server for their unix system.
>
>News to me.  Haven't seen hide nor hair of that yet.
>

I checked my saved copy of the "report on 4th DHCP backoff" from
the DHCP mailing list (host-conf@sol.cs.bucknell.edu), and it says
that HP only has a client implementation.

DHCP server implementations can be had from Sun, WIDE (whatever
that is), Microsoft, FTP, Competitive Automation (whatever that
is), and SGI, according to the same message.

-- 
-----------------------------------------------------------------------------
John Skovron
ILX Systems Inc.
skov@ilx.com

-----------[000293][next][prev][last][first]----------------------------------------------------
Date:      Mon, 12 Dec 1994 16:39:03 GMT
From:      iialan@iifeak.swan.ac.uk (Alan Cox)
To:        comp.os.vxworks,comp.protocols.tcp-ip
Subject:   Re: Simultaneous TCP/Socket Connections

In article <D0I6pF.EDr@aston.ac.uk> evansmp@mb51913.aston.ac.uk (Mark Evans) writes:
>None of the RFC's appear clear on this. 
>One posibility is to treat this as the simaltanious open.

If two people issue a connect() at each other at precisely the same moment
then it works with BSD just nicely. Of course its basically impossible
to achieve that synchronization !

Alan
-- 
  ..-----------,,----------------------------,,----------------------------,,
 // Alan Cox  //  iialan@www.linux.org.uk   //  GW4PTS@GB7SWN.#45.GBR.EU  //
 ``----------'`--[Anti Kibozing Signature]-'`----------------------------''
One two three: Kibo, Lawyer, Refugee :: Green card, Compaq come read me...

-----------[000294][next][prev][last][first]----------------------------------------------------
Date:      12 Dec 1994 17:30:49 GMT
From:      ian@spider.co.uk (Ian Heavens)
To:        comp.protocols.tcp-ip
Subject:   Interpretation of Nagle algorithm

I'm looking for clarification of the Nagle algorithm as described in
RFC 896 and RFC 1122.  It concerns behaviour when a buffer is
queued to TCP for transmission which exceeds the segment size (say 1.5
x segment size), when there are no unacknowledged segments.

The first full size segment is transmitted, according to Nagle; what
about the second?  I'm told that the correct interpretation is that
it should be sent, even though the 2nd segment is not full size,
and there is unacknowledged TCP data.  

Thanks

ian

---
	 Ian Heavens			ian@spider.co.uk                 
 	 Spider Software
 	 Spider Park, Stanwell Street		
	 Edinburgh, EH6 5NG, Scotland	+44 31 555 5166 (Ext 4735)
--

-----------[000295][next][prev][last][first]----------------------------------------------------
Date:      12 Dec 1994 17:52:35 GMT
From:      sbrog@drexler.nwest.mccaw.com (Steve Brog)
To:        comp.protocols.tcp-ip
Subject:   Re: Help! binding socket to address with bind dies with already..

In article <3ca6ac$q0d@news.eecs.uic.edu> anewman@ernie.eecs.uic.edu  
(Aaron Newman) writes:
> Hello.
> 
> I am developing a  small TCP-IP C/S app and when I try to bind()
> before listen()ing on a port, it works fine the first time.  But after I
> exit the app and restart, Bind dies with 'address in use' error code.
> This goes on for ~10 minutes, and then it works again.  I am closing
> the socket down after the app exit with 'close()'.  Is this the wrong
> way to close it?  The man pages are no help. (surprise)
> 
> This is on sparc Solaris 2.3
> 
> What stupid thing am I (not) doing?
> 

You might want to use the command "netstat-an |grep your_port_number" To  
figure out what state the port is in while it is hung.  A successfull  
close on the socket should go to TIME_WAIT state quickly, then dissappear  
in 10 seconds or so.  If you are hanging in CLOSE_WAIT, FIN_WAIT_1 or  
FIN_WAIT_2, TCP was not able to finish the shutdown of the connection.  I  
would guess the most likely problem is the other side of the connection  
did not get close()ed, so it ends up timing out after 10 min.  

Steve Brog



-----------[000296][next][prev][last][first]----------------------------------------------------
Date:      Mon, 12 Dec 94 18:41:24 GMT
From:      oberman@icaen.llnl.gov
To:        comp.protocols.tcp-ip
Subject:   RE: Thin/ThickNet

In Article <D0MFCI.Lsy@iglou.com>
bpeoples@iglou.iglou.com (Ben Peoples) writes:
>I've seen this term in a few places, and had considered it just a generic term,
>but then I recently saw it used specifically, so what is Thinnet and Thicknet,
>I know its not the name of a network, but a type of network...

Both are types of IEEE 802.3 (ISO 8802-3) networks commonly known as Ethernet.
They are identical except for the physical medium used. Thinnet is properly
called 10Base2 and Thicknet is 10Base5. 10Base2 uses RG-58 C/U or similar cable
while 10Base5 used the old yellow Ethernet cable.

Due to differences in the media, the maximum length of 10Base5 is 500 meters
and 10Base2 is 185 meters for a single, unrepeated segment. 10Base2 is aslso
restricted to 30 connections per segment.


To learn a lot more about this stuff, take a look at comp.dcom.lans.ethernet.

R. Kevin Oberman			
Energy Sciences Network (ESnet)
National Energy Research Supercomputer Center (NERSC)
Lawrence Livermore National Laboratory (LLNL)
Internet: koberman@llnl.gov		+1 510-422-6955

-----------[000297][next][prev][last][first]----------------------------------------------------
Date:      12 Dec 1994 18:46:37 GMT
From:      jstevens@vnet.ibm.com (Jeffrey S. Stevens)
To:        comp.protocols.tcp-ip,comp.protocols.ibm
Subject:   Re: AS/400 with TCP/IP

Well, you could also request the LPRMON PACKAGE from me, by sending a request,
which I make available as freeware.  Here's the pacakge description:

*************************  IBM Unclassified ****************************
* :nick.LPRMON     :sec.IBM Unclassified              :disk.JSTEVENS
* :title.LPR Monitor source code for V2R3M0/V3R0M5
* :version.1.1     :date.94/07/06   :summary.NENOUNCE :support.NO
* :oname.Jeff Stevens/Joe Lindsay   :onode.GDLVM6     :ouser.JSTEVENS
* :aname.Jeff Stevens/Joe Lindsay   :anode.GDLVM6     :auser.JSTEVENS
* :ops.OS/400                       :lang.C CL
* :kwd.AS/400 OS/400 TCP/IP LPR LPD V2R3M0 V3R0M5
* :abs.Source code for the LPR monitor.  Found in the AS/400 Redbook
* GG24-4028 "IBM AS/400 Printing III".
***********************************************************************
*                 Standard Official IBM Disclaimer
*
* This material contains programming source code for your consideration.
* These examples have not been thoroughly tested under all conditions.
* IBM, therefore, cannot guarantee or imply reliability, serviceability,
* performance or function of these programs.  All programs contained
* herein are provided to you "AS IS".  THE IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE EXPRESSLY
* DISCLAIMED.
***********************************************************************
* Similar code is built-in to V3R1, therefore this package is only needed
* on V2R3M0/V3R0M5 systems.  This code will run on V3R1 systems however.
*
* LPR Monitor provides the ability to reroute spooled files from a
* designated output queue to a specified LPD server.
*
* Internally requestable via:
*
*            REQUEST LPRMON PACKAGE FROM JSTEVENS AT GDLVM6
*
* Externally requestable by sending mail request to:
*
*            Internet: holodeck@vnet.ibm.com, jeff_s@vnet.ibm.com,
*                      jlindsay@vnet.ibm.com
*            Bitnet:   jstevens at vnet
*            Fax:      (607) 752-5421
*            IBMMail:  usib3ntq@ibmmail
************************************************************************
      LPRMON   USAGE    *  How to use the LPRMON tool
      LPRMON   CHANGES  *  Recent changes or fixes to this package
      LPRMON   V3R1     *  Describes V3R1 remote printing support
      LPRMON   H        *  LPR Monitor header file
                        *  (copy this file into QGPL/H.LPRMON)
      LPRMON   QCSRC    *  LPR Monitor C source file
                        *  (copy this file into QGPL/QCSRC.LPRMON)
      STRLPRMO QCMDSRC  *  LPR Monitor CMD source file
                        *  (copy this file into QGPL/QCMDSRC.STRLPRMON)
************************************************************************
* The following command will create the CL programs that compile
* the LPRMON build tools (below):
*
* V2R3M0:  CRTCLPGM PGM(BLDLPRMON) SRCFILE(QGPL/QCLSRC) SRCMBR(BLDLPRMON)
*          TEXT('Create LPR Monitor compilation tool')
*
* V3R0M5:  CRTCLPGM PGM(BLDLPR305) SRCFILE(QGPL/QCLSRC) SRCMBR(BLDLPR305)
*          TEXT('Create LPR Monitor compilation tool')
*
* Once the programs are created, compile the LPR Monitor code by:
*
*   V2R3M0:  CALL BLDLPRMON
*   V3R0M5:  CALL BLDLPR305
************************************************************************
      BLDLPRMO QCLSRC   *  V2R3M0 CL source to compile LPR Monitor
                        *  (copy this file into QGPL/QCLSRC.BLDLPRMON)
      BLDLPR35 QCLSRC   *  V3R0M5 CL source to compile LPR Monitor
                        *  (copy this file into QGPL/QCLSRC.BLDLPR305)


-- 
-jeff

------------------------------------------------------------------------
Jeffrey S. Stevens, AS/400 TCP/IP Applications
IBM Corporation, 40B/16-2N006,  1701 North Street,  Endicott, NY 13760 
Internet:  holodeck@vnet.ibm.com, jeff_s@vnet.ibm.com  
Bitnet:    jstevens at vnet
Phone: (607) 752-5488   Fax: (607) 752-5421   IBMMail: usib3ntq@ibmmail
------------------------------------------------------------------------

* This note carried on 100% recycled electrons *

-----------[000298][next][prev][last][first]----------------------------------------------------
Date:      12 Dec 1994 20:12:21 GMT
From:      aej@mikasa.WPI.EDU (Allan E Johannesen)
To:        comp.protocols.tcp-ip
Subject:   Re: Help! binding socket to address with bind dies with already..

int on = 1;

setsockopt(s, SOL_SOCKET, SO_REUSEADDR, (char *)&on, sizeof on);

-----------[000299][next][prev][last][first]----------------------------------------------------
Date:      12 Dec 1994 21:19:07 GMT
From:      carlson@Xylogics.COM (James Carlson)
To:        comp.protocols.tcp-ip
Subject:   Re: TCP RFC 793 to BERKLEY SOCKETS correlation  -  post.fil [1/1]

In article <3chl6b$jsk@getty.tekelec.com>, simha@tekelec.com (Ajay Simha) writes:
|> 
|> Hi Netters,
|> 
|> I have the following issues/questions that I would like to put out to you:
|> 
|> 1. RFC 793, Page 11, last paragraph
|>    "There are two principal cases for matching the sockets in the local passive
|> OPENs and an foreign active OPENs.  In the first case, the local passive OPENs 
|> has fully specified the foreign socket.  In this case, the match must be exact.
|> In the second case, the local passive OPENs has left the foreign socket 
|> unspecified.  In this case, any foreign socket is acceptable as long as the 
|> local sockets match.  Other possiblities include pareially restricted matches."
|> 
|> Question:  When using the Berkley Socket Interface to access the TCP protocol,
|> 		   where (when) does one specify the "foreign socket" that the RFC
|> 		   talks about (I mean which call)?

One doesn't do this.  For passive (aka "listen") sockets, only the local
address may be restricted through the bind() call.

|> 2. RFC 793, Page 12, first paragraph
|> 	"If there are several pending passive OPENs (recorded in TCBs) with the 
|> same local socket, an foreign active OPEN will be matched to a TCB with the
|> specific foreign socket in the foreign active OPEN, if such a TCB exists,
|> before selecting a TCB with an unspecified foreign socket."
|> 
|> Question: Assuming the "socket" the RFC is talking about means a logical
|> 		  end point in a TCP connection, which really translates to a pair
|> 		  of Hostaddress+Portnumber combination, how is it possible to open and
|> 		  bind using the same local hostaddress+portnumber combination.  Will
|> 		  we not get a "Address already in use" error?

No.  The only use for this mechanism is to match up incoming packets
with established sessions before matching against listen sockets.  For
example, suppose we have:

	Local addr/port		Foreign addr/port
A:	1.0.0.1/5		1.0.0.2/1000
B:	1.0.0.1/5		0.0.0.0/1000

The spec says that we match against the more specific tuple (A) first,
before matching against (B).  That way, incoming packets for the
established session represented by A are sent to that socket, and all
unrecognized packets are sent to the listen socket at B.

Theoretically, this could also be used to discriminate among multiple
listen sockets, but I've never seen that done.

|> 3. Question: How does one specify the PUSH flag on most TCPs?
|> 			 My understanding is, to use the TCP_NODELAY flag through
|> 			 the setsockoption type of call.  But, does this PUSH every
|> 			 time I do a send() or is just for that one time?

The TCP_NODELAY flag turns off the Nagle algorithm.  As far as I'm able
to tell, it has nothing to do with the PUSH bit in any implementation.

At least in the BSD4.3 code, PUSH is set when the data being sent by TCP
fits entirely within the window and there's no more to send.  That is to
say, it does it automatically by default.  (The comments in the code say
that this is done to work around bugs in implementations which need to
see either a full buffer or a PUSH before any data are delivered to the
application level.)

---
James Carlson <carlson@xylogics.com>            Tel:  +1 617 272 8140
Annex Software Support / Xylogics, Inc.               +1 800 225 3317
53 Third Avenue / Burlington MA  01803-4491     Fax:  +1 617 272 2618

-----------[000300][next][prev][last][first]----------------------------------------------------
Date:      12 Dec 1994 22:16:12 GMT
From:      tmt@hri.com (Thomas M. Talpey)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

Adam Goodfellow:
>Perhaps it is time for someone to re-examine the RPC used by NFS as I
>should think it could very usefully gain from an extension to deal with
>"large" messages (ie files, or parts of) by means of a sliding window or
>similar extension that is transparently used as appropriate.

Ric Werme:
> This is more a transport problem than a request-response problem.
> NFS over TCP can send very long packets - both the NFS V2 and V3 protocols
> have a 32 bit field for the data length!  Even the UDP 8800 byte limit
> is more arbitrary than mandated by UDP.

Vern Schruyver:
>However, NFS/TCP/IP is not significantly faster than NFS/UDP/IP.

Ric Werme:
> Which is one of the reasons why we did NFS V3 first.  I expect NFS over
> TCP to be quite a bit faster on lossy links, especially at large
> block sizes.

I did the NFS/TCP implementation in stock OSF/1, which is similar to
Rick Macklem's Guelph TCP but was developed independently. On OSF/1,
as with other NFS/TCP implementations at Connectathon, I found NFS/TCP
to be up to 15% faster than NFS/UDP. The largest gain is in bulk data
transfer, where the TCP protocol really shines through. The little
round-trippy stuff (stat, lookup, ...) is identical. This is over local
Ethernet, over low bandwidth or lossy links then I would consider TCP
to be a necessity. Vern is quoting Rick's Usenix paper which found that
TCP was the same speed, this was not my experience.

Now, over long latency links, another factor enters in, the ability of
the client to keep the pipeline full. One way is to make the RPC's bigger,
this is only feasible over a TCP transport. It gives the double benefit
of reducing the RPC overhead (fewer RPC's), and making better use of the
link by giving TCP enough to work with. Another way is to provide more
"parallelism" in the form of extra biod's, which will use the filesystem
readaheads to keep the pipe full. However, the characteristics of both
the network and the remote server will need to feed back to the biod
scheduler to make best use of them, since there are only a few. Some of
this feedback is in the stock OSF/1 code, in the form of congestion
control, but it could certainly be further improved.

Neither of these solutions is a protocol issue - Ric is right that there
is no length limit in the V2 protocol, and the V3 protocol has a large one.
Adam Goodfellow's suggestion of sliding windows is effectively provided by
the existence of one or more biod's, and is purely an implementation issue.
But running a large number of biod's will congest either the network or the
server without the feedback I mentioned. And, the number is different for
each server and each network path, plus it changes dynamically. In a stock
OSF/1 system you can run a LOT of biod's and the throughput curve will
simply flatten out. On other clients, it will rise to a certain point
and then fall of dramatically. This is why it's such a problem to tune
the biod count on such systems.

Tom Talpey
tmt@hri.com

-----------[000301][next][prev][last][first]----------------------------------------------------
Date:      Mon, 12 Dec 1994 23:05:16 GMT
From:      Olof.Ekedahl@englund.lu.se (Olof Ekedahl)
To:        comp.infosystems.www.users,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.infosystems.www.providers,comp.os.ms-windows.networking.tcp-
Subject:   Re: WWW Server/Client without INTERNET??

In article <3c9p4i$i2h@nntp.gov.ab.ca>, gordon@aec.env.gov.ab.ca says...
>
>briant@gateway.bsis.com (Brian Toole) wrote:
>>
>> You have to have a "NULL" WinSOCK DLL for this to work.
>> Read the faq that is included with the Windows version
>> of Mosaic and it'll tell you where to find a copy on the
>> net.
 
>> Brian Toole                            Broadway & Seymour, Inc.
>
>Hi Brian.  I believe you missed some of the messages on this thread.
>
>The original poster wanted to supply WWW services to the other PCs on the 
>same LAN.
>
>Therefore he needs a complete setup with a WWW server and WWW clients.
>
>Gordon McAndrew
>Vegreville, Alberta
>gordon@aec.env.gov.ab.ca
>
I haven't tried this, but wouldn't it be possible to run Mosaic or Netscape with a 
nullsock and access the HTML documents off the LAN server as local files?

Olof
-- 
__________________________________________________________________
_
Olof Ekedahl			Telephone: +46 46 107557
Department of English		Telefax: +46 46 107547 
Helgonabacken 14		E-mail: Olof.Ekedahl@englund.lu.se
S-223 62 LUND
Sweden


-----------[000302][next][prev][last][first]----------------------------------------------------
Date:      Mon, 12 Dec 1994 23:48:37 GMT
From:      lavoiema@IRO.UMontreal.CA (Martin Lavoie)
To:        comp.unix.solaris,comp.protocols.tcp-ip
Subject:   Re: Help! binding socket to address with bind dies with already..

Aaron Newman (anewman@ernie.eecs.uic.edu) wrote:
: Hello.
 
: I am developing a  small TCP-IP C/S app and when I try to bind()
: before listen()ing on a port, it works fine the first time.  But after I
: exit the app and restart, Bind dies with 'address in use' error code.
: This goes on for ~10 minutes, and then it works again.  I am closing
: the socket down after the app exit with 'close()'.  Is this the wrong
: way to close it?  The man pages are no help. (surprise)
 
: This is on sparc Solaris 2.3
 
: What stupid thing am I (not) doing?
 
: Thanks for saving me (again),
: Aaron

Hi Aaron,

	If you always bind the same reserved port, you can use 
the socket option SO_REUSEADDR like this:

        struct sockaddr_in      sock_addr;
        int                     tmp = 1;

        if ((ListenSocket = socket(AF_INET, SOCK_STREAM, 
					IPPROTO_TCP)) == -1)
                return(ERRSOCK);

        sock_addr.sin_family      = AF_INET;
        sock_addr.sin_port        = htons(Port);
        sock_addr.sin_addr.s_addr = INADDR_ANY;

        setsockopt(ListenSocket,SOL_SOCKET,SO_REUSEADDR,
			(char*)&tmp,sizeof(tmp));

        if (bind(ListenSocket, (struct sockaddr*) &sock_addr, 
			sizeof sock_addr) == -1)
                return(ERRBIND);


	If it does not answer your question and fix your problem, 
let me know, maybe I don't understand your question.

- martin

--
lavoiem@cmr.ca
Martin Lavoie
Centre de Recherche en Informatique Distribuee (CRID),
College militaire royal de Saint-Jean,
Richelain (Quebec), Canada.

-----------[000303][next][prev][last][first]----------------------------------------------------
Date:      13 Dec 1994 08:27:45 -0500
From:      andy@learnix.ca (Andy Barclay)
To:        comp.protocols.tcp-ip,comp.dcom.lans.fddi,comp.dcom.lans.ethernet
Subject:   Re: Max. # of routers on a network?

There is one other caveat about using 30 routers.
Remeber that RIP defines a route as unreachable if its more than
15 hops away.

-- 
Regards,
Andy W. Barclay.        andy@learnix.ca

   Isn't it great now that UNIX is user-friendly!

-----------[000304][next][prev][last][first]----------------------------------------------------
Date:      Tue, 13 Dec 1994 00:13:55 GMT
From:      pie@netcom.com (Michael Melo)
To:        comp.protocols.tcp-ip
Subject:   TCP-IP Response Times Over Gateways

I'm looking for some methods of testing TCP-IP response times over
multiple hops/gateways between one machine and another.


I'll be running an application in one state (GA) and making requests to
another machine in another state (MA) over a T1 link and am wondering
if anyone has written any apps that report time.

e-mail responses to: michael@heaven.com, jb@heaven.com





-- 
-------------+----------------+------------------------------------------------
Michael Melo | pie@netcom.com | "Often we are sad animals -- bored dogs,
-------------+----------------+  monkeys getting rained on." - R. Hass,
   Time Warner Interactive    |  _Human Wishes_ (Ecco Press) 
------------------------------+------------------------------------------------

-----------[000305][next][prev][last][first]----------------------------------------------------
Date:      13 Dec 1994 10:55:39 -0500
From:      esmith@access1.digex.net (Eric V. Smith)
To:        comp.protocols.tcp-ip
Subject:   Re: WinNT & Windows TCP/IP connection

In article <3cjdnr$5i1@news.halcyon.com>,
Timothy D. Hall <timh@winserve.com> wrote:
>Windows NT 3.5 supports TCP/IP via both PPP and SLIP/CSLIP with its
>Remote Access Services (RAS) Server.
>
>This will work only with another Windows NT 3.5 box or a Windows 95
>box.

This is not true.  I've connected a NT 3.5 beta to a Portmaster (I
believe, you'd have to ask my provider) using CSLIP with no problems.

Eric.

-- 
Eric V. Smith           | Technology ... the knack of so arranging the world
esmith@access.digex.net |  that we need not experience it.
Windsor Software Corp   |	- Max Frisch

-----------[000306][next][prev][last][first]----------------------------------------------------
Date:      Tue, 13 Dec 1994 04:45:06 GMT
From:      peisch@cfa.org (Peter Eisch)
To:        comp.protocols.tcp-ip
Subject:   Winsock/ODI on TokenRing

Does this work?  We're trying to coax it, but with no luck.  The cisco has
SNAP on it's token ring interface, but the pc's don't seem to be able to
get out...

Ideas?

peter

-- 
peisch@cfa.org


-----------[000307][next][prev][last][first]----------------------------------------------------
Date:      13 Dec 1994 05:51:27 GMT
From:      chou@pds.nchu.edu.tw (Chou Chaw Yi)
To:        comp.protocols.tcp-ip
Subject:   Some problem in install ipmulticast in Ultirx4.2a


Hi! Dear Networkers:
  Please pardon my audacity.
  I can make the vmunix for ipmulticast in Ultrix 4.2a. But, when I 
rebooted this DEC WORKSTAION 5000/200, it was appeared the following
messages:

    ufs_gunlock : gp unlocked, dev 0x1500 gno 2
    cpu 0 panic : ufs_gunlock locks held by cpu0

And it was repeatly rebooted.

  Thank you very much.

  Merry Christmas to everybody!

Yours sincerely,
Chaur-Yi Chou
13 Dec. 1994
---------------------------------------------------
My e. m. address : cijou@spring.nchu.edu.tw
---------------------------------------------------

-----------[000308][next][prev][last][first]----------------------------------------------------
Date:      13 Dec 1994 06:07:55 GMT
From:      tdh5821@halcyon.halcyon.com (Timothy D. Hall)
To:        comp.protocols.tcp-ip
Subject:   Re: WinNT & Windows TCP/IP connection

In article <3ca9jv$34a@serra.unipi.it>, Antonello Olla <olla@ssgrr.it> wrote:

>I would know how can I connect a Windows 3.1 (or WfW 3.1 or WfW3.11) 
>PC with an WinNT ( 3.1 or 3.5 ) PC using TCP/IP protocol.
>I know there are two methods: PPP and SLIP (or CSLIP), but I don't know
>which is supported by WinNT as server.
>Can someone tell me how can I realize this connection and using what 
>programs?

Windows NT 3.5 supports TCP/IP via both PPP and SLIP/CSLIP with its
Remote Access Services (RAS) Server.

This will work only with another Windows NT 3.5 box or a Windows 95
box.
--
Timothy D. Hall   | Winserve - Windows Internet Server
timh@winserve.com | winserve.001=204.118.34.11; net view \\winserve.001
206.827.8184      | http://www.winserve.com

-----------[000309][next][prev][last][first]----------------------------------------------------
Date:      13 Dec 1994 15:14:35 -0500
From:      barmar@nic.near.net (Barry Margolin)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

In article <D0pGGF.Aq4@calcite.rhyolite.com> vjs@calcite.rhyolite.com (Vernon Schryver) writes:
>File accesses tend to jump all over the file system.

While they may sometimes jump all over the system, in practice I think most
file accesses are sequential.  Random access is common for databases and
executables, but most user files are not of these types.  They tend to have
lots of text files or word processing documents, which are usually accessed
as a whole when they're read into word processors, printed, or piped to
various filters.

While it's good for a remote file system protocol to have reasonable
random-access operations, it would be very helpful if it had optimized
operations for reading a whole file.

Note that NFS, even over UDP, doesn't preclude such optimizations.  If the
application does a read() call that asks for 32K bytes, there's no reason
the kernel can't send out four 8K NFS_Read requests concurrently, rather
than waiting for each response before issuing the next NFS_Read.
(Unfortunately, this isn't possible for NFS_ReadDir operations, because
each sequential call includes a parameter that comes from the results of
the previous call.)
-- 

Barry Margolin
BBN Internet Services Corp.
barmar@near.net

-----------[000310][next][prev][last][first]----------------------------------------------------
Date:      13 Dec 94 16:37:42 -0500
From:      harvey@indyvax.iupui.edu (James Harvey)
To:        comp.protocols.tcp-ip
Subject:   Re: Time Servers?

In article <3ckor5$im7@mojo.eng.umd.edu>, mmeltzer@glue.umd.edu (Marc Meltzer) writes:
> Are there any publicly accessible time servers on the Internet?  If so,
> what are the addresses?

Yes.  For many of them, you should obtain permission before you start using
them.  Subscribe to comp.protocols.time.ntp to get more information on how
to do this correctly.  A list of stratum 1 and 2 ntp servers on the Internet
is available by anonymous ftp from louie.udel.edu in /pub/ntp/doc/clock.txt.
At this site, we synch a couple of our servers to ntp servers on the Internet
using ntp v3, and then synch all the rest of our stuff to those local servers.
Network service providers are providing time services now as well, and that
seems to be the direction that things are heading.
--
James Harvey   harvey@iupui.edu   IUPUI IT Networks and Systems
Disclaimer:  These are my own opinions.  I do not speak for Indiana University.

-----------[000311][next][prev][last][first]----------------------------------------------------
Date:      13 Dec 1994 09:00:45 GMT
From:      Ralf Hack <hack@pluto.rz.fh-reutlingen.de>
To:        comp.protocols.tcp-ip
Subject:   Solution: Linux-Router cannot cope with Sun-tcpip of our local stations

Ralf Hack <hack@pluto.rz.fh-reutlingen.de> wrote
> we setup a linux-router for our local network. With the local suns,
> we cannot route files (via ftp) above roughly 1k. However, we can
> receive them with our router. The same for xsessions. Routing sessions
> with netscape/mosaic .. does not work. However, routing an xterm-inal
> does work.
> 
> We use a 486/33 PC with Linux 1.1.23. Our subnet consists of SGI's and
> PC's. The suns are running SunOS 5.3 Generc sun4m sparc.
> 
Since linux cannot cope with fragmentation of packets, all I have to 
do  is to setup the mtu with ifconfig to 2000. Now it works. The suns
use some kind of fragmentation test, which allows them to change 
fragmentation according to the available net. A very interesting thing.

Thanks to everyone who respond. You were very helpful

Greetings from Germany
Ralf Hack


-----------[000312][next][prev][last][first]----------------------------------------------------
Date:      Tue, 13 Dec 1994 18:25:50 -0500
From:      Lyle_Seaman@transarc.com
To:        comp.protocols.nfs,comp.protocols.tcp-ip
Subject:   Re: NFS/SLIP

vjs@calcite.rhyolite.com (Vernon Schryver) writes:
> Over a correctly installed and operated Ethernet, NFS/TCP simply cannot
> be faster than NFS/UDP when UDP is built to go fast.  The best you can
> hope for is that your TCP congestion window will eventually open up all
> of the way and each bulk transfer will be blasted out exactly as fast
> as all NFS/UDP transfers.

"Full speed ahead and damn the torpedoes!"

I wonder if that's necessarily true.  I mean, I agree that it's
probably true if TCP is unmodified, and you're using the VJ congestion
control algorithms with switched ethernets, and so on.  But if you're
sharing the ethernet, surely the congestion caused by NFS/UDP will
significantly reduce the available bandwidth.  

VJ's slow-start was designed with certain goals in mind (namely,
speeding up FTP).  The criteria he used don't exactly apply to
distributed filesystems, but the principles do.  



-----------[000313][next][prev][last][first]----------------------------------------------------
Date:      13 Dec 1994 19:48:23 -0500
From:      msbeebe@mtu.edu (Matt Beebe)
To:        comp.protocols.tcp-ip
Subject:   Re: WinNT & Windows TCP/IP connection

Eric V. Smith (esmith@access1.digex.net) wrote:
: In article <3cjdnr$5i1@news.halcyon.com>,
: Timothy D. Hall <timh@winserve.com> wrote:
: >Windows NT 3.5 supports TCP/IP via both PPP and SLIP/CSLIP with its
: >Remote Access Services (RAS) Server.
: >
: >This will work only with another Windows NT 3.5 box or a Windows 95
: >box.
: 
: This is not true.  I've connected a NT 3.5 beta to a Portmaster (I
: believe, you'd have to ask my provider) using CSLIP with no problems.
: 
: Eric.

Actually, I think he's right, just unclear...

   NT3.5 should connect to any ppp service provider, but I don't think 
   (somebody please correct me if I'm wrong) that you can dial INTO an
   NT3.5 Box with anything other than WFWG (or W95) RAS setup on the
   remote end.

			-Matt


-----------[000314][next][prev][last][first]----------------------------------------------------
Date:      13 Dec 1994 11:46:32 GMT
From:      jim@cs.strath.ac.uk (Jim Reid)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

In article <D0I0nM.94D@calcite.rhyolite.com> vjs@calcite.rhyolite.com (Vernon Schryver) writes:

   In article <3c6lbf$jnu@cronkite.cisco.com> tli@cisco.com (Tony Li) writes:
   > ...
   >Second problem: many NFS implementations disable UDP checksums.  SLIP
   >provides no data integrity check.  Your data may be getting corrupted
   >too...  what little of it that gets there...

   Please name a common, current system with UDP checksums off by default.

SunOS


-----------[000315][next][prev][last][first]----------------------------------------------------
Date:      13 Dec 1994 14:46:44 GMT
From:      lapp@waterloo.hp.com (David Lapp)
To:        comp.protocols.tcp-ip
Subject:   Re: DHCP Info ???

John Skovron (skov@ilx.com) wrote:

: I checked my saved copy of the "report on 4th DHCP backoff" from
: the DHCP mailing list (host-conf@sol.cs.bucknell.edu), and it says
: that HP only has a client implementation.
 
: DHCP server implementations can be had from Sun, WIDE (whatever
: that is), Microsoft, FTP, Competitive Automation (whatever that
: is), and SGI, according to the same message.

As the author of that report I'd like to point out that
it wasn't intended to be a definitive list of implementations
nor a comment on their availability. The implementations listed
were those that attended the bakeoff. I understand that there
are a number of other implementations being worked on and
perhaps available. Some of the implementations that were
tested at the bakeoff may not yet be available.

I have to admit to being a little confused about WIDE myself :-)
So I'll just say that I beleive that they are a research group.

Competitive Automation is a company whose products include
DHCP implementations.

Dave Lapp

Standard Disclaimer: I speak for myself not HP.

-----------[000316][next][prev][last][first]----------------------------------------------------
Date:      Tue, 13 Dec 94 15:52:53 GMT
From:      srl@terminus.ericsson.se (Steve Langstaff)
To:        comp.protocols.tcp-ip
Subject:   RIP suspect operation, ideas?

Hi all.

I am evaluating an implementation of RIP and have come across a
strange situation involving 3 machines running RIP over a SLIP
and a LAN interface...

Box 1 - running the evaluation routed
Box 2 - running the evaluation routed
sparc - running SunOS/routed


 ---- .1     Net2         .2 ----
|Box1|----------------------|Box2|
 ----                        ----
  |.160                       |.162
--+--------------+------------+--------Net1
                 |.207
               -----
              |sparc|
               -----

Net1 192.9.200 mask 255.255.255.0 (ether)
Net2 192.10.1  mask 255.255.255.0 (point to point)


After running RIP on all nodes for a while, the routing tables for
the nodes are as follows:

Box1:

DEST                GATE                TYPE
--------------------------------------------
192.9.200.0         192.9.200.160       INTERFACE
192.10.1.1          192.9.200.162       HOST
192.10.1.2          192.10.1.1          INTERFACE (POINT TO POINT)

Box2:

DEST                GATE                TYPE
--------------------------------------------
192.9.200.0         192.9.200.162       INTERFACE
192.10.1.1          192.10.1.2          INTERFACE (POINT TO POINT)
192.10.1.2          192.9.200.160       HOST

sparc:

Destination          Gateway              Flags    Interface
192.10.1.1           192.9.200.162        UGH      le0
192.10.1.2           192.9.200.160        UGH      le0
192.9.200.0          192.9.200.207        U        le0

As can be seen, the routing tables have settled down to give the
somewhat strange situation that to get to 192.10.1.1 (Box1/slip0),
both Box1 and sparc should send datagrams to Box2, and vice-versa
for Box2. This is silly, since this means that all datagrams to the
ends of the SLIP link will have to traverse the link, even if the
destinations are local.

Is this a known situation that RIP can get routers into, or is this
a fault with the implementation of RIP? I don't think that routers
should normally advertise reachability of local interface addresses
(such as Box1 advertising 192.10.1.1) but in this situation it might
lead to these non-optimal routes being eliminated.

---

Steve L.


-----------[000317][next][prev][last][first]----------------------------------------------------
Date:      14 Dec 1994 00:20:41 -0500
From:      barmar@nic.near.net (Barry Margolin)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

In article <D0rvH0.J93@calcite.rhyolite.com> vjs@calcite.rhyolite.com (Vernon Schryver) writes:
>Agreed, the string of file system requests from single user process for
>a single file tend to be pretty sequential. ... The requests from a bunch of
>user processes are pretty random, which is why disk driver scheduling
>algorigthms are as they have been for the last 20 or 30 years.  

Those algorithms were designed for multiuser timesharing systems.  For
single-user systems, there's probably not so much interleaving of file
accesses that you need fancy support for it.  Since Unix tends to be used
on both timesharing systems and personal workstations, this indicates the
need for tuning facilities that change the optimization priorities.

>>Note that NFS, even over UDP, doesn't preclude such optimizations.  If the
>>application does a read() call that asks for 32K bytes, there's no reason
>>the kernel can't send out four 8K NFS_Read requests concurrently, rather
>>than waiting for each response before issuing the next NFS_Read.
>
>As I keep saying, some NFS vendors do (or did a few years ago) what some
>of us call "aggressive read aheads."  (4 is not very aggressive, being
>only 2 more than normal UNIX read-ahead.)  

I was explicitly *not* describing read-ahead.  Read-ahead is when the OS
requests data from the disk or server that the application hasn't asked
for, in anticipation of the need for it.  In my example, the application
only asked for 32K bytes, perhaps because that's how big the file was; if
the file were 1MB it would have asked for that much, perhaps resulting in
128 concurrent NFS_Read requests if the client has enough buffer space.

The point I was making is that the OS can satisfy such requests by issuing
the RPC's sequentially or in parallel.  Parallel NFS_Read calls can be used
to implement a form of sliding windows (but the same can't be done with
NFS_Write calls, because there's no way for the server to tell the client
what its "window" is).

>All such optimizations are bets that the next requests from the
>applications (plural!) in the system will be for the rest of the file
>(singular!).  

In my example there was no betting involved.  The application asked for
32K, and that's what the kernel requested from the server.
-- 

Barry Margolin
BBN Internet Services Corp.
barmar@near.net

-----------[000318][next][prev][last][first]----------------------------------------------------
Date:      Tue, 13 Dec 1994 16:16:18 GMT
From:      vjs@calcite.rhyolite.com (Vernon Schryver)
To:        comp.protocols.tcp-ip,comp.dcom.lans
Subject:   Re: packets sent to own address

In article <3cie85$md6@tools.near.net> barmar@nic.near.net (Barry Margolin) writes:

> ...
>>What happens exactly to a packet which is sent to it's own address --> 
 
> ...
>In most Unix implementations I believe the IP layer detects that the
>destination address is one of the local addresses, and simply moves the
>packet from the send queue to the receive queue.  I don't think it ever
>sends it to the interface driver.

In the 4.3BSD based code I've seen, the IP code does a route lookup and
notices a local address, and then sends the packet through the loopback
driver.  The interface structure of the loopback driver is used to
determine parameters for the higher layer protocols such as MTU (and so
TCP MSS).  The loopback driver tends to just shuffle queue pointers,
but in a system that has hardware doing TCP and UDP checksums, it may
do a little more.


Vernon Schryver    vjs@rhyolite.com

-----------[000319][next][prev][last][first]----------------------------------------------------
Date:      Tue, 13 Dec 1994 16:53:54 GMT
From:      vjs@calcite.rhyolite.com (Vernon Schryver)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

In article <3cii3c$apn@tribune.hri.com> tmt@hri.com (Thomas M. Talpey) writes:

> ...
>I did the NFS/TCP implementation in stock OSF/1, which is similar to
>Rick Macklem's Guelph TCP but was developed independently. On OSF/1,
>as with other NFS/TCP implementations at Connectathon, I found NFS/TCP
>to be up to 15% faster than NFS/UDP. The largest gain is in bulk data
>transfer, where the TCP protocol really shines through. ...

A difference of 15% on the Connectathon tests is far below the noise
threshold of that benchmark.  When taken over "other implementations"
anything that is within about 100% counts as "the same." To put it
another way, consider the enormous variation among LADDIS numbers for
various implementations.

A 15% difference for a single implementation resulting from only switching
between TCP and UDP would be interesting, but almost certainly only for
that implementation.  Again, variations among implementations dwarf 15%
by an order of magnitude.  For example, the HP and Digital reports here
of 5MByte/sec are more than 10 times larger than classical numbers for
UNIX workstations, and perhaps 100 times larger than a typical PC with
a typical junk Ethernet card using typically wrong parameters for the
junk (e.g. rsize and wsize=8K).

Depending on how your TCP and UDP code works, either can be faster on
a pure transfer benchmark such as `ttcp`.  For example, I know of systems
for which UDP on a bound socket using write() or send() is faster than
TCP, and TCP is faster than UDP on an unbound socket using sendto().
(Besides the route look-up for each sendto() that is not needed for
send(), there can be mutual-exclusion issues in multi-processors.)

Over a correctly installed and operated Ethernet, NFS/TCP simply cannot
be faster than NFS/UDP when UDP is built to go fast.  The best you can
hope for is that your TCP congestion window will eventually open up all
of the way and each bulk transfer will be blasted out exactly as fast
as all NFS/UDP transfers.


Vernon Schryver    vjs@rhyolite.com

-----------[000320][next][prev][last][first]----------------------------------------------------
Date:      13 Dec 1994 18:23:33 GMT
From:      mmeltzer@glue.umd.edu (Marc Meltzer)
To:        comp.protocols.tcp-ip
Subject:   Time Servers?

Are there any publicly accessible time servers on the Internet?  If so, 
what are the addresses?

Thanks

----------------------------------------------------------------------
Marc Meltzer                                     mmeltzer@glue.umd.edu
Department of Computer Science		   U. Maryland at College Park
----------------------------------------------------------------------


-----------[000321][next][prev][last][first]----------------------------------------------------
Date:      Tue, 13 Dec 1994 18:34:39 GMT
From:      vjs@calcite.rhyolite.com (Vernon Schryver)
To:        comp.protocols.tcp-ip,comp.dcom.lans.fddi,comp.dcom.lans.ethernet
Subject:   Re: Max. # of routers on a network?

In article <3ck7gh$436@lxotta.learnix.ca> andy@learnix.ca (Andy Barclay) writes:
>There is one other caveat about using 30 routers.
>Remeber that RIP defines a route as unreachable if its more than
>15 hops away.

That is true but irrelevant to the original question and to all likely
topologies involving only 30 RIP routers.  The original question involved
30 routers on a single network, connecting ohter networks that would be
only 2 hops away from each other.  With only 30 routers, you'd expect
the maximum hop count to be less than 6.

As I recall, it was not that long ago that the Internet TTL passed 30,
at which time there were quite a few more than 30 routers (of course,
not using RIP).


Vernon Schryver    vjs@rhyolite.com

-----------[000322][next][prev][last][first]----------------------------------------------------
Date:      13 Dec 1994 19:14:19 GMT
From:      garryh@seeding.apple.com (Garry Hornbuckle)
To:        comp.sys.mac.comm,comp.protocols.tcp-ip,comp.sys.mac.misc
Subject:   Re: [A] How to make a virgin copy of MacTCP!

Today (12/13)I have posted a variety of additional material in regards to
the MacTCP 2.0.6 update proces in response to your feedback.

* A MacTCP 2.0.2 -> 2.0.6 updater. 

  This will allow those of you who have MacTCP 2.0.2 masters to upgrade
  directly to 2.0.6 in a single step, i.e., without applying the 2.0.4 
  updater first.

* A MacTCP Admin 2.0.2 -> MacTCP 2.0.6 Admin updater.

  This was entirely our oversight. My apoligies, but here ya' go...

* Revised README

  These notes explain the update process for System 7.5 users, plus
  explain our choice of the ResCompare applications, and our decisions to
  require a virgin copy of the software.

I hope that this will make your upgrade to MacTCP 2.0.6 easier.

*** SYSTEM OPERATORS WHO HAVE REPOSTED PRIOR VERSIONS OF THIS UPDATER
    SHOULD GET THE NEW PACKAGE - ESPECIALLY THE README - ASAP. THE PROPER
    README IS DATED 12/13/94. NOTE THAT THE '.sea' FILE HAS ALSO CHANGED!

Best Regards

-------------------------------------------------------------------
Garry Hornbuckle    Product Manager, Communications & Collaboration
-------------------------------------------------------------------
"If I told you that I   | email      garryh@seeding.apple.com
 spoke only for myself  | applelink  HORNBUCKLE1
 would you believe me?" | fax        (408) 974-1211
-------------------------------------------------------------------

-----------[000323][next][prev][last][first]----------------------------------------------------
Date:      Tue, 13 Dec 1994 20:38:43 GMT
From:      mgamble@esys.co.uk (Mark Gamble)
To:        comp.protocols.tcp-ip
Subject:   Re: Interpretation of Nagle algorithm

In article <3ci1c9$9ht@crab.spider.co.uk> Ian Heavens says:
>
>I'm looking for clarification of the Nagle algorithm as described in
>RFC 896 and RFC 1122.  It concerns behaviour when a buffer is
>queued to TCP for transmission which exceeds the segment size (say 1.5
>x segment size), when there are no unacknowledged segments.
>
>The first full size segment is transmitted, according to Nagle; what
>about the second?  I'm told that the correct interpretation is that
>it should be sent, even though the 2nd segment is not full size,
>and there is unacknowledged TCP data.  
>
Don't think you should agonise about this too much - Nagle is there to
prevent congestion due to tinygrams, and the situation you're looking at
isn't going to do that.  On the nitty level, the key section is probably:

"The solution is to inhibit the sending of new TCP  segments  when
new  outgoing  data  arrives  from  the  user  if  any previously
transmitted data on the connection remains unacknowledged.   This
inhibition  is  to be unconditional; no timers, tests for size of
data received, or other conditions are required.   Implementation
typically requires one or two lines inside a TCP program."

You could argue this either way; your user data would generate two
segments, so is the second treated the same way as the first (as part of
'new outgoing data') or not?  Toss a coin.

I'd go for sending them both, but even if you hold the second segment
you'll most probably transmit it when the first is acked.

HTH

Mg

-----------[000324][next][prev][last][first]----------------------------------------------------
Date:      Tue, 13 Dec 1994 20:40:47 GMT
From:      mgamble@esys.co.uk (Mark Gamble)
To:        comp.protocols.tcp-ip
Subject:   Re: TCP/IP for msdos?

In article <D0Eot6.707@itex.jct.ac.il>, jacobsen@itex.jct.ac.il (Joel Jacobsen) says:
>
>Is there any good free or even Shareware TCP/IP package for msdos?
>
Try KA9Q from (amongst others) ftp.demon.co.uk in pub/ibmpc/something...

Mg

-----------[000325][next][prev][last][first]----------------------------------------------------
Date:      Tue, 13 Dec 1994 20:55:11 GMT
From:      jwa2n@galen.med.Virginia.EDU (James W. Adams)
To:        comp.protocols.ibm,comp.protocols.tcp-ip
Subject:   VM/SMTP V2R2 mail forwarding via TCP/IP?

How can one forward incoming email from VM/SMTP V2R2 under VM/ESA to an
Internet address?  The SMTP NAMES file seems to be for IBM-style RSCS
addresses.  Is there a way to do this (in the fashion of the UNIX
sendmail aliases or .forward features) without additional third-party
software?

--
  James W. Adams                        Medical Center Computing
  Box 512, HSC                          jwa2n@virginia.edu
  University of Virginia                
  Charlottesville, VA  22908            /* no comment */

-----------[000326][next][prev][last][first]----------------------------------------------------
Date:      Tue, 13 Dec 1994 21:27:49 +0000
From:      Andrew@cucumber.demon.co.uk (Andrew Gabriel)
To:        comp.protocols.tcp-ip
Subject:   Re: PPP via telnet...

In article <nac.72.2EE51728@sirius.com> nac@sirius.com "Nancy Cedeno" writes:

> In article <3buq21$mop@nyheter.chalmers.se> thed@mtek.chalmers.se (Thed Lyngmo)
>  writes:
> 
> >: Is it possible to run a ppp (or slip for that matter) connection over a 
> >: telnet connection?  
 
> >I only know how SLIP works and I can't see why that shouldn't be possible.
> >I guess the same goes for PPP.
 
> >What do you want to achieve by doing that?
> 
> I think it would be quite useful if you're stuck behind a firewall, and the 
> only way you can get out is via a telnet session to a host server.  
> 
> That's the setup we have at work right now.  The only machine you can ftp from 
> is that firewall server.  If we could run PPP from our MS Windows machines to 
> that host, we could use something like ws_ftp and wouldn't need to deal with 
> the middleman.

Do bear in mind that this can breach the firewall. When you have PPP running
to make access to external systems, you have also given the net access to
your internal system(s). You will need to consider very carefully what
daemons you run on your system and how the firewall is configured.
For example, if you are not careful, anyone on the net can telnet through
your PPP link, bypassing the security in your firewall, and you will very
likely be totally unaware.

Sadly, convenience often costs security.
-- 
Andrew Gabriel                        Home: Andrew@cucumber.demon.co.uk
Consultant Software Engineer          Work: Andrew.Gabriel@gpt.co.uk

-----------[000327][next][prev][last][first]----------------------------------------------------
Date:      Tue, 13 Dec 1994 22:39:45 GMT
From:      ussw@netcom.com (Don Dunstan)
To:        comp.protocols.tcp-ip
Subject:   FREE ftp utilities

Two free ftp utilities, DOS operable, are available by
anonymous ftp to ftp.netcom.com (pub/us/ussw/free).

-----------[000328][next][prev][last][first]----------------------------------------------------
Date:      13 Dec 1994 22:48:32 GMT
From:      debiso@ibm.net
To:        comp.protocols.tcp-ip
Subject:   Need DNS Software ASAP!!

HELP!!!

I need to get some DNS software for SCO unix asap!!!  I'm setting up and internet connection and my provider neglected to tell me I would need this!
Please E-Mail response if possible.

Thanx!
Joe DeBiso


-----------[000329][next][prev][last][first]----------------------------------------------------
Date:      Tue, 13 Dec 1994 23:19:48 GMT
From:      vjs@calcite.rhyolite.com (Vernon Schryver)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: NFS/SLIP

In article <3ckvbb$nac@tools.near.net> barmar@nic.near.net (Barry Margolin) writes:
>In article <D0pGGF.Aq4@calcite.rhyolite.com> vjs@calcite.rhyolite.com (Vernon Schryver) writes:
>>File accesses tend to jump all over the file system.
>
>While they may sometimes jump all over the system, in practice I think most
>file accesses are sequential.  Random access is common for databases and
>executables, but most user files are not of these types. ...

Agreed, the string of file system requests from single user process for
a single file tend to be pretty sequential.   This axiom is even older
than the approximate 3:1 reads vs. writes ratio or the fact that almost
all user files are tiny but there are a lot of bytes in the few large
files.  Unfortunately, except for the read:write ratio, these facts
are of limited use to a network file system.  The requests from a single
user process tend to jump around somewhat, if only because applications
often use more than one file at a time.  The requests from a bunch of
user processes are pretty random, which is why disk driver scheduling
algorigthms are as they have been for the last 20 or 30 years.  A network
file system for a real client operating system (i.e. not DOS) is like
the disk file system for the same client, and subject to the same
apparent randomness. 


>While it's good for a remote file system protocol to have reasonable
>random-access operations, it would be very helpful if it had optimized
>operations for reading a whole file.
>
>Note that NFS, even over UDP, doesn't preclude such optimizations.  If the
>application does a read() call that asks for 32K bytes, there's no reason
>the kernel can't send out four 8K NFS_Read requests concurrently, rather
>than waiting for each response before issuing the next NFS_Read.

As I keep saying, some NFS vendors do (or did a few years ago) what some
of us call "aggressive read aheads."  (4 is not very aggressive, being
only 2 more than normal UNIX read-ahead.)  Those vendors do quite well
on simplistic benchmarks of the form `cp /net/host/foo/bar /dev/null`,
but gratifyingly poorly (for competing salescritters) on more
representative customer applications.  The result of overly aggressive
read-ahead is worse than if you make your local disk file system always
suck up an entire file into the buffer cache every time a file is opened.

All such optimizations are bets that the next requests from the
applications (plural!) in the system will be for the rest of the file
(singular!).  If the bet pays off, you have saved the latency of the
operations, but if bet is wrong, you have wasted the network or disk
bandwidth and slowed down the transfers that the applications really
want.  If you have many applications to run, the payoff from the bet
can be zilch; you might be able to run other applications while one is
waiting for the file system.

In a network file system on a local area network, the best case payoff
from a read-ahead bet is at most the latency of an operation.  The
potential loss is the lost bandwidth.  Many networks are now much slower
than disks (e.g. few networks deliver 10 or 100 MBytes/sec), so that
the potential loss is much worse for a network file system.  If the cost
to read 1 sector from a disk is X, then the cost to read 50 sectors is
often still approximately X, because X is so dominated by seek and search
time not transfer time.  On a network, the cost to read extra data is
much larger compared to the cost to read one byte.  Log based file
systems are interesting on RAID's, but (I hope) no one is proposing them
for Ethernet based network file systems.

The familiar UNIX read-ahead mechanism for disk files can be used with
network files.  Unfortunately, it usually is not at aggressive enough.
The trick is to figure out a way to detect reference patterns that make
the read-ahead bet pay off often enough to keep the cost of bad bets
from killing system performance.


>(Unfortunately, this isn't possible for NFS_ReadDir operations, because
>each sequential call includes a parameter that comes from the results of
>the previous call.)

Few directories are (or should be) so large that matters.  The problem
with directories are the getattr's for individual files, not fetching
the big glob of all of the names.


Vernon Schryver    vjs@rhyolite.com

-----------[000330][next][prev][last][first]----------------------------------------------------
Date:      Wed, 14 Dec 1994 00:23:22 GMT
From:      sjaque@matta.die.udec.cl (Sandra Jaque)
To:        comp.protocols.tcp-ip
Subject:   Help !!! Wich port is wich

Hello, 

	I'm looking for some ports of services under TCP/IP  
that doesn't appear in my /etc/services file, please if someone 
know how or where can I get that information, send me a mail to 
sjaque@manet.die,udec.cl because is most reliable.  

For example I need to know which services use the port 
number 882,959,733,690,960,683,949,70 and so on.


Thanks you !!!


-----------[000331][next][prev][last][first]----------------------------------------------------
Date:      Wed, 14 Dec 1994 00:59:11 GMT
From:      cd4v@mamba.cs.Virginia.EDU (Christian  Dreke)
To:        comp.protocols.tcp-ip
Subject:   Re: Wher can I find a TCP/IP source code


Check out the FreeBSD implementation of TCP/IP. It uses Sockets.
It's available at ftp.cdrom.com 
under something like FreeBSD/FreeBSD-current/src/netinet (may not be
the exact directory).

Chris.

-----------[000332][next][prev][last][first]----------------------------------------------------
Date:      Wed, 14 Dec 1994 01:42:06 GMT
From:      vjs@calcite.rhyolite.com (Vernon Schryver)
To:        comp.protocols.nfs,comp.protocols.tcp-ip
Subject:   Re: NFS/SLIP

In article <0ivWpyGSMUgE4NvIFK@transarc.com> Lyle_Seaman@transarc.com writes:
>vjs@calcite.rhyolite.com (Vernon Schryver) writes:
>> Over a correctly installed and operated Ethernet, NFS/TCP simply cannot
>> be faster than NFS/UDP when UDP is built to go fast.  The best you can
>> hope for is that your TCP congestion window will eventually open up all
>> of the way and each bulk transfer will be blasted out exactly as fast
>> as all NFS/UDP transfers.
>
>"Full speed ahead and damn the torpedoes!"
>
>I wonder if that's necessarily true.  I mean, I agree that it's
>probably true if TCP is unmodified, and you're using the VJ congestion
>control algorithms with switched ethernets, and so on.  But if you're
>sharing the ethernet, surely the congestion caused by NFS/UDP will
>significantly reduce the available bandwidth.  
>
>VJ's slow-start was designed with certain goals in mind (namely,
>speeding up FTP).  The criteria he used don't exactly apply to
>distributed filesystems, but the principles do.  


That is wrong on several important counts.

Van Jacobson's congestion control, avoidance, and recovery algorithms
were and are designed for wide area networks, not local area networks.
The timers in his code are orders of magnitude too coarse for a local
area network of Ethernets.  What sense would it make to expect timers
with resolutions of 500 or 1000 milliseconds to do any good on a network
with round trip times of less than 5 milliseconds?

A properly running Ethernet loses essentially no packets.  Van Jacobson's
congestion avoidance mechanisms depend on noticing lost packets.  There
is no signal for the congestion avoidance mechanisms to use, because
there are no lost packets that signal congestion on a properly running
Ethernet, no matter how congested the Ethernet is.  (Of course, when an
Ethernet "collapses", you start losing packets, but transmission latencies
are intolerable a long time before that happens.)

Ethernets work fine because they have collisions to negotiate sharing
the bus.  On a properly running Ethernet, you cannot do any better than
queuing all of your packets on your MAC chip as soon as you have them,
and then letting the Ethernet protocol negotiate their delivery.

Using NFS/UDP does not and cannot cause any more or less congestion
using NFS/TCP, provided the NFS clients are offering the same load.
The contention resolution of the Ethernet protocol, "collisions," is
designed to deal with that congestion, and it does quite well.

It is not wise to put many NFS clients on a single Ethernet collision
domain, but that has nothing to do with UDP or TCP.  An NFS stream is
at least 250KByte/sec when running, and you cannot put many such streams
on an Ethernet at one time, regardless of whether you use TCP or UDP.


Vernon Schryver    vjs@rhyolite.com

-----------[000333][next][prev][last][first]----------------------------------------------------
Date:      14 Dec 94 02:19:16 GMT
From:      mferan@ego.psych.mcgill.ca (Shelly Feran)
To:        alt.sys.pdp11,comp.protocols.tcp-ip
Subject:   ftp for RT11, ver. 4.0 - 5.x??

The last of our 11/23s will be phased out of the data collection 
business soon - and we need to move a gig. or so of data from it
to a other machines. Does anyone know where we can get a hold of
tcp software that supports the 11's ethernet interface (which we
did locate, amazingly enough...)

thanks
Shelly Feran
Dept. of Psychology, Mcgill University


-----------[000334][next][prev][last][first]----------------------------------------------------
Date:      14 Dec 1994 02:58:16 GMT
From:      raj@cup.hp.com (Rick Jones)
To:        comp.protocols.tcp-ip
Subject:   Re: Interpretation of Nagle algorithm

I thought that Nagle also tended to rear his/its head in situations
where say someone presented 1500 bytes of data to a transport (assume
Ethernet, with a 1460 byte MSS). I've seen people have interesting
interactions between that, and the standalone ACK timer - especially
in places where there was a mismatch in the socket buffer sizes on
either end (say a PC and a Unix server)

I've been wondering if maybe we should re-define the criteria employed
in the implementation of the Nagle algorithm. My understanding was
that Nagle was intended to keep the ratio of data to headers
"reasonably" large. Presently, "reasonable" is defined (well, from
what I see in HP-UX/BSDish networking) as a fraction of the MSS. Back
when "Nagle" first started, MSSs didn't get much larger than 1460
bytes, so you could expect that Nagle would get you a ratio of headers
to data of say 40/730 (or whatever the fraction might be). However,
today, the MSSs are getting much larger - 4096 bytes (FDDI), 8192
bytes (ATM), 16384-32768 (FibreChannel). Maybe it is time to
add/change the critia to be "Do we have enough data to give us 12%
(6%) overhead?" instead of "Do we have data enough for 1/2 the MSS?"

I think that this would help with some of the problems that people
have experienced with ATM and needing/wanting to enable TCP_NODELAY.

Comments? Already done?

rick jones

-----------[000335][next][prev][last][first]----------------------------------------------------
Date:      14 Dec 1994 12:38:39 -0500
From:      lin@cs.purdue.EDU (John Chueng-Hsien Lin)
To:        comp.protocols.tcp-ip
Subject:   Re: Interpretation of Nagle algorithm

> I've been wondering if maybe we should re-define the criteria employed
> in the implementation of the Nagle algorithm. My understanding was
> that Nagle was intended to keep the ratio of data to headers
> "reasonably" large. Presently, "reasonable" is defined (well, from
> what I see in HP-UX/BSDish networking) as a fraction of the MSS. Back
> when "Nagle" first started, MSSs didn't get much larger than 1460
> bytes, so you could expect that Nagle would get you a ratio of headers
> to data of say 40/730 (or whatever the fraction might be). However,
> today, the MSSs are getting much larger - 4096 bytes (FDDI), 8192
> bytes (ATM), 16384-32768 (FibreChannel). Maybe it is time to
> add/change the critia to be "Do we have enough data to give us 12%
> (6%) overhead?" instead of "Do we have data enough for 1/2 the MSS?"
> 
> I think that this would help with some of the problems that people
> have experienced with ATM and needing/wanting to enable TCP_NODELAY.
> 
> Comments? Already done?

We have observed a circular-wait behavior that decreases TCP
throughput over ATM dramatically. The interaction of the following
items causes the anomalous behavior:

1. TCP send buffer size
2. TCP receive buffer size
3. ATM interface's MTU
4. TCP Maximum segment size (MSS)
5. The way user data is added to the TCP send buffer.
6. Sender's side Silly Window Syndrome (SWS) avoidance (Nagle's
   Algorithm)
7. Receiver side delayed ACKs.

Here is the abstract of the paper: (The paper also explains how SunOS
4.1.1 TCP implements Nagle's algorithm.)

-------------------------------------------------------------------
Title: TCP Buffering And Performance Over An ATM Network
       Douglas E. Comer and John C. Lin
       Purdue University
        
This paper reports a series of experiments to measure TCP performance
when transferring data through an Asynchronous Transfer Mode (ATM)
switch.  The results show that TCP buffer sizes and the ATM interface
maximum transmission unit have a dramatic impact on throughput.  We
observe a throughput anomaly in which an increase in the receiver's
buffer size decreases throughput substantially.  For example, when
using a 16K octet send buffer and ATM Adaptation Layer 5 on a 100
megabit per second (Mb/s) ATM path, the mean throughput for a bulk
transfer drops from 15.05 Mb/s to 0.322 Mb/s if the receiver's buffer
size is increased from 16K octets to 24K octets.  This paper analyzes
the performance, explains the anomalous behavior, and describes a
solution that prevent the anomaly from occurring.
-------------------------------------------------------------------
ftp://gwen.cs.purdue.edu/pub/lin/TCP.atm.ps.Z

John Lin




-----------[000336][next][prev][last][first]----------------------------------------------------
Date:      14 Dec 94 13:19:50 PST
From:      glittle@bcsc02.gov.bc.ca (Glen Little)
To:        comp.protocols.tcp-ip
Subject:   Telnet for 3270?

Does anyone know of a Windows based Telnet for 3270 program? Preferably
free or share-ware, but I'll consider commerical ones.

------------------------------------------------------------------------------
     Glen Little  //  glittle@bcsc02.gov.bc.ca  //  Vancouver, BC Canada
                       *** I speak only for myself! ***
            No longer "fishers of men" but "quickeners of mankind".



-----------[000337][next][prev][last][first]----------------------------------------------------
Date:      14 Dec 1994 06:52:37 GMT
From:      charap@seas.gwu.edu (Charalambos Pashiardis)
To:        comp.protocols.tcp-ip
Subject:   TCP/IP state machine

I am looking for information that will help me write the TCP state machine for
several TCP functions.

Please e-mail at: charap@seas.gwu.edu

Haris.

ps. you can refer me to any books you have in mind or any ftp locations or
    anything else.

:

-----------[000338][next][prev][last][first]----------------------------------------------------
Date:      Wed, 14 Dec 94 14:09:22 PDT
From:      Craig Nelson <craign@teleport.com>
To:        comp.protocols.tcp-ip,comp.os.ms-windows.programmer.networks
Subject:   IPX Packet driver


Hello, one and all.

I'm looking for an IPX-TCP/IP packet driver over a SLIP link. That is,
I have a SLIP link I can connect with. Thus I have TCP/IP. I would like
an IPX packet driver to lay on top of it that will take IPX packets and
"cook" them into TCP/IP to send them on their way.

Is there any way of doing this. Sorry if it is the wrong group.

Thanks,

Craig


-----------[000339][next][prev][last][first]----------------------------------------------------
Date:      Wed, 14 Dec 1994 07:55:12 +0000
From:      Andrew@cucumber.demon.co.uk (Andrew Gabriel)
To:        comp.protocols.tcp-ip
Subject:   Re: TCP RFC 793 to BERKLEY SOCKETS correlation  -  post.fil [1/1]

In article <3cieob$3lf@newhub.xylogics.com>
           carlson@xylogics.com "James Carlson" writes:
[snip]

> The TCP_NODELAY flag turns off the Nagle algorithm.  As far as I'm able
> to tell, it has nothing to do with the PUSH bit in any implementation.
> 
> At least in the BSD4.3 code, PUSH is set when the data being sent by TCP
> fits entirely within the window and there's no more to send.  That is to
> say, it does it automatically by default.
 
>                                            (The comments in the code say
> that this is done to work around bugs in implementations which need to
> see either a full buffer or a PUSH before any data are delivered to the
> application level.)
> 

I havn't looked at the BSD source, but I have to say this sounds wrong
to me. An implementation is only *required* to forward data to the
application when:
1.   PSH is set;
2.   enough data is received to fill the application's read buffer (no
     point waiting for any more data);
3.   FIN is set (no point waiting for any more data);
4.   Possibly if URG is set and PSH isn't (depends how URG is signalled).

An implementation might also choose to forward data to the application
at other times convenient to it (such as an IP segment boundary) if the
API allows, but it is not *required* to do so.

I have come across applications which don't set PSH when they should,
e.g. at the end of some data which is to elicit an application level
response. Funnily enough, rcp had this problem back in SunOS 3 days,
(and I presume BSD's rcp at that time also). Maybe the behaviour of the
BSD TCP implementation (if correctly reported) was an attempt to get
round such buggy applications?

-- 
Andrew Gabriel                        Home: Andrew@cucumber.demon.co.uk
Consultant Software Engineer          Work: Andrew.Gabriel@gpt.co.uk

-----------[000340][next][prev][last][first]----------------------------------------------------
Date:      Wed, 14 Dec 94 15:47:11 PDT
From:      kld@mudshark.sunquest.com
To:        comp.unix.programmer,comp.unix.aix,comp.unix.admin,comp.protocols.tcp-ip
Subject:   Encina Support


Hello Encina/DCE Support People --

I'm looking for someone to teach me how to:
  setup
  configure
Encina/DCE on an IBM RISC/6000 under AIX 3.2.5.
I don't have the time to learn it from scratch.
I've nearly finished installing it and would like
to spend next week doing the setup and configuration.

If you have substatial experience with Encina and would 
like a consulting position for a week in Salt Lake City, 
Utah (during ski season) please contact:

                      Karen Dickerson
                         (801)277-6946
                         kld@mudshark.sunquest.com



-----------[000341][next][prev][last][first]----------------------------------------------------
Date:      14 Dec 1994 10:22:37 GMT
From:      "Dr. Franz Pucher" <fpucher@iaik.tu-graz.ac.at>
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.domains,comp.protocols.tcp-ip.ibmpc,comp.protocols.iso
Subject:   XTP link

Hi!

Could someone mail me a link to an XTP (Xpress Transfer Protocol) site
(WWW, ftp ...)

Please send mail to fpucher@iaik.tu-graz.ac.at

Thank you in advance!

Ciao, F. Pucher



-----------[000342][next][prev][last][first]----------------------------------------------------
Date:      14 Dec 94 20:42:24 -0500
From:      yhsim@cc.memphis.edu (Yan H. Sim)
To:        comp.protocols.tcp-ip
Subject:   FAQ...

Is there a FAQ for this group?  

SIM


-----------[000343][next][prev][last][first]----------------------------------------------------
Date:      Wed, 14 Dec 94 11:58:36 MET
From:      stale@inenco.no
To:        comp.protocols.tcp-ip
Subject:   UUENCODE algorithm



This might not be the right newsgroup, but I try anyway:

Do anyone of you out there know of source or desription of the algorithm used 
in UUENCODE / UUDECODE utilities in unix.  These are extensivly used when 
transferring objects, bineries and such over the net.  We need a similar 
algorithm to ensure only "printable" characters in a byte stream generated from 
a data structure (that can include integers etc. as well as strings).

Is this algorithm publicly known / available?

Does any of you know of a ftp site or somewhere it is possible to get any 
source for such an algorithm?

Please email me if you can help me in any way.

Thanks,  Stale

*******************************************************************************
* Ståle Villumstad           * Tel  : +47 22230420       * Private:           *
* Inenco A/S                 * Fax  : +47 22184015       * Munkebekken 89     *
* Gjerdrums vei 10 A         * Email: stale@inenco.no    * 1061 OSLO          *
* 0486 OSLO                  *                           * Norway             *
* Norway                     *        @}-,-`--           * Tel: +47 22328289  *
*******************************************************************************


-----------[000344][next][prev][last][first]----------------------------------------------------
Date:      14 Dec 1994 23:18:37 -0500
From:      barmar@nic.near.net (Barry Margolin)
To:        comp.protocols.tcp-ip
Subject:   Re: Source Routing?

In article <3coe3q$7uf@mogan.cc.metu.edu.tr> ergul@ds5500.cc.boun.edu.tr (Salih Ergul) writes:
>	I'm hearing about source routing option of IP packets, but I
>don't know what it is and how to use it. Could anybody explain this?

This option allows an IP packet to specify the routers that should be used,
rather than letting each router pick the next router automatically.  There
are two forms: strict source routing, in which every hop is specified; and
loose source routing, in which only some routers need to be specified, and
the paths between the specified routers are determined automatically.

When source routing is used, the reply is supposed to follow the reverse
path.  This is why they can be a security problem when a simple packet
filter is used as a firewall.  Normally fake source addresses aren't a big
problem, because the reply packets won't be routed back to the faker.  But
if source routing is used, the faker can overcome this routing problem.
-- 

Barry Margolin
BBN Internet Services Corp.
barmar@near.net

-----------[000345][next][prev][last][first]----------------------------------------------------
Date:      Wed, 14 Dec 1994 15:21:07 GMT
From:      albaugh@agames.agames.com (Mike Albaugh)
To:        comp.protocols.tcp-ip
Subject:   Re: Thin ethernet cabling

	I didn't see the original, so I'm following up to a reply:

Perry Donham (donham@iii1.iii.net) wrote:
: For thinwire (10Base2), the tee connector must be attached
: directly to the computer...you can't have a 'stub' running from
: the wall.
: [...]
: If you really must have only one wire coming out of the wall,
: you need to use 10BaseT (twisted pair) in a star topology.

	Or, if money is less of an object, and/or you are wiring few
enough rooms, you could bury a transceiver in the wall and run
AUI cable to the device. The break-even point comes when the total
cost of transceivers equals the cost of a hub. Although I'd be
delighted to be proven wrong, I don't know of any 10BaseT hubs that
cost less than about half-a-dozen Transceivers :-)

	If money is more of an object than hassle, you could just
have a junction-box with the Tee in the wall, and only install a
transceiver in those rooms where use in imminent :-)

					Mike

| Mike Albaugh (albaugh@agames.com) Time Warner Interactive
| (The entertainment company formerly known as Atari Games (_NOT_ Tramiel's))
| 675 Sycamore Dr. Milpitas, CA 95035		voice: (408)434-1709
| The opinions expressed are my own (Boy, are they ever)

-----------[000346][next][prev][last][first]----------------------------------------------------
Date:      14 Dec 1994 15:25:34 GMT
From:      gem@cam-orl.co.uk (Glenford Mapp)
To:        comp.protocols.tcp-ip
Subject:   Fast TCP checksumming

Hello all,

I am interested in fast TCP checksumming algorithms in software.
Ones written in C would be acceptable. The architectures in which 
I am particularly interested are the little-endian machines, ARM, Alphas, etc.

Please email me your replies.  I will post a summary in due course.


Regards

GEM


-----------[000347][next][prev][last][first]----------------------------------------------------
Date:      Wed, 14 Dec 94 20:53:22 EST
From:      compu-assist@pacifier.com (Compu-Assist)
To:        comp.protocols.tcp-ip
Subject:   Lantastic & Winsock

I currently have Lantastic 6.0 W/NDIS Drivers for SMC Ethernet..

I would like to be able to run Winsock over this configuration I have
installed the pktdrv's and they seem to be loaded yet when I try to
load winpkt It cannot locate the proper vector, I have asigned 0x60 to
the packect driver yet winpkt says no packet driver found on 0x60...

Any help would be GREATLY appreciated!



Jeff


-----------[000348][next][prev][last][first]----------------------------------------------------
Date:      14 Dec 1994 14:27:36 +0100
From:      roque@master.di.fc.ul.pt (Pedro Roque Marques)
To:        comp.protocols.tcp-ip
Subject:   Re: Linux-Router cannot cope with Sun-tcpip of our local stations

Ralf Hack (hack@pluto.rz.fh-reutlingen.de) wrote:
: Hello world,
 
: we setup a linux-router for our local network. With the local suns,
: we cannot route files (via ftp) above roughly 1k. However, we can
: receive them with our router. The same for xsessions. Routing sessions
: with netscape/mosaic .. does not work. However, routing an xterm-inal
: does work.
 
: We use a 486/33 PC with Linux 1.1.23. 
Get a newer version... versions up to 1.1.3x had a small bug in the IP
fowarding code that made then discard packets superior to 1486 bytes.
SunOs 5.x attempts to send packets of 1500 bytes that the router should 
route. You won't have problems with most other OSs because infortunelly they
don't do MTU discovery and send 596(?) byte packets.

: Our subnet consists of SGI's and
: PC's. The suns are running SunOS 5.3 Generc sun4m sparc.
 
: Any hint to solve this problem is very much appreciated. 
 
: Thanks in advance.
 
: mfg
: Ralf Hack

--
	Pedro Roque (roque@di.fc.ul.pt)

Disclaimer: Any resemblance between the above views and those of my
employer, my terminal, or the view out my window are purely
coincidental.  Any resemblance between the above and my own views is
non-deterministic.  The question of the existence of views in the
absence of anyone to hold them is left as an exercise for the reader.

-----------[000349][next][prev][last][first]----------------------------------------------------
Date:      Wed, 14 Dec 94 16:17:20 GMT
From:      dlevinso@atuis.alaska.net (Don Levinson)
To:        comp.security.unix,comp.protocols.tcp-ip
Subject:   IP Network Address Translation daemon?

does anyone know of daemon code to provide Network Address Translation for a non-compliant IP
addressed network? can run on either BSD or SYSV (preferrably SunOS)

-----------[000350][next][prev][last][first]----------------------------------------------------
Date:      14 Dec 1994 18:35:43 GMT
From:      tmt@hri.com (Thomas M. Talpey)
To:        comp.protocols.nfs,comp.protocols.tcp-ip
Subject:   Re: NFS/SLIP

> Vern Schryver:
> Over a correctly installed and operated Ethernet, NFS/TCP simply cannot
> be faster than NFS/UDP when UDP is built to go fast.

You're taking the position that the Ethernet is the only source of packet
loss. A single modern client can saturate the network, thereby dropping
UDP packets before they are sent. Servers have limited response times,
thereby dropping packets reaching receive queues. To say nothing of the
many-clients/one-server imbalance. These losses are just congestion - and
without adaptive algorithms at the client, they make dramatic changes in
throughput. Plunk Van's algorithms in an NFS, AFS, DFS, whatever client
and it will go faster in any real-world situation. TCP is the easiest way
to do this, but it works for UDP too.

Tom Talpey
tmt@hri.com

-----------[000351][next][prev][last][first]----------------------------------------------------
Date:      Wed, 14 Dec 1994 20:02:59 GMT
From:      morgan@world.std.com (William M Stair)
To:        comp.protocols.tcp-ip
Subject:   HELP programming sockets under UNIX

I'm writing a UNIX daemon that other programs need to talk to.  For
now, these programs will all be running under UNIX on the same
machine, but that's going to change.

Can ANYONE give me a hint on how to start programming using sockets?
This seems to be the most portable method for interprocess
communication, but it seems like alot of setup is required, and I
can't seem to find the answers in one place.

1) Put a description entry in /etc/services

2) Put a respawn entry for the daemon in /etc/inittab

3) Write the daemon with the following code... I don't know what ANY
   of this really means:

	a - Create the socket with "socket()"?

	b - Bind the socket to an address with "bind()"?

	c - Listen for information on the socket with "listen()"?

	d - Get the first "connect()"ion with "accept()"?

	e - Read the data with "recv()", "recvfrom()", or "recvmsg()"?
		-Which to use???

	f - Send any data with "send()", "sendto()", or "sendmsg()"?
		-Which to use???

	g - How do I close the connection when I'm done?

Any help will be greatly appreciated, including plain old pointers to
FAQs and good books on the subject!

-Morgan

-----------[000352][next][prev][last][first]----------------------------------------------------
Date:      Wed, 14 Dec 1994 20:34:36 GMT
From:      fmeschbe@pax.eunet.ch (Meschberger)
To:        comp.protocols.tcp-ip
Subject:   Re: Winsock/ODI on TokenRing

Peter Eisch (peisch@cfa.org) wrote:
: Does this work?  We're trying to coax it, but with no luck.  The cisco has
: SNAP on it's token ring interface, but the pc's don't seem to be able to
: get out...
 
: Ideas?

We use TCP/IP (and IPX by the way) over TokenRing with ODI drivers at COOP
Switzerland. We didn't have any problems so far.

So I have some questions : What kind of ODI and TCP/IP stack do you use
(Novell's, FTP's, ...) ? If using Novell's ODI Stack and LAN WorkPlace
(their TCP/IP) : Do you use the native Novell TokenRing Driver (token.com)
or do you use lansup.com (ODI over IBM's LAN Support Program for TokenRing) ?
Have you configured in the file net.cfg that the TokenRing Section (whether
it's token or lansup) presents token-ring_snap for tcpip (Something
like 'protocol tcpip 8237 token-ring_snap' (I'm not sure of the number,
though) and that the tcpip Section binds to that particular Protocol ?

That's abaout all that comes to my mind of the shelf :-)

Greetings
Felix

--
fmeschbe@pax.eunet.ch or mefcs@coop.ch whatever you prefer ...

-----------[000353][next][prev][last][first]----------------------------------------------------
Date:      Wed, 14 Dec 1994 21:06:55 GMT
From:      vjs@calcite.rhyolite.com (Vernon Schryver)
To:        comp.protocols.nfs,comp.protocols.tcp-ip
Subject:   Re: NFS/SLIP

In article <3cndtv$njm@tribune.hri.com> tmt@hri.com (Thomas M. Talpey) writes:
>> Vern Schryver:
>> Over a correctly installed and operated Ethernet, NFS/TCP simply cannot
>> be faster than NFS/UDP when UDP is built to go fast.
>
>You're taking the position that the Ethernet is the only source of packet
>loss. A single modern client can saturate the network, thereby dropping
>UDP packets before they are sent.

That is wrong.  A reasonable NFS client will have enough queue to not
drop any plausible string of NFS/UDP requests, because NFS is a
request-response protocol.  If the NFS code in a client overruns the
device driver queue in the client, then that NFS client code is broken.
Such a bug is unlikely because of the fundamental request-response nature
of NFS (unless someone with fewer clues than they think they have enables
vastly too many biods).


>                                  Servers have limited response times,
>thereby dropping packets reaching receive queues.

Servers do have limited response times, but the request-response nature
of NFS ensures that reasonably designed servers do not lose NFS requests
due to such congestion.

>                                                  To say nothing of the
>many-clients/one-server imbalance.

NFS/TCP is not a solution to that problem.

>                                   These losses are just congestion - and
>without adaptive algorithms at the client, they make dramatic changes in
>throughput. Plunk Van's algorithms in an NFS, AFS, DFS, whatever client
>and it will go faster in any real-world situation. TCP is the easiest way
>to do this, but it works for UDP too.

Wrong.  Van Jacobson's timers in TCP are simply not applicable to LAN's,
if only because they are too coarse by a factor of 500.  Congestion does
occur on LAN's, but it is handled by the traffic admission mechanisms
of the LAN's, either waiting for the token or Ethernet collisions.

Adaptive timers based on round trip times similar to the TCP slow start
algorithm sound very interesting for NFS, which is why Sun has been
working on them in their implementation for several years.  However,
those timers are in NFS itself.  The TCP timers in NFS/TCP are completely
irrelevant and useless for NFS over a small internet of Ethernets and
802.5 and FDDI rings.  Old fashioned NFS/TCP has nothing to do with "NFS
with slow-start."  The first is a good thing on a WAN, where there is
significant packet loss.  The second would use either TCP or UDP.

In practice, absolutely no properly running LAN has a signficant packet
loss rate.  Congestion control and avoidance schemes based on packet
loss are entirely inapplicable to LANs and modest internets consisting
of individual LANs.  Van Jacobson's congestion control mechanisms
depend fundamentally on packet loss.  NFS on any link with significant
packet loss is a loser (which is why NFS/TCP is good on WANs).


Vernon Schryver    vjs@rhyolite.com

-----------[000354][next][prev][last][first]----------------------------------------------------
Date:      14 Dec 94 21:45:30 GMT
From:      werme@alingo.zk3.dec.com (Eric Werme)
To:        comp.protocols.tcp-ip,comp.dcom.lans
Subject:   Re: packets sent to own address

zhebu@svusenet.ubs.ch writes:

>What happens exactly to a packet which is sent to it's own address --> 
 
>1. The packet is sent from an ip-driver to it's own address (not the ip-
>   loopback address). Does the ip-layer construct a packet for the lan
>   (i.e. the ethernet)?

In DEC OSF/1, all Ether/FDDI messages go through ARP address resolution to
map IP address into link level address.  That code recognizes that the
IP address is for one of the local interfaces and gives the packet to
the loopback driver.  The packet never makes it onto the wire/fiber.
-- 
Eric (Ric) Werme         |  werme@zk3.dec.com
Digital Equipment Corp.  |  This space intentionally left blank.

-----------[000355][next][prev][last][first]----------------------------------------------------
Date:      Wed, 14 Dec 1994 22:24:46 GMT
From:      msimon7@ua1ix.ua.edu (Morris Simon)
To:        comp.protocols.tcp-ip
Subject:   PC-Router connection

We have just attached a Cisco router to our fiber backbone and are waiting
for VAX/VMS TCP/IP software to establish one or more Internet servers. In
the meantime I would like to connect a PC directly to the router using
MS-DOS, Windows or Linux TCP/IP software. Is it possible to do this on the
VAX backbone before connecting a VAX server? Would programs such as
Minuet, Mosaic, Chameleon, or Linux TCP/IP utilities be able to interface
with the router before we get the VAX/VMS software?
 
Also: Can someone suggest some good VAX/VMS utilities for Internet use?
 
Morris Simon <msimon7@ua1ix.ua.edu>


-----------[000356][next][prev][last][first]----------------------------------------------------
Date:      Wed, 14 Dec 1994 22:45:05 GMT
From:      thinman@netcom.com (Technically Sweet)
To:        comp.lang.misc,comp.protocols.tcp-ip
Subject:   Massively parallel programming for the Internet

Are there any projects attempting to build a massively parallel
programming environment using spare cycles on the Internet?
Something where you put a really simple interpreter on the
end machines, mid-level cluster managers spaced periodically,
and end-user control stations?

This was inspired by the project to crack large prime numbers by
mailing large lists of numbers around to volunteers running the
sieve program.  Such a thing could be entirely automated.
David Gelernter's "Mirror Worlds" is clear articulation of this vision.

It's apparent to me that the Internet is going to have to dramatically
increase the sophistication of its internal "metabolic" processes;
parallel computations which monitor traffic and alter routing tables
will have to be done at some point.  An Internet Nervous System if you will.

Social consequences: well, at some point, you may be required to
run the interpreter as part of your Internet Tax.

Back to tech talk: is anybody using multisets as a basis for
poarallel programming?  Multisets looked like a possibly pleasant
paradigm for programming the distributed interpretive system.

Newsgroups: comp.parallel,comp.lang.misc,comp.protocols.tcp-ip
Summary: Massively parallel programming for the Internet
Followup-To: 
Distribution: world
Organization: International Foundation for Internal Freedom
Keywords: 

Are there any projects attempting to build a massively parallel
programming environment using spare cycles on the Internet?
Something where you put a really simple interpreter on the
end machines, mid-level cluster managers spaced periodically,
and end-user control stations?

This was inspired by the project to crack large prime numbers by
mailing large lists of numbers around to volunteers running the
sieve program.  Such a thing could be entirely automated.
David Gelernter's "Mirror Worlds" is clear articulation of this vision.

It's apparent to me that the Internet is going to have to dramatically
increase the sophistication of its internal "metabolic" processes;
parallel computations which monitor traffic and alter routing tables
will have to be done at some point.  An Internet Nervous System if you will.

Social consequences: well, at some point, you may be required to
run the interpreter as part of your Internet Tax.

Back to tech talk: is anybody using multisets as a basis for
parallel programming?  Multisets looked like a possibly pleasant
paradigm for programming the distributed interpretive system.

-- 

Lance Norskog
thinman@netcom.com
Artisputtingtogether. Art  s th ow n  aw y.

-----------[000357][next][prev][last][first]----------------------------------------------------
Date:      14 Dec 94 22:56:37 GMT
From:      ddl@harvard.edu (Dan Lanciani)
To:        comp.protocols.tcp-ip
Subject:   Re: Interpretation of Nagle algorithm

In article <LIN.94Dec14123838@excalibur.cs.purdue.EDU>, lin@cs.purdue.EDU (John Chueng-Hsien Lin) writes:

| We have observed a circular-wait behavior that decreases TCP
| throughput over ATM dramatically. The interaction of the following
| items causes the anomalous behavior:

[description deleted]

This kind of problem has existed in BSD's tcp send policy for some
time and it is not confined to unusual MTUs.  Any time that the send
buffer (= ``send window'') is significantly smaller than the receiver's
window, and also ``small enough,'' trouble is likely to occur.  Superficially,
you might say that part of the problem is that the send policy implements an
approximation of anti-congestion and several other good ideas in a rather
convoluted and intertwined manner that is hard to follow and harder to
tweak. :)  The real problem is that the code makes comparisons of the
form ``is what I can send now reasonable in light of what the receiver
has allowed me to send in the past'' without taking into consideration
the possibility that the sender's own buffer size is (and always was) the
limiting factor.  As has been mentioned, this can interact badly with
the ACK policy (which makes an analogous send decision based only on
its own buffer size).

There are two reasons that this problem doesn't show up more often.  First,
most systems tend to set their send and receive buffers to the same size
(at least by default).  Communications between two similar systems running
with default buffer sizes will not provoke the degenerate behavior.  Second,
the BSD send policy implements at least one test which is independent of
the relative maximum send/receive buffer sizes and depends only on the MSS.
Most systems default the send buffer size to 4k or 8k (or even more) and
this is sufficiently larger than the typical MTU that the problem is avoided.

(Ok, I suppose you could claim that you need an unusual MTU to cause trouble,
but read on.)

On a typical Ethernet with standard MTU of 1500, you can see this problem
between a sender with a ~2k buffer size and a receiver with an 8k buffer
size.  Transfers are driven by the receiver's ACK timeout and performance
is terrible.  This is an extreme example and you could argue that such
buffer sizes are doomed from the start, but other combinations that look
more benign can lead to surprises as well.  Unfortunately, it isn't unusual
to see such buffer sizes in real life.  I've encountered them in two cases:
(1)  Small send buffers are required to limit the total memory resources
that can be consumed by connections that have effectively been suspended
by back-pressure.  This can come up in a context where many, many such
connections have to be maintained.  (2)  The default has been reduced to
improve (!) performance against receivers that have poor buffering capacity
yet advertise large receive windows.  The second case is rather ironic since
trading lost packets for ACK-timeout-driven transfers often makes the situation
worse.  (Especially when the default was reduced to accommodate a single system
and the change caused degraded performance to all other machines :()

				Dan Lanciani
				ddl@harvard.*

-----------[000358][next][prev][last][first]----------------------------------------------------
Date:      15 Dec 1994 12:43:31 -0600
From:      les@MCS.COM (Leslie Mikesell)
To:        comp.protocols.tcp-ip
Subject:   Re: Source Routing?

In article <3cog2t$p8j@tools.near.net>,
Barry Margolin <barmar@nic.near.net> wrote:
>In article <3coe3q$7uf@mogan.cc.metu.edu.tr> ergul@ds5500.cc.boun.edu.tr (Salih Ergul) writes:
>>	I'm hearing about source routing option of IP packets, but I
>>don't know what it is and how to use it. Could anybody explain this?
>
>This option allows an IP packet to specify the routers that should be used,
>rather than letting each router pick the next router automatically.  There
>are two forms: strict source routing, in which every hop is specified; and
>loose source routing, in which only some routers need to be specified, and
>the paths between the specified routers are determined automatically.

Does everything observe source routing when it is used (unless it is
specifically filtered)?  What equipment will generate source routes?
Can it be used to allow access out of a private net (say 10.x.x.x)
by specifying a route of the 'public' connection plus the private
address that other routers wouldn't know about?  Or does an RFC 1631
translating gateway exist that would be a better approach?

Les Mikesell
  les@mcs.com

-----------[000359][next][prev][last][first]----------------------------------------------------
Date:      Thu, 15 Dec 1994 11:38:18 -0500
From:      lfa@olie.wvitcoe.wvnet.edu (Larry F Armbruster)
To:        comp.protocols.tcp-ip
Subject:   [Q]Slip Service

I am looking for advice, pointers, experiences, etc in setting up a slip
server.  There may be 400+ potential users with a maximum of 15 users at
a time.

TAI
Larry


-----------[000360][next][prev][last][first]----------------------------------------------------
Date:      15 Dec 1994 03:44:58 GMT
From:      ergul@ds5500.cc.boun.edu.tr (Salih Ergul)
To:        comp.protocols.tcp-ip
Subject:   Source Routing?

	I'm hearing about source routing option of IP packets, but I
don't know what it is and how to use it. Could anybody explain this?
Thanks!
--


#==========================================================================#
         Salih Ergul               (student of)      Bogazici University 
  E-mail: Salih.Ergul@boun.edu.tr                    Dept. of Comp. Eng.
  URL   : http://www.boun.edu.tr/~ergul/           Bebek/Istanbul   TURKEY
#==========================================================================#


-----------[000361][next][prev][last][first]----------------------------------------------------
Date:      Thu, 15 Dec 1994 05:04:53 GMT
From:      rv@fiji.cs.brown.edu (rodrigo vanegas)
To:        comp.protocols.tcp-ip
Subject:   Re: FAQ...

In article <1994Dec14.204224.31083@msuvx1.memphis.edu> yhsim@cc.memphis.edu (Yan H. Sim) writes:

>  Is there a FAQ for this group?  
>  SIM

In the 

   Newsgroups: news.lists,news.answers
   Subject: List of Periodic Informational Postings, Part 1/9

FAQ i see the following entry.

|
| Newsgroups: comp.protocols.tcp-ip
| Subject: TCP/IP FAQ for *
| From: gnn@netcom.com (George Neville-Neil)
| Frequency: unknown
| Date: Wed, 7 Sep 1994 05:32:07 GMT
| Archive-name: tcp-ip/FAQ
|  

As you can see, it is not cross-posted to news.answers which is why
you may not have noticed it.

-- 
rodrigo vanegas
rv@cs.brown.edu

-----------[000362][next][prev][last][first]----------------------------------------------------
Date:      15 Dec 1994 06:43:45 GMT
From:      tdh5821@halcyon.halcyon.com (Timothy D. Hall)
To:        comp.protocols.tcp-ip
Subject:   Re: WinNT & Windows TCP/IP connection

In article <3ckg5r$n3q@access1.digex.net>,
Eric V. Smith <esmith@access1.digex.net> wrote:
>In article <3cjdnr$5i1@news.halcyon.com>,
>Timothy D. Hall <timh@winserve.com> wrote:
>>
>>This will work only with another Windows NT 3.5 box or a Windows 95
>>box.
>
>This is not true.  I've connected a NT 3.5 beta to a Portmaster (I
>believe, you'd have to ask my provider) using CSLIP with no problems.

You're right, my first response was incorrect; I must have assumed that
Antonello had wanted to use the NT machine's shared resources.

A detailed treatment of the correct response can be found at
http://www.luc.edu/~tbaltru/faq/.  Refer to question Q4B-11.  The same
FAQ is available via anonymous ftp from rtfm.mit.edu in the file:
ntwebfaq.zip.

Eric, please, what's a Portmaster and how's this related?
--
Timothy D. Hall   | Winserve - Windows Internet Server
timh@winserve.com | winserve.001=204.118.34.11; net view \\winserve.001
206.827.8184      | http://www.winserve.com

-----------[000363][next][prev][last][first]----------------------------------------------------
Date:      Thu, 15 Dec 1994 17:20:11 -0600
From:      resnick@uiuc.edu (Pete Resnick)
To:        comp.protocols.tcp-ip
Cc:        mscarton@mudshark.sunquest.com (Mark A. Scarton)
Subject:   Re: Time Servers?

In article <mscarton.214.000E243D@mudshark.sunquest.com>,
mscarton@mudshark.sunquest.com (Mark A. Scarton) wrote:

>>Are there any publicly accessible time servers on the Internet?  If so, 
>>what are the addresses?
>
>The US National Institute of Standards and Technology (NIST) provides daytime 
>service on port 13....

And regular NTP service on UDP port 123 and time service on port 37.

>...at time_a.timefreq.bldrdoc.gov (132.163.135.130).

That name is incorrect. The correct domain name is:

    time-a.timefreq.bldrdoc.gov

Underscores are illegal in domain names.

pr
-- 
Pete Resnick    (...so what is a mojo, and why would one be rising?)
Doctoral Student - Philosophy Department, Gregory Hall, UIUC
System manager - Cognitive Science Group, Beckman Institute, UIUC
Internet: resnick@uiuc.edu

-----------[000364][next][prev][last][first]----------------------------------------------------
Date:      15 Dec 1994 07:47:33 GMT
From:      bsmith@wci.com
To:        comp.unix.solaris,comp.protocols.tcp-ip
Subject:   Re: Help! binding socket to address with bind dies with already..


Hello Martin,

In article <D0q252.E6I@IRO.UMontreal.CA> you wrote:
: Aaron Newman (anewman@ernie.eecs.uic.edu) wrote:
 
: : I am developing a  small TCP-IP C/S app and when I try to bind()
: : before listen()ing on a port, it works fine the first time.  But after I
: : exit the app and restart, Bind dies with 'address in use' error code.


: 	If it does not answer your question and fix your problem, 
: let me know, maybe I don't understand your question.
 
: - martin

I have a similar problem on UnixWare.  Sometimes I can stop and restart
the program several times in quick succession without problems binding.
And sometimes it can not bind on the first try after a long pause.
The error code is always 'Addr in use', and when it fails the command
'netstat -s' fails to find any reference to the port specified.


The code (omitting error handling) follows...

    bzero((char *)&s_in, sizeof(s_in));
    s_in.sin_family     =       AF_INET;
    s_in.sin_addr.s_addr=       inet_addr(ip_addr);
    s_in.sin_port       =       htons(port);
    if (*fdp = socket(AF_INET, SOCK_STREAM, 0) < 0)
    if (setsockopt(*fdp,SOL_SOCKET,SO_REUSEADDR,&on,sizeof(on)) < 0)
    if (bind(*fdp, (struct in_addr *)&s_in, sizeof(s_in))<0)
    listen(*fdp, MAX_LISTEN);


Thanks in advance,
Bob Smith
2177 Augusta Place,  Santa Clara, CA  95051-1714
Voice: (408) 296-1546    FAX: (408) 296-1547
Email: bsmith@wci.com
-- 
Bob Smith
Wireless Connect, Inc
2177 Augusta Place,  Santa Clara, CA  95051-1714
Voice: (408) 296-1546    FAX: (408) 296-1547

-----------[000365][next][prev][last][first]----------------------------------------------------
Date:      15 Dec 1994 08:03:03 GMT
From:      AIR - System Design <air@teleport.com>
To:        comp.infosystems.www.users,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.infosystems.www.providers,comp.os.ms-windows.networking.tcp-
Subject:   Re: WWW Server/Client without INTERNET??

With Nullsock on each machine running NetScape and nothing on
the server at all, All machines can open the .htm files
in f:\public on the server.

I do not use forms or CGIs. I just share hypermedia information.


-> >The original poster wanted to supply WWW services to the other PCs on the 
-> >same LAN.
-> >
-> >Therefore he needs a complete setup with a WWW server and WWW clients.
-
-> I haven't tried this, but wouldn't it be possible to run Mosaic or Netscape with a 
-> nullsock and access the HTML documents off the LAN server as local files?
-> 
-> Olof



-----------[000366][next][prev][last][first]----------------------------------------------------
Date:      Thu, 15 Dec 1994 08:27:56 GMT
From:      ericma@netcom.com (Eric S. Ma)
To:        comp.protocols.tcp-ip
Subject:   Error compiling PPP on linux

Does anyone know why I got the following error when I compiled Linux 1.0.9
kernel with PPP support (ppp-2.1.2b)?

 drivers/net/net.a(ppp.o): Undefined symbol _dev_kfree_skb referenced from
 text segment 

Thanks.

Eric
ericma@netcom.com

-----------[000367][next][prev][last][first]----------------------------------------------------
Date:      15 Dec 1994 08:30:46 GMT
From:      etxmesa@eos.ericsson.se (Michael Salmon)
To:        comp.protocols.tcp-ip
Subject:   Re: Telnet for 3270?

In article <1994Dec14.131950.6067@venus.gov.bc.ca>
glittle@bcsc02.gov.bc.ca (Glen Little) writes:
|> Does anyone know of a Windows based Telnet for 3270 program? Preferably
|> free or share-ware, but I'll consider commerical ones.

Netmanage Chameleon has a good 3270 emulator, if any 3270 emulator can
be considered good that is.

-- 

Michael Salmon

#include	<standard.disclaimer>
#include	<witty.saying>
#include	<fancy.pseudo.graphics>

Ericsson Telecom AB
Stockholm

-----------[000368][next][prev][last][first]----------------------------------------------------
Date:      15 Dec 1994 08:42:44 GMT
From:      rob@dutecaj.et.tudelft.nl (Rob Wiers)
To:        comp.protocols.tcp-ip
Subject:   Telnet encription option rfc, anyone ?


Hello all,

I'm trying to implement a telnet server on a firewall. We would like
to support encryption, but I can't seem to find a rfc that specifies an
implementation of an encryption option for telnet. Can someony give me
some directions?

Please reply by email.

Thanx, Rob


rob@liberator.et.tudelft.nl
LRGWiers@Et.TUDelft.NL

-----------[000369][next][prev][last][first]----------------------------------------------------
Date:      Thu, 15 Dec 1994 13:50:20
From:      waseem@ftp.com
To:        comp.protocols.tcp-ip
Subject:   multicast applications?

I am looking for multicast shareware applications? Does anyone know of any and 
where to get them? (I already have the multicast version of dogfight)

any replies appreciated
waseem@ftp.com

-----------[000370][next][prev][last][first]----------------------------------------------------
Date:      Thu, 15 Dec 1994 14:08:23
From:      mscarton@mudshark.sunquest.com (Mark A. Scarton)
To:        comp.protocols.tcp-ip
Subject:   Re: Time Servers?

>Are there any publicly accessible time servers on the Internet?  If so, 
>what are the addresses?

The US National Institute of Standards and Technology (NIST) provides daytime 
service on port 13 at time_a.timefreq.bldrdoc.gov (132.163.135.130).  They even
provide sample source for how to set a Unix time of day clock from this server.
This provides access to the cesium clock ensemble in Boulder, CO.

There is much more additional information available on this server with respect
to time synchronization and protocols: documentation, code, and contacts.


Mark A. Scarton, ABD
Sunquest Information Systems
4505 South Wasatch Blvd, Suite 100
Salt Lake City, Utah  84124-4787

-----------[000371][next][prev][last][first]----------------------------------------------------
Date:      Thu, 15 Dec 1994 10:10:29 +0000
From:      Steve@starbug1.demon.co.uk (Steve Kay)
To:        comp.protocols.tcp-ip
Subject:   Re: POP server for SunOS?

In article <JIM.94Dec7130943@dewar.cs.strath.ac.uk>
           jim@cs.strath.ac.uk "Jim Reid" writes:

> In article <clyde.786507279@trojan> clyde@hitech.com.au (Clyde Smith-Stubbs)
>  writes:
> 
>    Does anyone know of a POP server that will run under SunOs?
> 
> The one that comes with MH works just fine. We've been using it for years.
> 
I recently asked Unipalm for a trial copy of Mail-It.  It arrived with the
necessary pop2/pop3 daemons for copying over to SunOS.
-- 
Steve Kay

-----------[000372][next][prev][last][first]----------------------------------------------------
Date:      Thu, 15 Dec 1994 10:41:37 +0000
From:      Alex@arcfan.demon.co.uk (Alex McLintock)
To:        comp.protocols.tcp-ip
Subject:   rpc library for winsock

Is there an rpc library for winsock? commercial or free.

I am currently using MS Visual C++ to access telnet via winsock.
I am currently trying to understand "goahead" since my telnet deamon
won't supress go aheads. (This is a slightly diferent project)

If this is the worng place to discuss this could people point me in the
right direction. Thanks. 

(and yes I am reading the rfcs :-)


-- 
Alex McLintock                           [also alexmc@biccdc.co.uk]
Editor for the Project Galactic Guide (alt.galactic-guide) Number 6

-----------[000373][next][prev][last][first]----------------------------------------------------
Date:      15 Dec 1994 11:15:25 GMT
From:      malcolm@orbotech.co.il (Malcolm)
To:        comp.protocols.tcp-ip
Subject:   [Q] Interop

Anyone have a schedule for INTEROP, or how I can get it ?



---
/* Malcolm Kavalsky             | malcolm@orbotech.co.il  */
/* Orbotech Inc.                | 972-842-3843 (Voice)    */
/* Industrial Zone, P.O.Box 215 | 972-843-8769 (FAX)      */
/* 70651 Yavne, Israel          |                         */



-----------[000374][next][prev][last][first]----------------------------------------------------
Date:      Thu, 15 Dec 1994 16:17:29
From:      ylwm0237@cyberstore.ca (Emerald Solutions Inc.)
To:        comp.protocols.tcp-ip
Subject:   Windows for Workgroups and WinSock won't work together ?

It seems that theWindows for Workgroups 3.11 network drivers prevent my 
WinSock PPP stack ( CHAMELEON SAMPLER ) from loading.

I have to start windows with WIN/N in order to dial up to the internet, 
otherwise I get bumped to the DOS prompt with no explaination when I try to 
load CUSTOM TCP-IP.

I need to dial up to the internet while connected to workgroup resources.  
I have tried both the ODI and usuall Windows Network Setups but the results 
are the same.

Any insight into this would be most helpful. 

Thanks very much.

Anthony Jones

PS  I am using a Xircom Ethernet Pocket adaptor on LPT1 if that seems relavant.

Thanks again.

AJ

-----------[000375][next][prev][last][first]----------------------------------------------------
Date:      Thu, 15 Dec 1994 13:13:58 +0000
From:      philc@hipstech.demon.co.uk (Phil Cramp)
To:        comp.protocols.tcp-ip
Subject:   Printing via Xyplex Terminal Servers

We have a requirement to print over a TCP/IP network where the remote 
device handling the printer will be a Xyplex terminal server (with the 
printer plugged into one of its serial ports).

We are given to understand that telnet End Of Record will need to be 
negotiated with the terminal server in order to do this - I would guess 
that this is not one of the "ordinary" negotiations completed when you 
open a Telnet connection.

Has anybody had any experience with Xyplex Terminal servers?
Will we need to develop our own interface script/program?
If we do, what should we look out for?

All help gratefully received.
----
Phil Cramp                          philc@hipstech.demon.co.uk
AT&T Istel Ltd., Redditch, UK
#include <std/disclaimer>
                That's it - we're platform shoes.


-----------[000376][next][prev][last][first]----------------------------------------------------
Date:      15 Dec 1994 13:40:31 GMT
From:      carlson@Xylogics.COM (James Carlson)
To:        comp.protocols.tcp-ip
Subject:   Re: TCP RFC 793 to BERKLEY SOCKETS correlation  -  post.fil [1/1]

In article <787391712snz@cucumber.demon.co.uk>, Andrew@cucumber.demon.co.uk (Andrew Gabriel) writes:
|> In article <3cieob$3lf@newhub.xylogics.com>
|>            carlson@xylogics.com "James Carlson" writes:
|> [snip]
|> 
|> > The TCP_NODELAY flag turns off the Nagle algorithm.  As far as I'm able
|> > to tell, it has nothing to do with the PUSH bit in any implementation.
|> > 
|> > At least in the BSD4.3 code, PUSH is set when the data being sent by TCP
|> > fits entirely within the window and there's no more to send.  That is to
|> > say, it does it automatically by default.
 
|> >                                            (The comments in the code say
|> > that this is done to work around bugs in implementations which need to
|> > see either a full buffer or a PUSH before any data are delivered to the
|> > application level.)
|> > 
|> 
|> I havn't looked at the BSD source, but I have to say this sounds wrong
|> to me. An implementation is only *required* to forward data to the
|> application when:
|> 1.   PSH is set;
|> 2.   enough data is received to fill the application's read buffer (no
|>      point waiting for any more data);
|> 3.   FIN is set (no point waiting for any more data);
|> 4.   Possibly if URG is set and PSH isn't (depends how URG is signalled).
|> 
|> An implementation might also choose to forward data to the application
|> at other times convenient to it (such as an IP segment boundary) if the
|> API allows, but it is not *required* to do so.
|> 
|> I have come across applications which don't set PSH when they should,
|> e.g. at the end of some data which is to elicit an application level
|> response. Funnily enough, rcp had this problem back in SunOS 3 days,
|> (and I presume BSD's rcp at that time also). Maybe the behaviour of the
|> BSD TCP implementation (if correctly reported) was an attempt to get
|> round such buggy applications?

I didn't want to speculate on which applications or implementations
might be getting this one right or wrong.  I just wanted to state what
the BSD world does.  There's no way for an application running on a
standard BSD system to request that the TH_PUSH bit be set.  Here's the
relevant part of sys/netinet/tcp_output.c from BSD4.4 (lines 358-365,
copyright by the Regents of the University of California, no UCC
warranties, et cetera):

	/*
	 * If we're sending everything we've got, set PUSH.
	 * (This will keep happy those implementations which only
	 * give data to the user when a buffer fills or
	 * a PUSH comes in.)
	 */
	if (off + len == so->so_snd.sb_cc)
		flags |= TH_PUSH;

---
James Carlson <carlson@xylogics.com>            Tel:  +1 617 272 8140
Annex Software Support / Xylogics, Inc.               +1 800 225 3317
53 Third Avenue / Burlington MA  01803-4491     Fax:  +1 617 272 2618

-----------[000377][next][prev][last][first]----------------------------------------------------
Date:      Thu, 15 Dec 1994 22:19:10 LOCAL
From:      cslater@omni.voicenet.com (charles slater)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.domains
Subject:   Re: Need help on SLIP line

Sorry, I can't answer your question.  But would you mind telling me what 
software you are using for your SLIP connection?
Thanks

In article <3cqi8q$d6@news.cais.com> duffy@cais2.cais.com (Duffy Men) writes:
>From: duffy@cais2.cais.com (Duffy Men)
>Subject: Need help on SLIP line
>Date: 15 Dec 1994 23:08:09 GMT
 
>I have a IBM RS6000 AIX 3.2.5 system and I can use SLIp to connect to my 
>Internet provider host.  Currently, I have following problem need help.
 
>1. I can "ping" my Internet provider host no problem, but I can not 
>"telnet" or "ftp" to my provider host.  If I try to "telnet" to my 
>provider host, after "connect to ...." message.  The cursor just stay in 
>there and hangup.
 
>2. Does anyone know how to setup default gateway and name server for SLIP 
>line on RS6000?
 
>Thank you for help.


-----------[000378][next][prev][last][first]----------------------------------------------------
Date:      Thu, 15 Dec 1994 17:49:36 GMT
From:      thinman@netcom.com (Technically Sweet)
To:        comp.parallel,comp.lang.misc,comp.protocols.tcp-ip
Subject:   Massively parallel programming for the Internet

Are there any projects attempting to build a massively parallel
programming environment using spare cycles on the Internet?
Something where you put a really simple interpreter on the
end machines, mid-level cluster managers spaced periodically,
and end-user control stations?

This was inspired by the project to crack large prime numbers by
mailing large lists of numbers around to volunteers running the
sieve program.  Such a thing could be entirely automated.
David Gelernter's "Mirror Worlds" is clear articulation of this vision.

It's apparent to me that the Internet is going to have to dramatically
increase the sophistication of its internal "metabolic" processes;
parallel computations which monitor traffic and alter routing tables
will have to be done at some point.  An Internet Nervous System if you will.

Social consequences: well, at some point, you may be required to
run the interpreter as part of your Internet Tax.

Back to tech talk: is anybody using multisets as a basis for
parallel programming?  Multisets looked like a possibly pleasant
paradigm for programming the distributed interpretive system.

-- 

Lance Norskog
thinman@netcom.com
Artisputtingtogether. Art  s th ow n  aw y.


-----------[000379][next][prev][last][first]----------------------------------------------------
Date:      Thu, 15 Dec 1994 19:29:21 GMT
From:      Gueorgui E. Vakhrouchev <george@lcs.krasnoyarsk.su>
To:        comp.protocols.tcp-ip,relcom.tcpip,vmsnet.networks.tcp-ip.wintcp
Subject:   Is there any TCP/IP solution of the problem?

Hello everybody,

   We are  needed in  some networking TCP/IP software product,  but  we
do not know where to obtain such  product  from. Below  I  have given a
describe  of such product functions. Maybe,  everyone knows such TCP/IP
product.

                              The Question
_________________________________________________________________________
     We live  in city where there is a high mountain and there located "a
brain" that  answers to the requirements of various subscribers from city
and grants  them  a  radio-channel  for  point-to-point  connection.  All
subscribers are  allowed from  that mountain.  The central  brain has  an
access to  28 frequency  ranges and it grants subscribers these frequency
channels by their requirements. Each connection is established in 250 mks
(microseconds). About 20 - 30 thousands of subscribers are served in all.
     There is  the following  problem: many users want to work with data-
base in their central office using client -- server mode. The others want
to have  Telnet or FTP sessions and so they fix the channel in "permanent
busy" state.  So, only  28 pairs  can be  served at  such mode (I call "a
pair" a  connection of  two local  networks) that is very few. Many users
have permanent  sessions but  these sessions  have little activities; the
others make  a query  from a  remote data-base that works for an hour and
there is  no an  exchange of a data over the channel for entire hour, but
the channel  is busy.  How to  be? Is it possible a solution of such kind
using your  products?: when  a session has no activity then disconnect is
made on  low layer,  but at that parameters of session have been retained
by server  and client.  So, an application of higher layer such as Telnet
will notice  nothing, but  the channel  can be used by other users. If an
activity appears  in session  (a pressing  of a  key and  so on) then low
layer requests  the channel  and has gotten it in 250 mks. The session is
restored, at  that an  application layer  does  not  see  a  break  of  a
connection was. A break of a connection can be made by timer.
     Entire connection  is "a  black box"  with  RS232C  interface,  full
analog of SLIP on direct line. An equipment is Motorola compatible.
______________________________________________________________________________

 Gueorgui Vakhrouchev                   e-mail: george@lcs.krasnoyarsk.su
 Chief of System Department
 Large Computer Systems Co Ltd          Fax:    +7 3912 343195
______________________________________________________________________________


-----------[000380][next][prev][last][first]----------------------------------------------------
Date:      15 Dec 1994 19:39:24 GMT
From:      Vineet Madan <75664.3400@CompuServe.COM>
To:        comp.infosystems,comp.protocols.appletalk,comp.protocols.misc,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc
Subject:   -->MacTCP and IBM TCP/IP over WAN?

Does anyone have any experience connecting Macs (on ethernet running
MacTCP) to a LAN or WAN used by IBM's (on Token Ring, also running 
TCP/IP)?  Specifically, I was wondering if there would be any 
additional difficulties due to the addition of a small Apple 
ethernet LAN to such a WAN.  I have been told by the corporate folks
that connecting my Mac LAN to the WAN would cause then additional 
headaches just because of the MacTCP and the fact that it isn't IBM 
TCP/IP.  This doesn't make a heck of a lot of sense to me.  Hardware
aside, is there a difference in the packets and their method of 
transmission over the WAN?  If anyone is currently running Macs and 
IBM's on a LAN or WAN, could you get back to me with the details of 
your configuration and whether the presence of MacTCP has cause any 
problems on your network?  Thanks a lot for your help.
    
     --Vineet Madan
       (vineet@crm.com, 75664.3400@compuserve.com)

-----------[000381][next][prev][last][first]----------------------------------------------------
Date:      Thu, 15 Dec 1994 21:34:45 GMT
From:      giza@apollo.HP.COM (Peter E. Giza)
To:        comp.protocols.tcp-ip
Subject:   Need info on RMON

Hi,

   I have a net dump from a DOS PC that says it can be
viewed with RMON compatible tools.  Can somebody point
to such a beast, preferably a un*x tool.  Thanks

-pete
-- 
                                              ~
                       |                                 ~
Peter E. Giza          |  "With my luck, the light at         ~ 
HP/Apollo              |   the end of the tunnel is an           ~
Chelmsford, MA         |   oncoming freight train."                ~
                       |                                            ~
PHO:(508) 436-5564     | -------+ +------+         ------          \ /
FAX:(508) 436-5140     |        | | H&P   \______   [?]  \__O_O_O__| |__
EML:giza@apollo.hp.com |        | | RR Lines     |  |    |  | | |  | |  \
                       | --(0)--+=+--(0)----(0)--+-=-(0)-(0)-(0)---(0)--\\
        ~:?)           |




-----------[000382][next][prev][last][first]----------------------------------------------------
Date:      Thu, 15 Dec 1994 22:53:03 GMT
From:      paul@atlas.abccomp.oz.au (Paul Brooks)
To:        comp.unix.solaris,comp.protocols.tcp-ip
Subject:   Re: Help! binding socket to address with bind dies with already..

In article <3ca6ac$q0d@news.eecs.uic.edu> anewman@ernie.eecs.uic.edu (Aaron Newman) writes:
|Hello.
|
|I am developing a  small TCP-IP C/S app and when I try to bind()
|before listen()ing on a port, it works fine the first time.  But after I
|exit the app and restart, Bind dies with 'address in use' error code.
|This goes on for ~10 minutes, and then it works again.  I am closing
|the socket down after the app exit with 'close()'.  Is this the wrong
|way to close it?  The man pages are no help. (surprise)

You are looking at the wrong man pages :-)
try looking up 'setsockopt'.
Before calling bind(), call setsockopt(SO_REUSEADDR).
The problem is that after you close() the connection, the socket structure
may hang around in TIME_WAIT state, probably for about 4 minutes, during which
time you can't bind to the same port number again. Provided the new connection
comes from a different host or port than the earlier one, ther eis no
confusion, so things will work correctly.

-- 
Paul Brooks              |paul@abccomp.oz.au       |Emerging Standard:
TurboSoft Pty Ltd        |pwb@newt.phys.unsw.edu.au|  one that has not yet
579 Harris St., Ultimo   |                         |  been superseded.
Sydney Australia 2007    |ph: +61 2 281 3155       |  

-----------[000383][next][prev][last][first]----------------------------------------------------
Date:      15 Dec 1994 23:08:09 GMT
From:      duffy@cais2.cais.com (Duffy Men)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.domains
Subject:   Need help on SLIP line

I have a IBM RS6000 AIX 3.2.5 system and I can use SLIp to connect to my 
Internet provider host.  Currently, I have following problem need help.

1. I can "ping" my Internet provider host no problem, but I can not 
"telnet" or "ftp" to my provider host.  If I try to "telnet" to my 
provider host, after "connect to ...." message.  The cursor just stay in 
there and hangup.

2. Does anyone know how to setup default gateway and name server for SLIP 
line on RS6000?

Thank you for help.

-----------[000384][next][prev][last][first]----------------------------------------------------
Date:      Fri, 16 Dec 1994 00:08:35 GMT
From:      mintz@cup.hp.com (Ken Mintz)
To:        comp.os.vxworks,comp.protocols.tcp-ip
Subject:   Re: Simultaneous TCP/Socket Connections

Mark Evans (evansmp@mb51913.aston.ac.uk) wrote:
> squeedy (hjb@netcom.com) wrote:
> > 2) you cannot listen and then do a connect on the same socket because
> > 	all 4.X BSD UNIX derived TCP/IP implementations specifically
> > 	prohibit this at the top kernel socket level.  it is
> > 	not a "host requirement" to provide this capability, and
>
> None of the RFC's appear clear on this. 

  This would not be covered by an RFC.  The RFCs cover protocols, not APIs.
  Sockets is an artifact of the BSD API, and how sockets behave is defined
  by the BSD implementation.  This limitation is specific to that API.

  Yes, another API could do something different, e.g. TLI and XTI.

-- Ken Mintz

-----------[000385][next][prev][last][first]----------------------------------------------------
Date:      16 Dec 1994 01:34:07 GMT
From:      sklower@oboe.CS.Berkeley.EDU (Keith Sklower)
To:        comp.protocols.tcp-ip
Subject:   Re: BSD support for Variable length subnet masks?

In article <3bk8io$gem@crab.spider.co.uk>,
Ian Heavens <ian@spider.co.uk> wrote:
}I'm trying to find out whether (and how, if so) BSD TCP/IP supports
}variable length subnet masks.  The code doesn't seem to support it
} - at least, the routines that convert network addresses to numbers
} don't - yet a netstat -r on our local BSD machines shows different
}subnet masks.

It depends on which version of BSD you're talking about.  Certainly,
4.2 and 4.3 BSD supported having the same netmask on all interfaces,
and the code since 4.3 Reno has been evolving to becomes smarter about
having different netmasks on different interfaces, and different routes.

If you put in something like
soda-2-net      128.32.33
wireA-net       128.32.112.132

in /etc/networks under 4.4 based code , you can get netstat -r to report

Internet:
Destination      Gateway            Flags     Refs     Use  Interface
default          inr-180-soda-33.Be UG          3 35144133  ef0
127              localhost.Berkeley UR          0        0  lo0
localhost.Berkel localhost.Berkeley UH          3     1969  lo0
soda-2-net       link#1             UC          0        0  ef0
inr-180-soda-33. 0:0:c:e:b:20       UHL         1        0  ef0
vangogh.CS.Berke 8:0:9:13:d:33      UHL         2     3596  ef0
orodruin.CS.Berk 0:0:3c:0:2d:8d     UHL         1        3  ef0
conviction.CS.Be link#1             UHL         2    18093  ef0
propaganda.CS.Be link#1             UHL         1     9125  ef0
sneezy.CS.Berkel 0:20:af:b:b5:6e    UHL         1       19  lo0
wireA-net.ttb.CS link#2             UC          0        0  ef1
wireA2           0:20:af:b:e6:c1    UHL         5      907  ef1
wireB-net.ttb.CS link#3             UC          0        0  ef2
wireB2           link#3             UHL         4      685  ef2
wireC-net.ttb.CS wireA2             UG          0        0  ef1
wireE-net.ttb.CS wireA2             UG          0        0  ef1
wireF-net.ttb.CS wireB2             UG          0        0  ef2
wireG-net.ttb.CS wireA2             UG          1    18632  ef1
wireH-net.ttb.CS wireB2             UG          0        0  ef2
128.32.112.168   sneezy-f           U           0        0  fake0
128.32.112.172   wireB2             UG          0        0  ef2
128.32.112.180   wireA2             UG          0       75  ef1
128.32.112.188   wireA2             UG          0        0  ef1


-----------[000386][next][prev][last][first]----------------------------------------------------
Date:      16 Dec 1994 03:15:22 GMT
From:      mjr@tis.com (Marcus J Ranum)
To:        comp.security.unix,comp.protocols.tcp-ip
Subject:   Re: IP Network Address Translation daemon?

>does anyone know of daemon code to provide Network Address Translation

	One approach is to put the network behind a firewall that
only has one (legal) connection to the Internet. Other approaches
involve expensive routers; I think several router makers do some form
of address translation but it doesn't work very well. Remember
that some protocols on TOP of TCP encode addresses in them: FTP is
just one example. Dual-homed bastion firewalls effectively do the
translation but don't give you direct connectivity.
	If direct connectivity is your goal, The Correct Answer is
to fix the addresses -- otherwise the network will just keep growing
and sooner or later it'll come back to haunt you.

mjr.

-----------[000387][next][prev][last][first]----------------------------------------------------
Date:      Fri, 16 Dec 1994 03:39:24 GMT
From:      gzhang@mprgate.mpr.ca (Greg Zhang)
To:        comp.protocols.tcp-ip
Subject:   Runtime access of routing table entries?

Hello,

    I am wondering whether there is some mechanism to access routing
table entries at run time. I am aware of the memory dump of gated. What
I need is the next hop router's address for a destination network as the
packet arrives.

    Thanks in advance.


				Greg


-----------[000388][next][prev][last][first]----------------------------------------------------
Date:      16 Dec 1994 04:35:00 GMT
From:      bealar@ndlc.occ.uky.edu (Andy Beal)
To:        comp.protocols.tcp-ip
Subject:   Netmask on seperate network

I hope this isn't in the FAQ, I couldn't find it, so flames welcomed.

My question is this, can one netmask server more than one physical network?

Here's the situation, at Daviess County Schools in Owensboro, KY, we 
currently have a Netware 3.11 server, and an RS6000.  The State Bd. of 
Edu, is bringing in their statewide WAN.  We have been assigned 62 IP 
addresses, (example 170.182.38.10 - 170.182.38.72 these aren;t the actual 
ones, I can't recite them from memory yet.)  Anyway, we have 62 IP 
addresses, and one Netmask for the building network.  Our Netware Router 
has 1 Token-Ring network card, 1 Ethernet Network card, and 1 ArcNet 
card.  We have been told from one source that each network must have it's 
own netmask.  We have been told from a slightly more reliable source, 
that one netmask can be used on several networks.  
Our Question is, can we use one netmask throughout the entire building, 
or do we need 3 individual netmasks, one for each physical network?


--
                /^\      /^\               
___________/\  /   \    /   \  /\________  Andy Beal
             \/     \  /     \/            bealar@ndlc.occ.uky.edu
                     \/                    

-----------[000389][next][prev][last][first]----------------------------------------------------
Date:      16 Dec 1994 04:36:06 GMT
From:      bealar@ndlc.occ.uky.edu (Andy Beal)
To:        comp.protocols.tcp-ip
Subject:   Re: Netmask on seperate network

Andy Beal (bealar@ndlc.occ.uky.edu) wrote:
: I hope this isn't in the FAQ, I couldn't find it, so flames welcomed.
 
: My question is this, can one netmask server more than one physical network?
 
: Here's the situation, at Daviess County Schools in Owensboro, KY, we 
: currently have a Netware 3.11 server, and an RS6000.  The State Bd. of 
: Edu, is bringing in their statewide WAN.  We have been assigned 62 IP 
: addresses, (example 170.182.38.10 - 170.182.38.72 these aren;t the actual 
: ones, I can't recite them from memory yet.)  Anyway, we have 62 IP 
: addresses, and one Netmask for the building network.  Our Netware Router 

Ahh, sorry that should be a Netware SERVER, not router!

: has 1 Token-Ring network card, 1 Ethernet Network card, and 1 ArcNet 
: card.  We have been told from one source that each network must have it's 
: own netmask.  We have been told from a slightly more reliable source, 
: that one netmask can be used on several networks.  
: Our Question is, can we use one netmask throughout the entire building, 
: or do we need 3 individual netmasks, one for each physical network?


: --
:                 /^\      /^\               
: ___________/\  /   \    /   \  /\________  Andy Beal
:              \/     \  /     \/            bealar@ndlc.occ.uky.edu
:                      \/                    

--
                /^\      /^\               
___________/\  /   \    /   \  /\________  Andy Beal
             \/     \  /     \/            bealar@ndlc.occ.uky.edu
                     \/                    

-----------[000390][next][prev][last][first]----------------------------------------------------
Date:      16 Dec 1994 05:58:51 GMT
From:      cozmen@osf1.gmu.edu (Cenk Ozmen)
To:        comp.protocols.tcp-ip
Subject:   TIA



	Hello, I was wondering if anyone knows where I can get "tia".  I
use Mosaic for Windows and I need tia to run it from home.  Thanks a
bunch.  If you could please e-mail, that would be very nice:

cozmen@osf1.gmu.edu

-----------[000391][next][prev][last][first]----------------------------------------------------
Date:      16 Dec 1994 06:40:24 GMT
From:      ccrowley@coho.halcyon.com (Calvin L. Crowley)
To:        comp.protocols.tcp-ip
Subject:   rarp client sun nit(4) questions

Not sure if this is the right newsgroup for this or not but ...
[ cross posted to comp.sys.sun.misc, comp.protocols.tcp-ip ]

Have need to move rarp client software that currently works with
either HP LLA or SCO LLI to Sun-OS 4.1.3.  Sun has NIT(4n) which appears
to be STREAMS based as the SCO LLI is.  Anyway, have gotten as far
as obtaining the physical address via SIOCGIFADDR and interpreting
it correctly with the ether_ntoa routines. But, what next?  Is there
a sun-os equivalent to the SCO "Version 4.0 Device Driver Writer's
Guide Supplement", Chapter 5 which describes LLI or the HP LLA guide? 
We are kind of broke and short of sun-os support and manuals at our
site but i can scrape up the $$$ if i know what i am looking for.
The sun-os if(4) and nit(4) man pages have some interesting
pointers but not enough examples.

Pointers to nit code would be excellent.  To a working sun-os rarp client
even better.  Have already seen parpd-1.2 (if.c and nit.c) and
the getethers code but they are not quite specific about
pushing a packet (like the rolled up RARP frame) down the stream.

Do i want to basically implement a packet filter to reap the
rarpd response like the arp example in the nit_xx(4) man page
or what?

Any help would be appreciated, TIA.
Please email your response as i don't read this group regularly.

regards, cal
- --
Cal Crowley  (hm: ccrowley@halcyon.com,  wk: crowley@lewis-gw1.army.mil)
"One size fits all, doesn't fit anything"
           ... a package or class or library developer's standard nightmare.

--
Cal Crowley  (hm: ccrowley@halcyon.com,  wk: crowley@lewis-gw1.army.mil)
"One size fits all, doesn't fit anything"
           ... a package or class or library developer's standard nightmare.

-----------[000392][next][prev][last][first]----------------------------------------------------
Date:      16 Dec 1994 18:43:30 -0700
From:      pete@dswi.com (Pete Kruckenberg)
To:        comp.protocols.tcp-ip
Subject:   Info needed on digital interfaces/protocols

Hi!

I'm making my first venture into the world of Unix drivers, and need
to get some information for this project. I'm writing a driver for a
synchronous card to be used with the free-ware Unix clone called
Linux. What I really need is some pointers to books, RFC's, on-line
stuff, etc, that would help me understand how to interface with DDS
(and eventually frame-relay and ISDN) connections. Specifically, I
need to know what protocols are used, how those protocols work, and
how network protocols are shipped out over digital lines.

I guess I'm looking for information in two areas: general digital
communications information that is geared towards the protocols that
are used (sort of a tutorial), and specific information about how
those protocols work, and how protocols from higher layers are
accommadated/encapsulated.

I would appreciate any recommendations of books, or pointers to
on-line resources to help me in a rather quick introduction to this
field before I start to actually work on this driver.

By the way, for those of you who are interested, the driver is for an
ISA-bus board that handles 2.4kbps to 6Mbps synchronous traffic, and my
ultimate goal is to have kernel-space drivers for raw communication
with the card, then user-space drivers to handle HDLC, frame-relay,
and ISDN connections.

Thanks.
Pete Kruckenberg
pete@dswi.com


-----------[000393][next][prev][last][first]----------------------------------------------------
Date:      16 Dec 1994 20:28:52 -0800
From:      shunt@eis.calstate.edu (Steve J. Hunt)
To:        comp.protocols.tcp-ip
Subject:   Re: Urgent: SMTP Server for DOS

juanm@ac.upc.es (Juan Manuel Bernabeu) writes:

> I'm looking desperatly for a DOS-TSR based SMTP server program.
> 
> We have no NetWare server, so we can't use pegasus/mercury. :(
> 

How about running Linux on a PC?  That would do what you want.  And 
Linux is pretty cheap.  Pegasus doesn't require a Netware server 
anymore, but it can't act as an "SMTP server" anyway.

I don't think there are any DOS or Windows based sendmail type 
programs.

Steve Hunt
shunt@eis.calstate.edu
    

-----------[000394][next][prev][last][first]----------------------------------------------------
Date:      16 Dec 1994 18:57:33 -0500
From:      barmar@nic.near.net (Barry Margolin)
To:        comp.protocols.tcp-ip
Subject:   Re: Source Routing?

In article <3cq2oj$oub@Mercury.mcs.com> les@MCS.COM (Leslie Mikesell) writes:
>Does everything observe source routing when it is used (unless it is
>specifically filtered)?  

RFC 1009 says:

      It is important for gateways to implement both the Loose and
      Strict Source Route options.  The Record Route and Timestamp
      options are useful diagnostic tools and must be supported in all
      gateways.

>			  What equipment will generate source routes?

Any TCP/IP host should be able to.  RFC 1122 says:

                 A host MUST support originating a source route and MUST
                 be able to act as the final destination of a source
                 route.

                 If host receives a datagram containing a completed
                 source route ...
                 the option as received (the recorded route) MUST be
                 passed up to the transport layer (or to ICMP message
                 processing).  This recorded route will be reversed and
                 used to form a return source route for reply datagrams
                 (see discussion of IP Options in Section 4).

>Can it be used to allow access out of a private net (say 10.x.x.x)
>by specifying a route of the 'public' connection plus the private
>address that other routers wouldn't know about?  

If the gateway doesn't filter these addresses, it probably could.  However,
there are probably many parts of the net that wouldn't be reachable that
way.  Many sites disable source routing on their Internet gateway, to
prevent them from spoofing the packet filters.  And I suspect that many
UDP-based applications don't use the reversed route when sending replies;
if the UDP API doesn't use "pseudo-connections", it's up to the application
to get the source route and reverse it, and I can't recall seeing any Unix
applications that do this (I'm not even sure *how* to do it on Unix -- you
may have to use the pseudo-connection interface to accomplish it).

>						  Or does an RFC 1631
>translating gateway exist that would be a better approach?

Either that, or application-level gateways.
-- 

Barry Margolin
BBN Internet Services Corp.
barmar@near.net

-----------[000395][next][prev][last][first]----------------------------------------------------
Date:      16 Dec 1994 19:32:49 -0500
From:      barmar@nic.near.net (Barry Margolin)
To:        comp.protocols.tcp-ip
Subject:   Re: TFTP

In article <3cscde$6u5@xmission.xmission.com> tlacy@xmission.com (Terry Lacy) writes:
>My server was identifying a conversation by the client's IP address and port number.
>I discovered that some TFTP clients seem to start multiple connections using the
>same port number, so I had to start identifying the conversation by the socket number.
>I suppose this is possible in a situation where the client starts a new task for
>every transfer, and generates port numbers randomly.  Is it possible to identify the
>conversation by the combination of server port number, client IP address and client
>port?  Is there ANY possibility that this would not be unique?

If the server generates a unique local port number for each initial request
it receives (on port 69), it should be possible to identify a conversation
simply by the server port number.  The simplest solution is to iterate
through the port number space, so that it will take a while for a port to
be reused.  This way, it's unlikely that a packet from an old connection
will be viewed as a new request on a recent connection.

>8.  Server receives an error packet on the second conversation.  NOW, do I terminate
>    this conversation???  According to RFC1350, "TFTP recognizes only one error
>    condition that does not cause termination, the source port of a received packet
>    being incorrect.  In this case, an error packet is sent to the originating host."
>    Logic would dictate that the server should terminate the connection, but the RFC
>    seems to contradict this.  In any event, the second conversation will NEVER
>    receive an ACK for the data packet it sent, and would eventually time-out and
>    terminate.

You shouldn't terminate the connection because your scenario isn't the only
reason why the server might receive a packet from the wrong source port.
It could be a packet from an old connection that was using the same server
port (perhaps the client system hung in mid-session, and the server
eventually terminated due to a timeout).  You need to tell that old client
that its server no longer exists, but the server must stick around to
continue responding to the current client.  There's no way to tell the
difference between these two cases, so the server must make the optimistic
assumption.

-- 

Barry Margolin
BBN Internet Services Corp.
barmar@near.net

-----------[000396][next][prev][last][first]----------------------------------------------------
Date:      16 Dec 1994 13:40:00 -0000
From:      josh@nagos.lif.icnet.uk (Joshua Kafeero)
To:        comp.protocols.tcp-ip
Subject:   FAQ Available ..???


Is there an faq for this group? If so could
you please tell me where I could get a copy?

Josh.



-----------[000397][next][prev][last][first]----------------------------------------------------
Date:      16 Dec 1994 14:25:00 GMT
From:      ntaib@silver.ucs.indiana.edu (Iskandar Taib)
To:        comp.infosystems.www.users,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.infosystems.www.providers,comp.os.ms-windows.networking.tcp-ip,comp.os.ms-windows.networking.misc
Subject:   Re: WWW Server/Client without INTERNET??

In article <3bkmgo$b0u@nntp.gov.ab.ca>,
Gordon McAndrew  <gordon@aec.env.gov.ab.ca> wrote:
>maad@iastate.edu (Madhu Rajamani) wrote:
>> 
>> 
>> I need to run a World Wide Web server (on a PC or PC based network)
>> within an organization which does NOT have internet access.  Also, I'd like
>> to be able to dial in to the local network (or standalone PC) and access 
>> HTML documents using MOSAIC or NETSCAPE.
 
>> Madhu Rajamani
>> 
>
>There is no requirement for an Internet connection to run a WWW server.
>But you probably have to be using the TCP/IP protocol on the LAN.  If you
>have a SLIP server of some kind you will be able to dial in and use whatever
>services are on the LAN.  You will probably have to use the IP addresses
>directly or set up a DNS on the LAN.
>
>We use local FTP & WWW servers to give out info to the staff.

If you have a Novell Lan or equivalent just have a drive mapping that
points to a place on the server where the html files live, and you can
use the Open Local File option to get to files. You can also write
links using the file://  option for "local file" links.




-- 
-------------------------------------------------------------------------------
Iskandar Taib                          | The only thing worse than Peach ala
Internet: ntaib@silver.ucs.indiana.edu |    Frog is Frog ala Peach
Home page: http://bigwig.geology.indiana.edu/iskandar/isk2.html

-----------[000398][next][prev][last][first]----------------------------------------------------
Date:      16 Dec 1994 14:33:19 GMT
From:      jspectre@netaxs.com (johnathan spectre)
To:        comp.protocols.tcp-ip
Subject:   bootp..?

I'm interested in knowing more about bootp servers. Could someone
point me to a book or net-available reference..? Is the server software
freely available (for AIX)? I know very little about them but understand
it would act as a dynamic address server for a variety of devices (solving
some of my problems).


--
Johnathan Spectre                  |             the truth is out there...
                                   |                finger for PGP 2.6 key
jSpectre@netaxs.com                |            support public key privacy

-----------[000399][next][prev][last][first]----------------------------------------------------
Date:      16 Dec 1994 15:05:38 GMT
From:      bwc@phoenix.cs.uga.edu (Brantley Coile)
To:        comp.protocols.tcp-ip
Subject:   Re: IP Network Address Translation daemon?

Marcus J Ranum writes:
>>does anyone know of daemon code to provide Network Address Translation
>
>	One approach is to put the network behind a firewall that
>only has one (legal) connection to the Internet. Other approaches
>involve expensive routers; I think several router makers do some form
>of address translation but it doesn't work very well. Remember
>that some protocols on TOP of TCP encode addresses in them: FTP is
>just one example. Dual-homed bastion firewalls effectively do the
>translation but don't give you direct connectivity.

We at Network Translation have a drop-in solution that includes
hardware and software that take 5min to install and does
address translation on the fly at near ethernet speeds.
We handle the FTP case and also provides screening based on
connections thru the box.  We don't use proxies.  The box is not
a router.  It's call PIX, private Internet Exchange.

>	If direct connectivity is your goal, The Correct Answer is
>to fix the addresses -- otherwise the network will just keep growing
>and sooner or later it'll come back to haunt you.
>

This is NOT the correct answer.  We are running out of addresses and
why should a thousand PC and MACs in a company consume addresses
when the only want to connect to outside servers?

A product like the PIX allow machines to share addresses because addresses are
dynamicly assigned.  You can use reserved addresses from rfc1597
and make the layout of your network much easier.  How would you like
a class A address?  You can have one it you use 10.0.0.0 and translate
addresses.

Our PIX is a PBX analogue.  If every phone in the contry had to have
an inword dialing number we would have run out of numbers long ago.

Brantley Coile
Network Translation
bwc@translation.com


-----------[000400][next][prev][last][first]----------------------------------------------------
Date:      Fri, 16 Dec 1994 15:31:16 GMT
From:      ytan@xvnews.unconfigured.domain (Yiwen Tan - x7236)
To:        comp.protocols.tcp-ip
Subject:   clean up tcp/ip ports


I have a utility running TCP/IP on a Sun workstation.  When I killed it by Ctrl-C and tried to run it again, I got the message "bind address already in use".  Is there a utility in the SunOS,  similar to the ipcrm, to clean it up? Thanks.

Yiwen


-----------[000401][next][prev][last][first]----------------------------------------------------
Date:      Fri, 16 Dec 94 15:41:03 GMT
From:      tlacy@xmission.com (Terry Lacy)
To:        comp.protocols.tcp-ip
Subject:   TFTP

Has anyone here had experience programming a TFTP server or client?  I'm looking for
someone who knows about possible pitfalls that are not in RFC1350.

I have been working on a TFTP server that runs in MS-Windows using Winsock.
Here's just one of my problems:

My server was identifying a conversation by the client's IP address and port number.
I discovered that some TFTP clients seem to start multiple connections using the
same port number, so I had to start identifying the conversation by the socket number.
I suppose this is possible in a situation where the client starts a new task for
every transfer, and generates port numbers randomly.  Is it possible to identify the
conversation by the combination of server port number, client IP address and client
port?  Is there ANY possibility that this would not be unique?

I am also interested in knowing what happens in the following situation:

1.  Client sends a read request.
2.  Client doesn't get a response within it's REXMT time, so it sends another
    read request.
3.  Server receives the first read request and sends the first data packet.
4.  Client receives the data packet and sends an ACK.
5.  Server receives the second read request and sends another first data packet.
6.  Client receives the first data packet again.  This time, assuming the server set
    up a new conversation with a different port number, the client sees the wrong
    server port number with that packet, so it sends an error back to the server.
7.  Server receives an ACK on the first conversation, so it sends the second data
    packet.
8.  Server receives an error packet on the second conversation.  NOW, do I terminate
    this conversation???  According to RFC1350, "TFTP recognizes only one error
    condition that does not cause termination, the source port of a received packet
    being incorrect.  In this case, an error packet is sent to the originating host."
    Logic would dictate that the server should terminate the connection, but the RFC
    seems to contradict this.  In any event, the second conversation will NEVER
    receive an ACK for the data packet it sent, and would eventually time-out and
    terminate.

What other pitfalls are there?  Any help would be greatly appreciated.

Terry Lacy
tlacy@xmission.com


-----------[000402][next][prev][last][first]----------------------------------------------------
Date:      16 Dec 1994 16:24:43 GMT
From:      dzoey@apexgrp.com (Joe Herman)
To:        comp.protocols.tcp-ip
Subject:   CIDR Question


Hello, 

I am a little confused about IP network number assignment and CIDR.
When I first read about CIDR, I was concerned that it meant that
you could not switch IP providers without switching your network 
number (and renumbering all your hosts (DHCP isn't available yet
for enough OS types here))

Several people on the net told me that was the wrong interpretation,
and you could in fact switch providers and keep a CIDR network number
from your old provider.  However, I've been helping one of our
customers connect to the internet, and I keep hearing from the
service provider's sales people that you have to renumber if you
switch vendors.

So, in practice, has anyone with a CIDR number been able to switch
providers and keep their number?  Are the salesfolks mistaken?

Joe Herman
dzoey@apexgrp.com
--
Everything is wonderful until you know something about it.


-----------[000403][next][prev][last][first]----------------------------------------------------
Date:      16 Dec 1994 16:57:10 GMT
From:      juanm@ac.upc.es (Juan Manuel Bernabeu)
To:        comp.protocols.tcp-ip
Subject:   Urgent: SMTP Server for DOS

Hi All! :)

I'm looking desperatly for a DOS-TSR based SMTP server program.

We have no NetWare server, so we can't use pegasus/mercury. :(

If possible, it should also be free/shareware.

Please reply! Thankx! :)

--
Salutacions, 
Pere.
perec@freenet.fsu.edu

-----------[000404][next][prev][last][first]----------------------------------------------------
Date:      Fri, 16 Dec 1994 17:03:20 GMT
From:      mcummin@relay.nswc.navy.mil (Mary A. Cummings)
To:        comp.protocols.tcp-ip
Subject:   looking for bootp source code

I am in search of some source code (perferably a client and server 
implementation) of the bootstrap protocol, bootp, to be used in conjunction
with the trivial file transfer protocol, tftp.  Does anyone know
where I can get this?

If this is not the right newsgroup for this, can you tell me which
newsgroup would be better?  thanks.

Mary Ann Cummings
mcummin@relay.nswc.navy.mil



-----------[000405][next][prev][last][first]----------------------------------------------------
Date:      Fri, 16 Dec 94 17:57:15 GMT
From:      arodwin@Shiva.COM (Andrew Rodwin)
To:        comp.protocols.tcp-ip
Subject:   RIPV2

Does anyone have a list of router vendors whose products support
RIPV2, or who have announced intentions to implement RIPV2?

Thanks,
Andrew Rodwin

-----------[000406][next][prev][last][first]----------------------------------------------------
Date:      Fri, 16 Dec 1994 18:37:26 GMT
From:      wunder@nosferatu.hpl.hp.com (Walter Underwood)
To:        comp.protocols.tcp-ip
Subject:   Re: Interpretation of Nagle algorithm

Rick Jones (raj@cup.hp.com) wrote:
: ... My understanding was that Nagle was intended to keep the ratio
: of data to headers "reasonably" large.

The opposite, actually. The most important thing is to keep the ratio
from getting "unreasonably small". I was working in a neighboring
department when John Nagle ran into the problem and proposed his fix.

At Ford Aerospace, we were running BSD 4.1 with the UNET networking
software from Ungermann/Bass. UNET sent data immediately, so holding
down a repeating key on a telnet connection would send a series of
packets with one byte of data and forty bytes of header. Queue up a
whole bunch of these tinygrams for your 9600 baud link (that's how
were were connected to the Internet), and you're in big trouble. When
you get something from the other end, there is no way to get an ACK
back in time, since you've already got many seconds worth of data in
the transmit queue. Byebye connection.

I remember watching John demonstrate the problem. He'd log into a
DEC-20 at Stanford, hold down the 'x' key, wait for the echo to stop
coming back, then wait for the connection to drop. Worked every time.

The algorithm is helpful for situations other than 9600 baud
connections of course, but those are coming back into vogue.  All
those folks with "Internet Starter Kit for Macintosh" and a V.32 modem
probably care a lot about tinygrams, whether they know it or not.

wunder

-----------[000407][next][prev][last][first]----------------------------------------------------
Date:      Fri, 16 Dec 1994 21:52:50 GMT
From:      rachel@ritz.mordor.com (Rachel Panzino)
To:        comp.protocols.tcp-ip
Subject:   Good book on sockets

Are there any good books out there on UNIX Networking and Sockets
programming??

thanks!
Rachel

-----------[000408][next][prev][last][first]----------------------------------------------------
Date:      16 Dec 1994 23:14:28 GMT
From:      Bob Stein <stein@gcomm.com>
To:        comp.protocols.tcp-ip,comp.dcom.lans
Subject:   Re: packets sent to own address

> What happens exactly to a packet which is sent to it's own address --> 

I was just wrestling with a similar issue regarding Ipswitch's 
Piper/IP for DOS.  It appears the packet driver interface, pk0,
is smart enough to reflect packets with the host IP address.
But the SLIP driver, sl0, is not.  It sends the packets through
the SLIP connection.  For most of our customers, this is fine 
because the SLIP server reflects the packet.  But for one customer,
the SLIP server wouldn't do this.  This guy could PING everyone
on the Internet but himself...

-- Bob Stein, Galacticomm

-----------[000409][next][prev][last][first]----------------------------------------------------
Date:      Sat, 17 Dec 1994 00:35:02 GMT
From:      morgan@world.std.com (William M Stair)
To:        comp.protocols.tcp-ip
Subject:   HELP! My first TCP/IP server program doesn't respond :^(

I've gotten just about EVERYTHING working, except my TCP/IP stream server
doesn't allow anyone to connect to it.
 
It's basically just a daytime server called "TCPdaytimed" that uses
argv[1] to determine what port to listen to.  I added the line:
 
        masterq   256/tcp
 
to /etc/services and everything in the server seems to run just fine
(if it's root) until it gets to the "accept" call.  Then it just sits
there blocked for ever.
 
It's clear that it's resolving the machine name, and what port # to
connect to just fine, so it's not an obvious installation problem, and
I know it's NORMAL for "accept" to block, but when I to CONNECT to
that port with a "telnet localhost 256" the server continues to sit
there, and my telnet session reports:
 
        % telnet localhost 256
        Trying 127.0.0.1 ...
        telnet: connect: Connection refused
also
        % telnet localhost masterq
        Trying 127.0.0.1 ...
        telnet: connect: Connection refused
 
but when I telnet to the normal daytime port (13) I get:
 
        % telnet localhost 13
        Trying 127.0.0.1 ...
        Connected to localhost.
        Escape character is '^]'.
        Fri Dec 16 18:53:42 1994
 
which is what I'd expect from MY server.  ARGH!  If you get the chance
to look at the code for the server, I included part of it at the
bottom of this post.  It's pretty much straight out of the book
"Internetworking with TCP/IP" volume III by Douglas E. Comer and David
L. Stevens.
 
I didn't want to include the whole thing, because (a) it's really not
my code, and (b) saving bandwidth is a good thing.
 
Could the problem be in the sockaddr parameter to bind()???  Many
thanks for all the feedback everyone.  I'd never have gotten this far
without the help!!
 
-Morgan <morgan@bur.visidyne.com>                          "Chaos is great!"
----------------------------------------------------------------------------
Visidyne Inc, 10 Corporate Place, South Bedford Street, Burlington MA, 01803
 
 
 
/********************************************************************/
/* TCPdaytimed.c - main */
 
#include <sys/types.h>
#include <sys/socket.h>
#include <netinet/in.h>
 
#include <stdio.h>
 
extern int errno;
extern char *sys_errlist[];
 
long time ();
 
#define QLEN 5
 
/*------------------------------------------------------------
 * main - Iterative TCP server for DAYTIME service
 *------------------------------------------------------------
 */
int
main (argc, argv)
     int argc;
     char *argv[];
{
  struct sockaddr_in fsin;      /* the from address of a client */
  char *service = "daytime";    /* service name or port number */
  int msock, ssock;             /* master & slave sockets */
  int alen;                     /* from-address length */
 
  switch (argc)
    {
    case 1:
      break;
    case 2:
      service = argv[1];
      break;
    default:
      IpcErrExit ("usage: TCPdaytimed [port]\n");
    }
 
        /**************************************************************/
        /* IpcPassiveTCP basically does:                              */
        /*                                                            */
        /*   struct sockaddr_in sin = (bunch of stuff);               */
        /*                                                            */
        /*   s = socket(PF_INET, SOCK_STREAM, TCP);                   */
        /*   bind (s, (struct sockaddr *)sin, sizeof(sin));           */
        /*   listen (s, QLEN);                                        */
        /*   return s;                                                */
        /*                                                            */
        /* According to all the debugging, each of these steps        */
        /* APPARENTLY works okay.                                     */
        /**************************************************************/
 
  msock = IpcPassiveTCP (service, QLEN);
 
  while (1)
    {
      ssock = accept (msock, (struct sockadr *) &fsin, &alen);
      if (ssock < 0)
        IpcErrExit ("accept failed: %s\n", sys_errlist[errno]);
      (void) TCPdaytimed (ssock);
      (void) close (ssock);
    }
}
 
/*------------------------------------------------------------
 * TCPdaytimed - do TCP DAYTIME protocol
 *------------------------------------------------------------
 */
int
TCPdaytimed (fd)
     int fd;
{
  char *pts;                    /* pointer to time string */
  time_t now;                   /* current time */
  char *ctime ();
 
  (void) time (&now);
  now = now;
  pts = ctime (&now);
  (void) write (fd, pts, strlen (pts));
  return 0;
}

-----------[000410][next][prev][last][first]----------------------------------------------------
Date:      Sat, 17 Dec 1994 09:40:48 LOCAL
From:      rymerson@post.QueensU.CA (Jim Rymerson)
To:        comp.protocols.tcp-ip
Subject:   Re: Telnet for 3270?

In article <3courm$j97@erinews.ericsson.se> etxmesa@eos.ericsson.se (Michael Salmon) writes:
>From: etxmesa@eos.ericsson.se (Michael Salmon)
>Subject: Re: Telnet for 3270?
>Date: 15 Dec 1994 08:30:46 GMT
 
>In article <1994Dec14.131950.6067@venus.gov.bc.ca>
>glittle@bcsc02.gov.bc.ca (Glen Little) writes:
>|> Does anyone know of a Windows based Telnet for 3270 program? Preferably
>|> free or share-ware, but I'll consider commerical ones.
 
>Netmanage Chameleon has a good 3270 emulator, if any 3270 emulator can
>be considered good that is.
 
>-- 
 
>Michael Salmon
 
>#include        <standard.disclaimer>
>#include        <witty.saying>
>#include        <fancy.pseudo.graphics>
 
>Ericsson Telecom AB
>Stockholm

You may want to look at QWS3270.  It is a winsock compliant TN3270 emulator.  
It is available from ftp.ccs.queensu.ca in the pub/msdos/tcpip directory or 
from the winsock directory at CICA or any of it's mirrors.  

Jim.

-----------[000411][next][prev][last][first]----------------------------------------------------
Date:      17 Dec 1994 01:41:11 GMT
From:      vql0608@is.nyu.edu (Vincent  Lai)
To:        alt.winsock,comp.infosystems.www.users,comp.protocols.ppp,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc
Subject:   Chameleon and Mosaic trouble

Hi:

Having a little trouble running Mosaic (20a7 version, I believe).  I am doing
so using Chameleon 4.012 from NetManage and win32s from Microsoft (version 
1.18 I think.  I do know it is the latest version from the Microsoft FTP 
site, filename is pw1118.exe).  When I attempt to run Mosaic, I get the 
following GPF error:

Application error: Growstub caused a GPF in module POINTER.DLL at 0001:0F2E

Mosaic is still up and running and probably purrs like a kitten, but my mouse 
pointer completely goes away, and I never really get a chance to get 
that pointer back.

I have tried contacting NCSA, but no response so far.  

Thanks

Vincent Lai


-----------[000412][next][prev][last][first]----------------------------------------------------
Date:      17 Dec 1994 10:08:50 -0500
From:      randall@uspto.gov (Rick Randall)
To:        comp.protocols.tcp-ip
Subject:   Re: WinNT & Windows TCP/IP connection

>Windows NT 3.5 supports TCP/IP via both PPP and SLIP/CSLIP with its
>Remote Access Services (RAS) Server.
>
>This will work only with another Windows NT 3.5 box or a Windows 95
>box.

	Typical proprietary garbage from Micro$oft.  *Any* PPP or CSLIP
software on a particular host __should__ be able to connect to *any* PPP/CSLIP
software on a remote system, given the right handshaking parameters.


 ~ rr



-----------[000413][next][prev][last][first]----------------------------------------------------
Date:      17 Dec 1994 04:43:10 GMT
From:      arice2@mason1.gmu.edu (Adam L Rice)
To:        comp.protocols.tcp-ip
Subject:   who has ppp for mac?

t are the names of a few good ppp programs for mac?  I have no problem
with the cost (under $300)

arice@gmu.edu


-----------[000414][next][prev][last][first]----------------------------------------------------
Date:      Sat, 17 Dec 1994 13:19:25 -0500
From:      rstory@revelstone.com (Robert Story)
To:        comp.protocols.tcp-ip
Subject:   Re: Good book on sockets

In article <D0xBG2.67r@ritz.mordor.com>, rachel@ritz.mordor.com (Rachel
Panzino) wrote:

:Are there any good books out there on UNIX Networking and Sockets
:programming??

Yep, "UNIX Network Programming" by W. Richard Stevens..

-- 
|   Robert Story    | PEI FAQ maintainer |#include <std/disclaimer.h>|
| Atlanta, Georgia  |--------------------|    War Damn Eagle! ' 90   |
|    Go Braves !    |   rstory@crl.com   |   rstory@mindspring.com   |

-----------[000415][next][prev][last][first]----------------------------------------------------
Date:      17 Dec 1994 19:38:29 -0800
From:      bob@osomil.ipinc.com (Robert Crowe)
To:        comp.protocols.tcp-ip
Subject:   Re: FIREFOX

In article <9412171348.0JE4500@infoboard.be>,
 <ronny.huybrechts@infoboard.be> wrote:
>
>Has anybody ever heart of
>the FIREFOX protocol ?
>It *should* be better than TCP/IP ?
>But it's brandnew ?
>Does anybody use it already ?
>Thanks for replying
>

If your talking about the Firefox from Novix, its not really a
protocol per se, rather it allows clients connected to an IPX 
network access to TCP/IP services.  Its a novell NLM and some client
software which acts as a sort of protocol translator.

In theory, you can allocate a pool of IP addresses on your Novell
server, and as clients require access to TCP/IP it assigns them
out dynamically.  This allows for a few nice features:  you don't
have to run a TCP/IP stack on each PC, and you don't have to assign
each machine a static IP address.   At least this is my understanding
of how it works.  Unfortunately, we had an eval copy, but never 
quite got it working properly, so I can't say how well it actually
works.

Cheers,

Bob.


-- 
-------------------------------------------------------------------------
Robert Crowe                                                bob@ipinc.com
Internet Products, Inc,                               (619) 576-4100 x101
8947-A Complex Drive San Diego, Ca. 92123             FAX: (619) 576-4111

-----------[000416][next][prev][last][first]----------------------------------------------------
Date:      17 Dec 1994 07:45:52 GMT
From:      colin@mayfield.hp.com (Colin Wynd)
To:        comp.protocols.tcp-ip
Subject:   Re: HELP! My first TCP/IP server program doesn't respond :^(

: It's basically just a daytime server called "TCPdaytimed" that uses
: argv[1] to determine what port to listen to.  I added the line:
:  
:         masterq   256/tcp

1) Are you passing argv[1] to it?

2) You can check by doing

      netstat -an | grep 256

   Once the server is running and you should see a line like:-

       tcp        0      0  *.256   *.*      LISTEN

   If you don't then you're not bind'ing to that port and you should check
   your source.


Cheers
Colin
-//-


-----------[000417][next][prev][last][first]----------------------------------------------------
Date:      17 Dec 1994 09:14:26 GMT
From:      tli@cisco.com (Tony Li)
To:        comp.protocols.tcp-ip
Subject:   Re: Netmask on seperate network

In article <3cr5dk$4ca@sparc.occ.uky.edu> bealar@ndlc.occ.uky.edu (Andy
Beal) writes: 
    
    My question is this, can one netmask server more than one physical network?
    
    Here's the situation, at Daviess County Schools in Owensboro, KY, we 
    currently have a Netware 3.11 server, and an RS6000.  The State Bd. of 
    Edu, is bringing in their statewide WAN.  We have been assigned 62 IP 
    addresses, (example 170.182.38.10 - 170.182.38.72 these aren;t the actual 
    ones, I can't recite them from memory yet.)  Anyway, we have 62 IP 
    addresses, and one Netmask for the building network.  Our Netware server
    has 1 Token-Ring network card, 1 Ethernet Network card, and 1 ArcNet 
    card.  We have been told from one source that each network must have it's 
    own netmask.  We have been told from a slightly more reliable source, 
    that one netmask can be used on several networks.  
    Our Question is, can we use one netmask throughout the entire building, 
    or do we need 3 individual netmasks, one for each physical network?
    
You can one netmask throughout the entire building.  You need one _subnet_
for each physical network.  

Tony Li
cisco

-----------[000418][next][prev][last][first]----------------------------------------------------
Date:      17 Dec 1994 09:21:09 GMT
From:      tli@cisco.com (Tony Li)
To:        comp.protocols.tcp-ip
Subject:   Re: CIDR Question

In article <3csf0b$kv3@terminus.apexgrp.com> dzoey@apexgrp.com (Joe Herman)
writes: 
    
    I am a little confused about IP network number assignment and CIDR.
    When I first read about CIDR, I was concerned that it meant that
    you could not switch IP providers without switching your network 
    number (and renumbering all your hosts (DHCP isn't available yet
    for enough OS types here))
    
    Several people on the net told me that was the wrong interpretation,
    and you could in fact switch providers and keep a CIDR network number
    from your old provider.  However, I've been helping one of our
    customers connect to the internet, and I keep hearing from the
    service provider's sales people that you have to renumber if you
    switch vendors.
    
The sales people are pushing FUD.  Unless they are stating policies for
their own service provider, you do NOT have to renumber if you change
service provider.

Not renumbering does require some effort on the part of your new service
provider, and does cause other parts of the network to possibly see an
extra route, so there may be extra costs incurred by your new service
provider.  

Tony


-----------[000419][next][prev][last][first]----------------------------------------------------
Date:      17 Dec 1994 10:39:14 GMT
From:      penguin@ucc.gu.uwa.edu.au (Paul Wyatt Wagland)
To:        comp.protocols.tcp-ip
Subject:   ftp and the SYST command (HELP)

Hi all,

I am writing a ftpDaemon for OS/2, but instead of using the standard OS/2
output I am using a unix-like output, using Peter Lewis's ftp listing
guidelines.

The problem with this is:
Automated clients do a SYST, get OS/2, and then expect the LIST format to
be in OS/2 format, (this is a real drag BTW), and usually refuse to show
the output.

My question is:
Can I return UNIX to the SYST command, instead of OS/2. If I can't (and I
know the rfc says I can't), what is so evil about it? Is there any other
way around this problem?

If this is not the appropiate news group, could you tell me where to send
this question?

Many Thanks.

BTW. Could you please CC: any replies to penguin@tartarus.uwa.edu.au as
my news site tends to expire fairly fast :-(
The Penguin.
--
------------------------------+------------------------------------------------
Paul Wagland                  |Can _you_ get calorie free water?
penguin@ucc.gu.uwa.edu.au     |
penguin@tartarus.uwa.edu.au   |Team OS/2 member...and proud of it!

-----------[000420][next][prev][last][first]----------------------------------------------------
Date:      Sat, 17 Dec 94 11:20:23 CET
From:      olla@ssgrr.it (Antonello Olla)
To:        comp.protocols.ppp,comp.protocols.tcp-ip,comp.os.ms-windows.nt.misc,comp.os.ms-windows.setup
Subject:   HELP!!!    WfW 3.1 -> WinNT 3.5  PPP

I tried to connect a WfW 3.1 with Trumpet 2.0b to a Win NT 3.5 beta2 
using PPP trough two modems Zoom v32bis.
The Remote Access server is configured to accept dialin , netbeui and 
TCP/IP protocols, IP remap.
I have also made the modifies found at http://www.luc.edu/~tbaltru/faq/  
Q4B-11, but I amn't able to realize the connection.
After dialing I receive "connect", enter <ESC> and after few seconds, 
during Trumpet show the PPP evolution, the line go down.
Then in the NT Event Viewer I read Remote Access error 20077:
An error occurred in the Point to Point Protocol Module. The error code is 
the data:
0000: dc 02 00 00
I haven't the possibility to try the singole PCs, then I don't know if the 
problem is in the server or the client.
Can someone suggest me a method to resolve the problem?
Thank you in advance
				Antonello



-- 
Antonello Olla
Scuola Superiore G. Reiss Romoli
S.P. per Coppito Km 0.300
67010  L'AQUILA
ITALY
Tlx 600870 SSGRR I
Tel +39-862-336-384
Fax +39-862-336-363 (o 491)
Email olla@ssgrr.it


-----------[000421][next][prev][last][first]----------------------------------------------------
Date:      Sat, 17 Dec 94 13:48:14 +0200
From:      ronny.huybrechts@infoboard.be
To:        comp.protocols.tcp-ip
Subject:   FIREFOX


Has anybody ever heart of
the FIREFOX protocol ?
It *should* be better than TCP/IP ?
But it's brandnew ?
Does anybody use it already ?
Thanks for replying


-----------[000422][next][prev][last][first]----------------------------------------------------
Date:      Sat, 17 Dec 1994 17:42:26 GMT
From:      larry@rn.com (Larry Snyder)
To:        comp.protocols.tcp-ip,comp.protocols.ppp
Subject:   Livingston Number (for Portmasters)

I need the telephone number for Livingston to inquire about
the Portmaster...


larry


-- 
Larry Snyder, System Adminstrator, GatorNet Internet Connectivity
Lake Mary (Orlando), Florida
larry@rn.com

-----------[000423][next][prev][last][first]----------------------------------------------------
Date:      17 Dec 1994 18:20:12 GMT
From:      Larry Himes <lehimes@access.digex.net>
To:        comp.protocols.tcp-ip,comp.protocols.ppp
Subject:   Re: Livingston Number (for Portmasters)

larry@rn.com (Larry Snyder) wrote:
>
> I need the telephone number for Livingston to inquire about
> the Portmaster...
> 
> 
> larry

You can get all that from http://www.livingston.com.

Larry Himes.

-----------[000424][next][prev][last][first]----------------------------------------------------
Date:      17 Dec 1994 19:39:01 GMT
From:      Bob Stein <stein@gcomm.com>
To:        comp.protocols.tcp-ip
Subject:   Re: ftp and the SYST command (HELP)

> Automated clients do a SYST, get OS/2, and then expect the LIST format to
> be in OS/2 format, (this is a real drag BTW), and usually refuse to show
> the output.
> 
> Can I return UNIX to the SYST command, instead of OS/2. If I can't (and I
> know the rfc says I can't), what is so evil about it? Is there any other
> way around this problem?

I faced a similar situation when I wrote the FTP daemon for the 
Internet Connectivity Option for The Major BBS Bulletin Board System.  
The reply to SYST is supposed to come from the list of operating 
systems in the Numbers RFC, which is now RFC-1700.  I registered 
"THE-MAJOR-BBS" as an operating system (it's there in RFC-1700), and
use that in the reply to SYST.

Then I ran into the exact problem you describe with the FTP client
that America Online runs.  In a dramatic display of responsiveness,
AOL modified their software *overnight* to do the following:  Issue
SYST; if the response is among a small set (Unix, VAX, ...) treat
accordingly; if not, issue PWD; if the reply begins with a forward
slash (/), treat as if it's a Unix system and move on; if not, 
generate an error message.  The next day, we were able to log into 
our FTP site:

	ftp://gcomm.com/

from America Online.  Those guys are doing something right...

In other drama, I was able to achieve compatibility with the
following FTP clients simply by mimicing the Unix 9-word directory
listing (ls -l) format:

	NCSA Mosaic's FTP client for Windows and for X Window Servers
	Netscape's FTP client
	NcFTP by Mike Gleason
	NetManage's Chameleon for Windows
	Intercon's TCP/Connect II for Windows
		(BUT, TCP/Connect users have to enable "Don't use
		extra flags" unless your server supports non-standard
		"LIST -F" option, as most Unix sites appear to do.)

What FTP clients are giving you grief?  Can you tell me if they work
on gcomm.com?   :-)

> BTW. Could you please CC: any replies to penguin@tartarus.uwa.edu.au as
> my news site tends to expire fairly fast :-(

I'll try, but Netscape's Newsreader doesn't appear to be able to CC 
:-(  I'll find another way...

-- Bob Stein, Galacticomm


-----------[000425][next][prev][last][first]----------------------------------------------------
Date:      Sat, 17 Dec 1994 23:31:08 GMT
From:      creed@taligent.com (Creed Erickson)
To:        comp.protocols.tcp-ip
Subject:   Re: who has ppp for mac?

In article <3ctq8u$f9k@portal.gmu.edu>, arice2@mason1.gmu.edu (Adam L Rice)
wrote:

> t are the names of a few good ppp programs for mac?  I have no problem
> with the cost (under $300)
> 
> arice@gmu.edu

Available via anon ftp: 

ftp.uu.net:archive/systems/mac/info-mac/Communication/MacTCP/mac-ppp-201.hqx

Also availible from any other InfoMac mirror.

Requires (i.e., does not bundle) MacTCP and Comm Toolbox.

-- 
Creed Erickson - creed@taligent.com        Voice: +1 408 255 2525
10201 N. De Anza Blvd., Cupertino, CA 95014  FAX: +1 408 777 5380

-----------[000426][next][prev][last][first]----------------------------------------------------
Date:      18 Dec 1994 11:49:24 -0600
From:      vohra@willis.cis.uab.edu (Anuj Vohra)
To:        comp.lang.c++,comp.protocols.tcp-ip
Subject:   Class Lib. for Socket comm??

Hi,
  I am new to this news group and do not know if this information was 
posted before. I am interested in class libraries for socket (or TLI) 
communications.
Any pointers or information will be greatly appriciated.
Regards
Anuj

-- 
			vohra@cis.uab.edu
			vohra@atax.eng.uab.edu

-----------[000427][next][prev][last][first]----------------------------------------------------
Date:      Sun, 18 Dec 1994 04:06:50 GMT
From:      jtara@cts.com (Jon Tara)
To:        comp.infosystems.www.users,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.infosystems.www.providers,comp.os.ms-windows.networking.tcp-ip,comp.os.ms-windows.networking.misc
Subject:   Re: WWW Server/Client without INTERNET??

In article <3cs7vs$7nj@usenet.ucs.indiana.edu> ntaib@silver.ucs.indiana.edu (Iskandar Taib) writes:
>From: ntaib@silver.ucs.indiana.edu (Iskandar Taib)
>Subject: Re: WWW Server/Client without INTERNET??
>Date: 16 Dec 1994 14:25:00 GMT
 
>In article <3bkmgo$b0u@nntp.gov.ab.ca>,
>Gordon McAndrew  <gordon@aec.env.gov.ab.ca> wrote:
>>maad@iastate.edu (Madhu Rajamani) wrote:
>>> 
>>> 
>>> I need to run a World Wide Web server (on a PC or PC based network)
>>> within an organization which does NOT have internet access.  Also, I'd like
>>> to be able to dial in to the local network (or standalone PC) and access 
>>> HTML documents using MOSAIC or NETSCAPE.
 
>>> Madhu Rajamani
>>> 
>>
>>There is no requirement for an Internet connection to run a WWW server.
>>But you probably have to be using the TCP/IP protocol on the LAN.  If you
>>have a SLIP server of some kind you will be able to dial in and use whatever
>>services are on the LAN.  You will probably have to use the IP addresses
>>directly or set up a DNS on the LAN.
>>
>>We use local FTP & WWW servers to give out info to the staff.
 
>If you have a Novell Lan or equivalent just have a drive mapping that
>points to a place on the server where the html files live, and you can
>use the Open Local File option to get to files. You can also write
>links using the file://  option for "local file" links.

Yes, the latter approach will work fine as long as you don't need to run any 
CGI scripts. If you do, then you will need a real server, and will need to run 
TCP/IP over your LAN.

In either case, you will need a null Winsock, or a real one with it's SLIP or 
PPP interface disabled. I tried the null winsock that's floating around, and 
it bombed with Netscape, so I suspect that the real Winsock, disabled approach 
is the only way to go if you're not running TCP/IP.
________________________
 A new picture of San Diego Bay every half hour:
 <A HREF ="http://www.cts.com/~jtara/baycam.html">San Diego BayCam</A>
  jtara@cts.com

 

-----------[000428][next][prev][last][first]----------------------------------------------------
Date:      Sun, 18 Dec 1994 14:12:18 +0000
From:      Steve@starbug1.demon.co.uk (Steve Kay)
To:        comp.protocols.tcp-ip
Subject:   Re: FIREFOX

In article <3d0arl$lnu@osomil.ipinc.com>
           bob@osomil.ipinc.com "Robert Crowe" writes:

> In article <9412171348.0JE4500@infoboard.be>,
>  <ronny.huybrechts@infoboard.be> wrote:
> >
> >Has anybody ever heart of
> >the FIREFOX protocol ?
> >It *should* be better than TCP/IP ?
> >But it's brandnew ?
> >Does anybody use it already ?
> >Thanks for replying
> >
> 
> If your talking about the Firefox from Novix, its not really a
> protocol per se, rather it allows clients connected to an IPX 
> network access to TCP/IP services.  Its a novell NLM and some client
> software which acts as a sort of protocol translator.
> 
Bob's quite right - Firefox just sits on top of Novell and allows TCP/IP as well
as IPX.

It's an alternative to converting to Netware IP, which makes all Netware traffic
TCP/IP based (it puts a TCP/IP 'wrapper' around all IPX packets).  By using
Firefox, you only install the software on your fileserver, and all users can
connect to that server using IPX and then use TCP/IP for telnet/ftp etc from
there (as Firefox package includes Novell's LAN Workplace).

We're currently doing an eval on it right now - if anyone wants to know how it
goes, email me.

-- 
Steve Kay                   Tech Support: VME, UNIX, DOS, WINDOWS
Mirrlees Blackstone Ltd     Tel  :  0161 483 1000 Ext 2331
Bramhall Moor Lane          Fax  :  0161 487 1465
Hazel Grove                 email:  steve@starbug1.demon.co.uk
Stockport                   "Build a system any idiot can use and only
SK7 5AH                      an idiot will want to use it".

-----------[000429][next][prev][last][first]----------------------------------------------------
Date:      Sun, 18 Dec 1994 14:42:54 GMT
From:      adam@comptech.demon.co.uk (Adam Goodfellow)
To:        comp.protocols.tcp-ip
Subject:   Re: ftp and the SYST command (HELP)

In article <3cuf4i$vtg@styx.uwa.edu.au>,
                Paul Wyatt Wagland <penguin@ucc.gu.uwa.edu.au> wrote:

>Hi all,
>
>I am writing a ftpDaemon for OS/2, but instead of using the standard OS/2
>output I am using a unix-like output, using Peter Lewis's ftp listing
>guidelines.
>
>The problem with this is:
>Automated clients do a SYST, get OS/2, and then expect the LIST format to
>be in OS/2 format, (this is a real drag BTW), and usually refuse to show
>the output.
>
 
>My question is:
>Can I return UNIX to the SYST command, instead of OS/2. If I can't (and I
>know the rfc says I can't), what is so evil about it? Is there any other
>way around this problem?
>

Surely the sensible thing is to return a name that indicates how to
interpret the output.

There is no point in returning OS/2 for something that is going to look like
Unix - it would only lead to confusion with software that tries to parse a
directory listing IMHO.

Unfortunately for me, even after much pestering, the stupid poeple who make
my computer and OS havent even bothered to register the OS name, so I just
return UNIX for everything (FTP, Indent etc) and give unix like output for
everything - This solves all problems. (The standard dir output seems to do
a good job of crashing web browsers so I changed it to a unix style)

In the case of FTP, it would perhaps help if a standard was specified for
specific listing formats, and while typical unix output contains a lot of
garbage that most people arnt interested in, it is the most common, and
IHMO, should be adopted by all. Either that, or a new list command is needed
that indicates type (dir/file/link), size, date and name in a strict format.

-- 
Adam

=======================================================================
| Computech  Tel/Fax: 0181 673 7817  email: adam@comptech.demon.co.uk |
=======================================================================

-----------[000430][next][prev][last][first]----------------------------------------------------
Date:      Mon, 19 Dec 1994 02:29:56 -0600
From:      Tyson Douglas Cooper <tdc115@herald.usask.ca>
To:        comp.protocols.misc,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.os.os2.misc,comp.os.os2.programmer.misc
Subject:   Need help re. SLIP, REXX etc.

Ok, here's my situation: I currently access the internet via a shell 
account at the University that I attend. Our dial-up access system is 
incredibly limited, and because of this our computing services department 
isn't officially allowing SLIP access. I say "officially" because they do 
acknowledge that SLIP access is possible, but if we want to use it we 
must figure it out for ourselves. So, I have a friend in computing 
services who gave me his REXX script which he uses with OS/2 for a SLIP 
connection. Because he could get in trouble for this he said that 
everything I need to know is in this script, but that's all he could give me.

So, after all that my request is this: does anyone out there think they 
might be able to help me out on this? I need to know what I need to do to 
use this script that was given to me. I intend to use either Trumpet 
Winsock or Netmanage's Chameleon for this task, and if anyone out there 
knows how I would configure either of these programs with the info at the 
bottom, then I would certainly appreciate it. Please e-mail me directly.

Thank-you for your help.

Tyson Cooper
University of Saskatchewan

Here is the script that was sent to me:

From: ####################################
Date: Mon, 12 Sep 1994 13:45:09 -0600 (CST)
From: ####################################
To: Tyson.Cooper@sask.usask.ca
Subject: Re: how to SLIP? (fwd)


Here is the relevant parts of my REXX script for doing a SLIP connection:

parse arg interface , dialcmd username password

/*--------------------------------------------------------------------------*/
/*		     Initialization and Main Script Code		    */
/*--------------------------------------------------------------------------*/

/* Set some definitions for easier COM strings */
cr='0d'x
crlf='0d0a'x
lf='0a'x
response ='OK'

say ''
say 'SLIPUOFS - SLIP Script for U of S Terminal Server ',
    '(interface' interface')'

dialcmd="ATM1DT933-9400"
tsb_prompt ='Local>'
slipstr = 'SLIP Mask:'
slip_parse = 'SLIP Address:'
port_cmd = 'show port alt chara'||cr
serv_cmd = 'show server'||cr
slip_cmd = 'set port internet slip enabled'||cr
username="Kevin Lowey"

/* Flush any stuff left over from previous COM activity */
call flush_receive

/* Dial the remote server */
call charout , 'Now Dialing...'

/* Wait for connection */

redial_done = 0
do until (redial_done = 1)
  call send dialcmd || cr
  call flush_receive 

  listen_done = 0
  do until (listen_done = 1)
    call waitfor crlf 

    if (pos('CONNECT',waitfor_buffer) > 0) then do
      listen_done = 1
      redial_done = 1
    end

    if (pos('NO CARRIER',waitfor_buffer) > 0) then do
      listen_done = 1
    end

    if (pos('BUSY',waitfor_buffer) > 0) then do
      listen_done = 1
    end

    if (pos('RINGING',waitfor_buffer) > 0) then do
      listen_done = 1
    end
  end /* listen_done */
end

/* Handle login.  We wait for standard strings, and then flush anything */
/* else to take care of trailing spaces, etc..				*/
call send cr
call waitfor 'username>' ; call flush_receive 'echo'
call send username || cr
call waitfor 'Local>' ; call flush_receive 'echo'

call send port_cmd
call waitfor ('SLIP Mask:') ; call flush_receive 'echo'
parse var waitfor_buffer . 'SLIP Address:' a '.' b '.' c '.' d 'SLIP'  .

ip_addr = a||'.'||b||'.'||c||'.'||d

call send serv_cmd
call waitfor ('Ethernet:') ; call flush_receive 'echo'
say 'JANUS name received from terminal server .......'
parse var waitfor_buffer . 'JANUS' a .
j_addr = "128.233.20.1"||a
/*
  d = d - 100
  if (d > 20) then
    d = d - 20
*/
pname = "janus"||a||"_"||d

say 'The port for this session is ' pname '(' ip_addr ') on' j_addr

/* Flush anything else */
call flush_receive 'echo'

/* Now configure this host for the appropriate address, */
/* and for a default route through the Annex.		*/

call send slip_cmd
say 'SLIP Connection Established'

'ifconfig sl0 inet' ip_addr j_addr 'netmask 255.255.255.0'
'route -f add default' j_addr '1'
'start pmx'
'detach c:\mmos2\playwav.cmd c:\mmos2\sounds\startrek\lockcomp.wav'

/* All done */
exit 0

- Kevin Lowey (Kevin.Lowey@USask.CA, Client Services Educational Services Team)
  http://jester.usask.ca/~lowey/  --- gopher skynet.usask.ca

     I'm just collateral damage of the Political Correctness Movement

 
Steven Wright Says ...
You can't have everything ... where would you put it?




-----------[000431][next][prev][last][first]----------------------------------------------------
Date:      Sun, 18 Dec 1994 16:49:45 GMT
From:      shai@shekel.jct.ac.il (Shai Fultheim-System Assistant)
To:        comp.protocols.tcp-ip,comp.unix.programmer,comp.unix.questions
Subject:   nit_pf que

in the nit_pf manual it writes:
					      The  packet  filter
                itself  is  specified in the Pf_Filter array as a
                sequence   of   two-byte   commands,   with   the
                Pf_FilterLen  field giving the number of commands
                in the sequence.  This  implementation  restricts
                the  maximum  number  of  commands  in  a  filter
                (ENMAXFILTERS) to 40.  The next section describes
                the available commands and their semantics.

if i need to define filetr which i need for that more than 40
commands, i have big problem.
can i change it or i have to build new kernel???
can i use ioctl to push the first 40 cmd and after that push the other
commands, or if i do
  if (ioctl(control->nit_fd, I_PUSH, "pf"))
    fatal("\n\t    initdevice: I_PUSH pf");
it will overwrite the previos info???

what should i do???
pls help.
--
Shai Fultheim    				E-mail: shai@shekel.jct.ac.il
System Assistant				Tel (W):       (972)-2-751160
Jerusalem College of Technology - Israel	Tel (H):       (972)-2-963476

-----------[000432][next][prev][last][first]----------------------------------------------------
Date:      Sun, 18 Dec 1994 16:50:57 +0000
From:      Steve@scotia.co.uk (Steve Holden)
To:        comp.protocols.tcp-ip
Subject:   Re: Urgent: SMTP Server for DOS

In article <3ctpe4$l7t@eis.calstate.edu>
           shunt@eis.calstate.edu "Steve J. Hunt" writes:

> juanm@ac.upc.es (Juan Manuel Bernabeu) writes:
> 
> > I'm looking desperatly for a DOS-TSR based SMTP server program.
> > We have no NetWare server, so we can't use pegasus/mercury. :(
> How about running Linux on a PC?  That would do what you want.  And 
> Linux is pretty cheap.  Pegasus doesn't require a Netware server 
> anymore, but it can't act as an "SMTP server" anyway.
> 
And it's perfect for the DOS environment ... ;-)

> I don't think there are any DOS or Windows based sendmail type 
> programs.

The KA9Q variant I obtained from ftp.demon.co.uk as a part of their
DIS216C package includes DOS-compatible SMTP send and receive code.
I believe this is shareware, and I don't think it varies too radically
from the standard KA9Q code. So you should be able to get source.

If you don't mind some assembly being required you could probably
write a TSR using the KA9Q source as a base.
-- 
Steve Holden     Scotia Electronic Publishing  +---------------------------+
Tel: +44 436 678962            29 John Street  |    Tools, Training and    |
Fax: +44 436 677814 (bureau)      Helensburgh  | Technology for Successful |
Internet: steve@scotia.co.uk         Scotland  |   Electronic Publishing   |
Compuserve: 100343,2205               G84 8XL  +---------------------------+

-----------[000433][next][prev][last][first]----------------------------------------------------
Date:      Sun, 18 Dec 1994 17:05:53 +0000
From:      Steve@scotia.co.uk (Steve Holden)
To:        comp.protocols.tcp-ip
Subject:   Re: Good book on sockets

In article <D0xBG2.67r@ritz.mordor.com>
           rachel@ritz.mordor.com "Rachel Panzino" writes:

> Are there any good books out there on UNIX Networking and Sockets
> programming??
> 
> thanks!
> Rachel
> 
The whole scoop on sockets client/server code can be found in:

	Internetworking with TCP/IP Volume III:
	Client/Server Programming and Applications BSD Socket Version;
	Douglas E. Comer and David L. Stevens;
	Prentice-Hall International, 1993

The first volume in this series is an excellent introduction to the more
important elements of the Internet family of protocols, and the second gives
lots of good implementation examples.  Volume three is also available in an
alternative edition exemplifying the TLI interface instead of sockets.

I also have a lot of time for:

	UNIX System V Network Programming;
	Stephen A. Rago;
	Addison-Wesley, 1993

This book is rather broader in scope, and takes the experienced programmer
right through the UNIX networking APIs, including sockets and the Transport
Layer Interface.

Just two of the many available, and only my personal opinion.  You will
probably get many other equally good but different recommendations.
-- 
Steve Holden     Scotia Electronic Publishing  +---------------------------+
Tel: +44 436 678962            29 John Street  |    Tools, Training and    |
Fax: +44 436 677814 (bureau)      Helensburgh  | Technology for Successful |
Internet: steve@scotia.co.uk         Scotland  |   Electronic Publishing   |
Compuserve: 100343,2205               G84 8XL  +---------------------------+

-----------[000434][next][prev][last][first]----------------------------------------------------
Date:      Sun, 18 Dec 1994 17:22:34 GMT
From:      weiss@pesach.jct.ac.il (efi weiss)
To:        comp.protocols.tcp-ip
Subject:   internet "firewall" & proxy servers

what is a internet "firewall" ??
how can proxy servers help bypassing it ?
and what software needed to run this proxy server? (where can I find this
program? )
tnx
efi

-----------[000435][next][prev][last][first]----------------------------------------------------
Date:      18 Dec 1994 19:04:27 GMT
From:      mjr@tis.com (Marcus J Ranum)
To:        comp.protocols.tcp-ip
Subject:   Re: internet "firewall" & proxy servers

>what is a internet "firewall" ??

	It's a system that protects the boundary between 2 networks.

>how can proxy servers help bypassing it ?

	Bypassing it? As it "getting around it?"   That implies
attempting to defeat whatever protections it's put in place for,
which is a policy issue that should be taken up with whoever is
responsible for network security.

mjr.

-----------[000436][next][prev][last][first]----------------------------------------------------
Date:      Sun, 18 Dec 1994 22:21:10 +0000
From:      mpdillon@halcyon.com (Michael Dillon)
To:        comp.protocols.tcp-ip
Subject:   Re: who has ppp for mac?

In article <3ctq8u$f9k@portal.gmu.edu>,
arice2@mason1.gmu.edu (Adam L Rice) wrote:
> t are the names of a few good ppp programs for mac?  I have no problem
> with the cost (under $300)

Buy "The Internet Starter Kit for Macintosh" by Adam Engst, 2nd edition.

---------------------------------------------------------------------
Cool cats, brick bats, bad boys wearin' big hats
Surf's up, my cup, floating, flying, rising up.

Michael Dillon                    mpdillon@halcyon.com
C-4 Powerhouse, RR #2             michael@junction.net
Armstrong, BC   V0E 1B0           Fido: 1:353/350
Canada                             BBS: +1-604-546-2705

-----------[000437][next][prev][last][first]----------------------------------------------------
Date:      Mon, 19 Dec 1994 06:49:36
From:      m_hazel@enet.net (Mike Hazel)
To:        comp.protocols.tcp-ip,comp.os.ms-windows.programmer.networks
Subject:   Re: IPX Packet driver

In article <NEWTNews.5284.787443130.craign@craign.teleport.com> Craig Nelson <craign@teleport.com> writes:
>From: Craig Nelson <craign@teleport.com>
>Subject: IPX Packet driver
>Date: Wed, 14 Dec 94 14:09:22 PDT


>Hello, one and all.
 
>I'm looking for an IPX-TCP/IP packet driver over a SLIP link. That is,
>I have a SLIP link I can connect with. Thus I have TCP/IP. I would like
>an IPX packet driver to lay on top of it that will take IPX packets and
>"cook" them into TCP/IP to send them on their way.

Two ways....

ODIPKT.DRV (available at ftp.novell.com[I think])

Novell's LAN Workplace for DOS using SLIP_PPP and IPTUNNEL. IPTUNNEL is a 
driver for encapsulating IPX in IP and must be loaded at both the workstation 
and the server.. 

Mike

-----------[000438][next][prev][last][first]----------------------------------------------------
Date:      Mon, 19 Dec 1994 01:50:03 +0000
From:      Steve@scotia.co.uk (Steve Holden)
To:        comp.protocols.tcp-ip
Subject:   Re: WinNT & Windows TCP/IP connection

In article <3cuuu2$308@pioneer.uspto.gov>
           randall@uspto.gov "Rick Randall" writes:

> >Windows NT 3.5 supports TCP/IP via both PPP and SLIP/CSLIP with its
> >Remote Access Services (RAS) Server.
> >
> >This will work only with another Windows NT 3.5 box or a Windows 95
> >box.
> 
>         Typical proprietary garbage from Micro$oft.  *Any* PPP or CSLIP
> software on a particular host __should__ be able to connect to *any* PPP/CSLIP
> software on a remote system, given the right handshaking parameters.
>
I would agree.  Sadly, while doing a review of SunSelect's PC-NFS for the
Sun UK User Group Magazine I discovered that the PC-NFS PPP implementation
relies on seeing a fixed character string which, guess what, the Solaris
PPP login daemon just happens to provide, whereas the Demon Internet
server does not.

I don't know whether the incompatibility commented on previously goes
deeper than that.  I had (as an ex-employee) formerly held quite a good
opinion of Sun's approach to standards-compliance.

The problem seems to be that, while SLIP and PPP are standardised in RFC
documentation, there is little concensus about establishing the necessary
dial-in connection before the standardised protocol comes into play.
-- 
Steve Holden     Scotia Electronic Publishing  +---------------------------+
Tel: +44 436 678962            29 John Street  |    Tools, Training and    |
Fax: +44 436 677814 (bureau)      Helensburgh  | Technology for Successful |
Internet: steve@scotia.co.uk         Scotland  |   Electronic Publishing   |
Compuserve: 100343,2205               G84 8XL  +---------------------------+

-----------[000439][next][prev][last][first]----------------------------------------------------
Date:      Mon, 19 Dec 1994 06:55:04
From:      m_hazel@enet.net (Mike Hazel)
To:        comp.protocols.tcp-ip
Subject:   Re: WinNT & Windows TCP/IP connection

In article <3cuuu2$308@pioneer.uspto.gov> randall@uspto.gov (Rick Randall) writes:
>From: randall@uspto.gov (Rick Randall)
>Subject: Re: WinNT & Windows TCP/IP connection
>Date: 17 Dec 1994 10:08:50 -0500
 
>>Windows NT 3.5 supports TCP/IP via both PPP and SLIP/CSLIP with its
>>Remote Access Services (RAS) Server.
>>
>>This will work only with another Windows NT 3.5 box or a Windows 95
>>box.
 
>        Typical proprietary garbage from Micro$oft.  *Any* PPP or CSLIP
[snipped by reader]

NT 3.5 RAS SLIP/PPP will work with -many- not -all- PPP implementations, the 
trick is writing the script, particularly if PAP authentication is in use.

Mike

-----------[000440][next][prev][last][first]----------------------------------------------------
Date:      Mon, 19 Dec 1994 03:42:18 GMT
From:      davidp@qpsx.oz.au (David Pascoe)
To:        comp.protocols.tcp-ip
Subject:   any RIP-2 sources

Hi,

I have gotten hold of the ucsd hamradio copy of RIPv2 for NOS (whatever that 
is) and was wondering if there are any other RIP implementations available out 
there.

All I am really interested in, is a simple usage where RIP is used for 
exchanging routing information between peer routers internal to a network. We 
will only ever have say about 16 routing stacks inside our network, and no 
routing information will be allowed to be seen on a customers network.

cheers,
davidp.
--
David Pascoe, Jtec Pty. Limited, davidp@qpsx.oz.au, Fax:+61-9-321-5216

Yinkel, n.: A person who combs his hair over his bald spot, hoping no one
will notice.                                         -- Rich Hall, "Sniglets"

-----------[000441][next][prev][last][first]----------------------------------------------------
Date:      Mon, 19 Dec 1994 06:02:42 GMT
From:      paul@atlas.abccomp.oz.au (Paul Brooks)
To:        comp.protocols.tcp-ip
Subject:   Re: Winsock/ODI on TokenRing

In article <D0qG9y.4HC@cfa.org> peisch@cfa.org writes:
|Does this work?  We're trying to coax it, but with no luck.  The cisco has
|SNAP on it's token ring interface, but the pc's don't seem to be able to
|get out...

Depend's on who's "Winsock" you are using. "Windows Sockets" is an API
standard - each vendor implements their own Windows Sockets modules.

Our product has a Winsock interface, and runs quite happily over 802.5
token ring. Also 802.3 SNAP ethernet, if anyone cares. It is entirely
possible the implementation you bought does not - contact the people you
got it from and ask them, or post here with the name of the firm that 
created your "Winsock".


-- 
Paul Brooks              |paul@abccomp.oz.au       |Emerging Standard:
TurboSoft Pty Ltd        |pwb@newt.phys.unsw.edu.au|  one that has not yet
579 Harris St., Ultimo   |                         |  been superseded.
Sydney Australia 2007    |ph: +61 2 281 3155       |  

-----------[000442][next][prev][last][first]----------------------------------------------------
Date:      19 Dec 1994 14:35:42 -0500
From:      mskucher@noc.tor.hookup.net (Murray S. Kucherawy)
To:        comp.protocols.tcp-ip
Subject:   FTP RFC

Which documents define the FTP protocol?  I looked at the latest RFC inded
from nic.ddn.mil and there were none (on file) that were just protocol
definitions - at least none that I found.  There were lots that discussed
FTP or modified it, but no basic protocol specification.

Please reply by e-mail.

Thanks in advance...
-- 
Murray S. Kucherawy ==========================================================
Software Manager            HookUp Communication Corporation - Ontario, CANADA
E-mail: mskucher@hookup.net          NIC Handle "MSK"           (905) 847-8000

-----------[000443][next][prev][last][first]----------------------------------------------------
Date:      19 Dec 1994 16:03:01 -0500
From:      jhawk@panix.com (John Hawkinson)
To:        comp.protocols.tcp-ip
Subject:   Re: Source Routing?

In <ricard.36.006B524C@axis.se> ricard@axis.se (Ricard Wolf) writes:

>	In article <3cog2t$p8j@tools.near.net> barmar@nic.near.net (Barry Margolin) 
>writes:>From: barmar@nic.near.net (Barry Margolin)
>>Subject: Re: Source Routing?
>>Date: 14 Dec 1994 23:18:37 -0500
 
>>In article <3coe3q$7uf@mogan.cc.metu.edu.tr> ergul@ds5500.cc.boun.edu.tr (Salih Ergul) writes:
>>>       I'm hearing about source routing option of IP packets, but I
>>>don't know what it is and how to use it. Could anybody explain this?
 
>>This option allows an IP packet to specify the routers that should be used,
>...
 
>Is IP source routining really used in practice? I've seen it in the IP RFC, 
>and some implementations of 'ping' seem to be able to set up various source 
>routing fields, but I've never seen it being used in the real world.

Many versions of traceroute implement the -g option, which uses LSRR
traceroute.

The BSD4.4 telnet, the Cray telnet, and the MIT Athena telnet all
support @middlehost:destinationhost syntax for loose source routing,
and an extension for strict.

NetBSD's native telnet (the 4.4 telnet) supports this.

I find myself using it about once a week, and LSRR traceroute more
often than that.

--
John Hawkinson
jhawk@panix.com

-----------[000444][next][prev][last][first]----------------------------------------------------
Date:      Mon, 19 Dec 1994 13:21:13 GMT
From:      angusr@festival.ed.ac.uk (Angus Rae)
To:        alt.winsock,comp.infosystems.www.users,comp.protocols.ppp,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc
Subject:   Re: Chameleon and Mosaic trouble

Vincent  Lai (vql0608@is.nyu.edu) wrote:

: Application error: Growstub caused a GPF in module POINTER.DLL at 0001:0F2E
: Mosaic is still up and running and probably purrs like a kitten, but my mouse
: pointer completely goes away, and I never really get a chance to get
: that pointer back.

It's a bug with the Microsoft Mouse drivers I'm afraid, it happens with
a lot (if not all) Win32s applications.

An upgraded POINTER.DLL is available from Microsoft which turns off some
features while Win32s programs are running, thus "solving" the problem.

The file needed from the Microsoft software library is the wonderfully
named HD1061.EXE

--
Angus G Rae            Biological User Support Team, Edinburgh University
Email: Angus.Rae@ed.ac.uk     Personal Page: http://www.ed.ac.uk/~angusr/
The above views are mine, and Edinburgh Uni can't have any of them.


-----------[000445][next][prev][last][first]----------------------------------------------------
Date:      19 Dec 1994 15:11:41 GMT
From:      ceham@w3eax.umd.edu (Maurice De Vidts NE3S)
To:        comp.protocols.tcp-ip
Subject:   Packet Frame interference ?

Hi,

I have setup Trumpet winsock using packet drivers ( -n )
on a Netware 3.11 over our company LAN.  It seems to work
but our network administrator has prohibited me from using it
since he claims it crashes the server" 

I know I may be using a different frame type, but it seems
a little far fetched that different ethernet frames will
cause such a crisis. 

I would like some techincal pointers to understand
if I am truly creating a problem.  The novell winsock
driver he has installed does not work with any
of the winsock telnets I have found on the net. 


the other PC's on the LAN are using the typical IPXODI
interface and a typical net.cfg follows:

	Frame Ethernet_II
	Frame Ehternet_802.3
	Protocol IPX 0 Ehternet_802.3


I would appreciate a clarification on this issue !
 
ceham@w3eax.umd.edu 



-----------[000446][next][prev][last][first]----------------------------------------------------
Date:      19 Dec 1994 15:29:47 GMT
From:      colin@mayfield.hp.com (Colin Wynd)
To:        comp.protocols.tcp-ip
Subject:   Re: Good book on sockets

Rachel Panzino (rachel@ritz.mordor.com) wrote:
: Are there any good books out there on UNIX Networking and Sockets
: programming??

UNIX Network Programming by W. Richard Stevens  ISBN: 0-13-949876-1

or

Internetworking with TCP/IP (BSD Socket version, volume 3) by Douglas Comer and
David L. Stevens  ISBN: 0-13-474222-2

Cheers
Colin
-//-

-----------[000447][next][prev][last][first]----------------------------------------------------
Date:      19 Dec 1994 15:34:38 GMT
From:      keithng@hk.super.net (Mr Keith Ng)
To:        comp.protocols.tcp-ip
Subject:   Any articles on networking protocols?

I want to broaden my knowledge in interconnecting pc lan to host computer.
I have no previous knowledge in things like TCPIP, DECNET, TELNET, LAT, 
SLIP, etc.  Does anybody know any articles or good books I can get from 
the Internet?

-----------[000448][next][prev][last][first]----------------------------------------------------
Date:      19 Dec 94 15:34:50 GMT
From:      jcmorris@mwunix.mitre.org (Joe Morris)
To:        alt.winsock,comp.infosystems.www.users,comp.protocols.ppp,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc
Subject:   Re: Chameleon and Mosaic trouble

vql0608@is.nyu.edu (Vincent  Lai) writes:

>       When I attempt to run Mosaic, I get the following GPF error:
 
>Application error: Growstub caused a GPF in module POINTER.DLL at 0001:0F2E

Any chance you're using MS Mouse driver version 9?  There's a known
problem in the POINTER.DLL file that produces the symptom you describe;
a fix is supposedly available at ftp.microsoft.com but I don't know
the filename or directory.  (I don't use V9 so I haven't looked for it.)

Joe Morris / MITRE

-----------[000449][next][prev][last][first]----------------------------------------------------
Date:      20 Dec 1994 05:07:59 -0800
From:      cgi@crl.com (Paul Smith)
To:        comp.protocols.tcp-ip
Subject:   2->N Tasks write to UDP socket?

I have a need for a "concurrent" server that uses UDP.  

1) A client sends a UDP pkt to the servers port.

2) Server forks a child and goes back to recv() on it's port.

3) Child does data base stuff, sends 1-->N UDP pkts on the 
socket (fd) inherited from the parent useing send_to() which
includes the required routing info to reach the original client.

Child does not do recv().

Question:  If there are N child tasks writing to the same
fd, will all UDP pkts be kept atomic?

If its possible I'm interested in any thoughts on a true concurrent
server that uses UDP.  The requirement would be that it must be 
transparent to the client.  I.E. can't cause a re-bind() to
a new port for the child server task's new port #...

Thanks..

-----------[000450][next][prev][last][first]----------------------------------------------------
Date:      Mon, 19 Dec 1994 15:58:03 GMT
From:      werner@sarducci (John Werner x25054)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc
Subject:   Call Back SLIP/CSLIP/PPP ?


I am looking for a Callback implementation of SLIP/CSLIP/PPP to run on
a SUN host.  Basically, we are looking to add dial in access for PCs
running windows sockets to a couple of our Sun Workstations.  I have
seen the SLIP/etc. stuff for generic dial-in ports, but we are
required to have Callback systems for all of our computers with
modems.  It seems like it should be an easy thing to setup a SLIP
driver to initiate a connection, then drop the line, and call it back.

Before we go off and write it ourselves, has anyone already done this?

I have not really used SLIP, CSLIP or PPP, so I can't really be sure I
am not missing something (i.e. if one has to login into the machine
before starting up SLIP, then the Callback needs to be at an earlier
level).

Comments?

-- 
John Werner			| Disclaimer: The opinions expressed here
  INet werner.wbst311@xerox.com	|   are not necessarily those of my employer,
  XNS: werner:wbst311:xerox	|   or any of its affiliates.
	     Work: (716)422-5054 	Home: (716)436-3607
--
John Werner			| Disclaimer: The opinions expressed here
  INet werner.wbst311@xerox.com	|   are not necessarily those of my employer,
  XNS: werner:wbst311:xerox	|   or any of its affiliates.
	     Work: (716)422-5054 	Home: (716)436-3607

-----------[000451][next][prev][last][first]----------------------------------------------------
Date:      Mon, 19 Dec 1994 15:46:08 LOCAL
From:      ricard@axis.se (Ricard Wolf)
To:        comp.protocols.tcp-ip
Subject:   Re: Source Routing?

	In article <3cog2t$p8j@tools.near.net> barmar@nic.near.net (Barry Margolin) 
writes:>From: barmar@nic.near.net (Barry Margolin)
>Subject: Re: Source Routing?
>Date: 14 Dec 1994 23:18:37 -0500
 
>In article <3coe3q$7uf@mogan.cc.metu.edu.tr> ergul@ds5500.cc.boun.edu.tr (Salih Ergul) writes:
>>       I'm hearing about source routing option of IP packets, but I
>>don't know what it is and how to use it. Could anybody explain this?
 
>This option allows an IP packet to specify the routers that should be used,
...

Is IP source routining really used in practice? I've seen it in the IP RFC, 
and some implementations of 'ping' seem to be able to set up various source 
routing fields, but I've never seen it being used in the real world.

/Ricard
--
Ricard Wolf                   / | \  / | /-           email: ricard@axis.se
Axis Communications AB       /__|  \/  | \__          uucp:  axisab.se!ricard
S-223 70 LUND               /   |  /\  |    \         Tel:   +46 46 19 18 63
SWEDEN                     /    | /  \ | \__/         Fax:   +46 46 13 61 30

-----------[000452][next][prev][last][first]----------------------------------------------------
Date:      19 Dec 1994 19:07:31 GMT
From:      pnor@se.bel.alcatel.be (Philip Norris)
To:        comp.infosystems.www.users,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.infosystems.www.providers,comp.os.ms-windows.networking.tcp-ip,comp.os.ms-windows.networking.misc
Subject:   Re: WWW Server/Client without INTERNET??



= jtara@cts.com (Jon Tara) writes:
= > Gordon McAndrew wrote:
=
= > If you have a Novell Lan or equivalent just have a drive mapping that
= > points to a place on the server where the html files live, and you can
= > use the Open Local File option to get to files. You can also write
= > links using the file://  option for "local file" links.
=
= Yes, the latter approach will work fine as long as you don't need to run any 
= CGI scripts. If you do, then you will need a real server, and will need to run 
= TCP/IP over your LAN.
=
= In either case, you will need a null Winsock, or a real one with it's SLIP or 
= PPP interface disabled. I tried the null winsock that's floating around, and 
= it bombed with Netscape, so I suspect that the real Winsock, disabled approach 
= is the only way to go if you're not running TCP/IP.


Been following this thread for a while now. Maybe somebody can answer this.

Scenario: I want to rent space on a Web server. So I need to develop Web pages.
          So I need to test Web pages. Lets develop the pages locally, on my
          stand-alone Windows PC. Lets test 'em there too. Got no TCP/IP,
          got no network card either (got a modem though). I know what the
          ultimate URL of the pages will be. But for now, wanna develop pages
          locally.

Question: How do I develop pages locally, then transfer them to a server,
          *WITHOUT* having to change every godamned URL?

I think what I want is to spoof the server part of a URL. In other words, I
want to say "Hey, browser. When you see http://expensive.xxx.com/myfile.html,
replace it with http://localhost/myfile.html, then go get the file". As an
alternative, a local server which answers to the name of expensive.xxx.com
would be fine.

Is it possible? How close can I get? What do others do?

-- 
Regards

Philip Norris               "An electric twitch ... sometimes I swear
pnor@se.bel.alcatel.be       this body's got a mind of its own" - Chris Rea

-----------[000453][next][prev][last][first]----------------------------------------------------
Date:      19 Dec 1994 19:22:44 GMT
From:      ganguly@sw.stratus.com (Deb Dutta Ganguly)
To:        comp.protocols.tcp-ip
Subject:   t_look() internals

I wish to implement a function similar to t_look() for a API I am
developing. Can someone tell me the function(s) used by t_look() ?

Thanks
debdutta ganguly

-----------[000454][next][prev][last][first]----------------------------------------------------
Date:      19 Dec 1994 19:59:45 GMT
From:      cary@scripps.edu (Steve Cary)
To:        comp.protocols.tcp-ip
Subject:   IPng/IPv6 Documentation Pointer Request

Can somebody point me at the appropriate documents from 
the IETF which describe the new standard??

	THANKS!
	-Steve

(Light holiday reading, anyone?)





-----------[000455][next][prev][last][first]----------------------------------------------------
Date:      Mon, 19 Dec 1994 21:00:18 GMT
From:      patrick@cs.arizona.edu (Patrick T. Homer)
To:        comp.parallel,comp.lang.misc,comp.protocols.tcp-ip
Subject:   Re: Massively parallel programming for the Internet

In article <thinmanD0toH9.B94@netcom.com>, thinman@netcom.com (Technically Sweet) writes:
|> Are there any projects attempting to build a massively parallel
|> programming environment using spare cycles on the Internet?
|> Something where you put a really simple interpreter on the
|> end machines, mid-level cluster managers spaced periodically,
|> and end-user control stations?

One project along these lines is the Legion project, being headed by some
folks at the U. of Virginia.  They have a Mosaic WWW server that can
provide some additional details:

	http://uvacs.cs.virginia.edu/~mentat/legion/legion.html

Patrick


---------------------------------------------------------------------------
|     Patrick T. Homer     | Dept. of Computer Science |    cat lover     |
| Post-Doc. Research Assoc.| The University of Arizona | backpacker/hiker |
|  patrick@cs.arizona.edu  | Tucson, AZ 85721          |    raconteur     |
|      (602) 621-4252      | (602) 621-4246 (FAX)      |  sf enthusiast   |
---------------------------------------------------------------------------


-----------[000456][next][prev][last][first]----------------------------------------------------
Date:      19 Dec 1994 21:08:43 GMT
From:      dykster@ibm.net
To:        comp.protocols.tcp-ip
Subject:   Mosaic server for an AS/400

We are looking for software for an AS/400 that makes it a Mosaic server?

If one does exist, does it allow access to the Internet or only over a local
LAN?  What means would the AS/400 use to connect to the Internet other
than going through a Unix (or equivalent) gateway?

-----------[000457][next][prev][last][first]----------------------------------------------------
Date:      Mon, 19 Dec 1994 21:10:49 GMT
From:      jdonham@us.oracle.com (Jake Donham)
To:        comp.parallel,comp.lang.misc,comp.protocols.tcp-ip
Subject:   Re: Massively parallel programming for the Internet

"Lance" == Technically Sweet <thinman@netcom.com> states:

    Lance> Are there any projects attempting to build a massively
    Lance> parallel programming environment using spare cycles on the
    Lance> Internet?  Something where you put a really simple
    Lance> interpreter on the end machines, mid-level cluster managers
    Lance> spaced periodically, and end-user control stations?

    Lance> This was inspired by the project to crack large prime
    Lance> numbers by mailing large lists of numbers around to
    Lance> volunteers running the sieve program.  Such a thing could
    Lance> be entirely automated.  David Gelernter's "Mirror Worlds"
    Lance> is clear articulation of this vision.

Gelernter is working on this kind of thing at Yale. His system (or one
of his systems) is called Pirhana, and it grabs spare cycles off of
the Yale network for computation.

Pirhana programs are written in Linda. A description of parallel
programming in Linda can be found in Gelernter and Carriero's
_How_to_Write_Parallel_Programs_.

Jake


-----------[000458][next][prev][last][first]----------------------------------------------------
Date:      Mon, 19 Dec 1994 21:11:56 GMT
From:      callahan@xmission.com (Michael Callahan)
To:        comp.parallel,comp.lang.misc,comp.protocols.tcp-ip
Subject:   Re: Massively parallel programming for the Internet

Technically Sweet (thinman@netcom.com) wrote:
: Are there any projects attempting to build a massively parallel
: programming environment using spare cycles on the Internet?
: Something where you put a really simple interpreter on the
: end machines, mid-level cluster managers spaced periodically,
: and end-user control stations?

I have seen sevaral systems which may qualify at some point, but
I think the technology is still in it's infancy, and will take
some time to become widely available or used, if it ever is.

The contenders that I know of:
scheme-48:  There is work on turning this implementation of scheme
            into an internet scripting language.  A sample WWW server
            has been set up, see the scheme-48 home page for the
            exact location.
        http://www-swiss.ai.mit.edu/~jar/s48.html

safe-tcl:  I'm not sure of the details of this one.  It appears that
           the idea is to extend tcl in such a way that it can
           be used as an internet scripting language.

GUILE:  The new GNU extention language.  There was talk of making a
        safe GUILE, similar to safe-tcl.  So far, this is still
        vaporware, but I'm sure that at some point, someone will
        create a GUILE-capable WWW server.

Also, on Jeff Dalton's home page, I found a link to a project called
O-Plan, but I'm not sure if it's actually doing any distributed
computation besides what the server does.

     http://www.aiai.ed.ac.uk/~oplan/
     http://www.aiai.ed.ac.uk/~jeff/

: This was inspired by the project to crack large prime numbers by
: mailing large lists of numbers around to volunteers running the
: sieve program.  Such a thing could be entirely automated.
: David Gelernter's "Mirror Worlds" is clear articulation of this vision.

I'd be supprised if anything as 'organized' as Gelernter's vision
ever springs to life from the internet.  Here's a link I found to 
Yale's Linda home page:
      http://www.cs.yale.edu/HTML/YALE/CS/Linda/linda.html
There is also a trellis page somewhere at the same site.


   mike
http://www.xmission.com/~callahan


-----------[000459][next][prev][last][first]----------------------------------------------------
Date:      Mon, 19 Dec 1994 21:14:46 GMT
From:      loss@husky.bloomu.edu (Doug Loss)
To:        comp.parallel,comp.protocols.tcp-ip
Subject:   Re: Massively parallel programming for the Internet

In article <thinmanD0toH9.B94@netcom.com> thinman@netcom.com (Technically Sweet) writes:

>Are there any projects attempting to build a massively parallel
>programming environment using spare cycles on the Internet?
>Something where you put a really simple interpreter on the
>end machines, mid-level cluster managers spaced periodically,
>and end-user control stations?

This sounds reminiscent of Linda, a distributed parallel computing system I 
read about a while ago.  I'm sure there's material about it somewhere on the 
internet.  It seems to me that I remember a company named Torque that was 
selling a "compute server" designed for a Linda-type environment; the system 
was to sit on a network and provide computing power for client programs as 
needed.


Doug Loss                        Cogito Eggo Sum
Data Network Coordinator         I think.  I am a waffle.
Bloomsburg University                                                        
loss@husky.bloomu.edu                                                        
Voice (717) 389-4797                                                         


-----------[000460][next][prev][last][first]----------------------------------------------------
Date:      Mon, 19 Dec 1994 23:11:10 GMT
From:      paul@atlas.abccomp.oz.au (Paul Brooks)
To:        comp.protocols.tcp-ip
Subject:   Re: TCP RFC 793 to BERKLEY SOCKETS correlation  -  post.fil [1/1]

In article <787391712snz@cucumber.demon.co.uk> Andrew@cucumber.demon.co.uk writes:
|In article <3cieob$3lf@newhub.xylogics.com>
|           carlson@xylogics.com "James Carlson" writes:
|[snip]
|
|> The TCP_NODELAY flag turns off the Nagle algorithm.  As far as I'm able
|> to tell, it has nothing to do with the PUSH bit in any implementation.
|> 
|> At least in the BSD4.3 code, PUSH is set when the data being sent by TCP
|> fits entirely within the window and there's no more to send.  That is to
|> say, it does it automatically by default.
 
|>                                            (The comments in the code say
|> that this is done to work around bugs in implementations which need to
|> see either a full buffer or a PUSH before any data are delivered to the
|> application level.)
|> 
|
|I havn't looked at the BSD source, but I have to say this sounds wrong
|to me. An implementation is only *required* to forward data to the
|application when:

It is probably to get around receivers which don't forward data up to
the application until it gets a PUSH signal. The TCP v 2.2 on our
IBM AS400 is like this - unless you put a PUSH flag in, it WILL NOT
forward the data up to the telnet server, so option negotiation
never occurs, no matter how much data you try to send to force it.

-- 
Paul Brooks              |paul@abccomp.oz.au       |Emerging Standard:
TurboSoft Pty Ltd        |pwb@newt.phys.unsw.edu.au|  one that has not yet
579 Harris St., Ultimo   |                         |  been superseded.
Sydney Australia 2007    |ph: +61 2 281 3155       |  

-----------[000461][next][prev][last][first]----------------------------------------------------
Date:      Mon, 19 Dec 1994 23:17:51 GMT
From:      seth@news.dorsai.org (Seth Bromberger)
To:        comp.protocols.tcp-ip
Subject:   Re: Urgent: SMTP Server for DOS

Also, doesn't NOS (POP3NOS) include an SMTP server?

Seth Bromberger

Steve Holden (Steve@scotia.co.uk) wrote:
: In article <3ctpe4$l7t@eis.calstate.edu>
:            shunt@eis.calstate.edu "Steve J. Hunt" writes:
 
: > juanm@ac.upc.es (Juan Manuel Bernabeu) writes:
: > 
: > > I'm looking desperatly for a DOS-TSR based SMTP server program.
: > I don't think there are any DOS or Windows based sendmail type 
: > programs.
 
: The KA9Q variant I obtained from ftp.demon.co.uk as a part of their
: DIS216C package includes DOS-compatible SMTP send and receive code.
: I believe this is shareware, and I don't think it varies too radically
: from the standard KA9Q code. So you should be able to get source.
 
: If you don't mind some assembly being required you could probably
: write a TSR using the KA9Q source as a base.

-----------[000462][next][prev][last][first]----------------------------------------------------
Date:      20 Dec 1994 13:48:43 -0800
From:      phil@lykos.netpart.com (Phil Trubey)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc
Subject:   IP layer encryption available?

Does anyone know of any PD or commercial products that implement any form
of encryption at the IP layer?  Specifically, I am looking for
a stack that can run in DOS or Windows that can encrypt IP packets
as they go onto the net - obviously this stack would have to work
hand in hand with another stack, or device somewhere which
stripped off this encryption capability.

The place I envision this being used is in a company that has many
scattered employees that need to access TCP/IP based corporate
data - and they would like to use the Internet to access this data.
To keep the data away from prying eyes, some form of encryption would be
needed (to defeat packet sniffer attacks).

Thanks for any and all info.  I'll post a summary if I get responses.

-- 
Phil Trubey                 | 
NetPartners                 | Providing Internet products and services. 
E-mail: phil@netpart.com    |   Home Page: http://www.netpart.com/
Phone:  714-759-1641        |

-----------[000463][next][prev][last][first]----------------------------------------------------
Date:      20 Dec 1994 01:00:18 GMT
From:      jamesc@osaiphp1.osa.csi.itg.telecom.com.au (James Crawford)
To:        comp.protocols.tcp-ip
Subject:   Re: Summary: How to detect dead socket under HP-UX 9.03 (and more generally)?

I have a slightly different question.  How can I detect that the
socket connection has been closed (or better still shutdown) from
the other end when writing to the socket.

It seems that I have no problem detecting this when doing a read,
but when I write to the socket, the write will work fine, even if
the other end has shutdown or closed their end of the conversation.

I also am running HPUX 9.03, but would prefer a generic solution
if there is one.

Thanks for any advice.

James Crawford.


-----------[000464][next][prev][last][first]----------------------------------------------------
Date:      20 Dec 1994 02:34:27 GMT
From:      cengiz@cybernetics.net (Cengiz Akinli)
To:        comp.infosystems.www.users,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.infosystems.www.providers,comp.os.ms-windows.networking.tcp-
Subject:   Re: WWW Server/Client without INTERNET??

In article <3cot7n$md0@desiree.teleport.com>, AIR - System Design <air@teleport.com> says:
>
>With Nullsock on each machine running NetScape and nothing on
>the server at all, All machines can open the .htm files
>in f:\public on the server.
>
>I do not use forms or CGIs. I just share hypermedia information.
>
>
>-> >The original poster wanted to supply WWW services to the other PCs on the 
>-> >same LAN.
>-> >
>->>Therefore he needs a complete setup with a WWW server and WWW clients.
>-

I think he means he was going to put the hypermedia on a drive available to
the LAN, in which case each browser would load it as a local file.

That can be done, and has been done.  If you already have a networking package
installed that works, there's not always a need to install another one (TCP/IP).

>-> I haven't tried this, but wouldn't it be possible to run Mosaic or Netscape with a 
>-> nullsock and access the HTML documents off the LAN server as local files?
>-> 

Yes, it sure would.  We've done that and it works great.
>-> Olof
>
>

A major issue here, which I don't THINK has been addressed, is the topology of
your LAN.  Is this a peer-to-peer, or client-server LAN?

If you only need to store the hypermedia information on the server then
doing it that way will work fine.  If, on the other hand, your file server
is under heavy load, or you need to be able to make information on drives
not mounted on the network available, then implementing a TCP/IP protocol
stack on the workstations is a tremendously efficient, low-overhead solution.

A LAN networking package and a TCP/IP stack can coexist quite happily on the
same ethernet.  I know because ours do.  Doing this gives us the ability to 
use a workstation on the LAN as an HTTP server.  Also, the data on the HTTP
server is much more secure because it is inaccessible by the LAN.  The only
way anyone can modify the hypermedia is to either sit down in front of that
workstation, which sits 10' away from my desk, or the ftp into it - but I 
don't run an FTP daemon on it, so problem solved!

Doing this takes a good load off the server.



Incidentally, I've tried using the nullsock at NCSA but it never worked.  I
just loaded and ran Trumpet Winsock without configuring it.  As long as
you only refer to local files, you'll never make a network system call and
it'll stay quiet.

-----------[000465][next][prev][last][first]----------------------------------------------------
Date:      Tue, 20 Dec 1994 02:48:36 GMT
From:      sidd@pacific.mps.ohio-state.edu (Siddhartha Mukherjee)
To:        comp.lang.misc,comp.protocols.tcp-ip
Subject:   Re: Massively parallel programming for the Internet

In article <thinmanD0toJ5.BGH@netcom.com> thinman@netcom.com (Technically Sweet) writes:
>Are there any projects attempting to build a massively parallel
>programming environment using spare cycles on the Internet?
>Something where you put a really simple interpreter on the
>end machines, mid-level cluster managers spaced periodically,
>and end-user control stations?
>

amoeba (by Andrew Taanenbaum ?) i think is an implementation of
such a system

-----------[000466][next][prev][last][first]----------------------------------------------------
Date:      20 Dec 1994 02:51:20 GMT
From:      Michael Noe <mnoe@interport.net>
To:        comp.protocols.tcp-ip
Subject:   Re: WinNT & Windows TCP/IP connection

> >Windows NT 3.5 supports TCP/IP via both PPP and SLIP/CSLIP with its
> >Remote Access Services (RAS) Server.
> >
> >This will work only with another Windows NT 3.5 box or a Windows 95
> >box.
> 
> 	Typical proprietary garbage from Micro$oft.  *Any* PPP or CSLIP
> software on a particular host __should__ be able to connect to *any* PPP/CSLIP
> software on a remote system, given the right handshaking parameters.
                                                               
                                                            
                                                               
The Windows NT Internet FAQ at http://www.luc.edu/~tbaltru/faq/ explains
how to setup RAS to be a generic PPP server.  RAS doesn't support
SLIP.  I'm doing some preliminary testing with it, and so far it
works as advertised.  I'm connecting a DOS/Windows 3.1 PC (without 
RAS) using Trumpet's winsock & built in dialer.

Michael Noe


-----------[000467][next][prev][last][first]----------------------------------------------------
Date:      Tue, 20 Dec 94 05:11:34 GMT
From:      peterjt@hookup.net (Peter Thomas)
To:        comp.protocols.tcp-ip
Subject:   AS/400 TCP/IP Limits

I've been doing some reading of the AS/400 TCP/IP manuals and have come across a limit of 160 data
bufffers for usa by the TCP/IP support, along with a note that TWO to THREE buffers are required per
TELNET session.

We're looking at changing all of our access to our main AS/400 from being X25 based (through Eicon
Gateways, and ACCESS/5250) to being TCP/IP based, and this limit makes me wonder whether this
is possible...

Is this limit still in effect on the current release (V3R5), and if so, would anyone know if IBM has any plans
to increase this?

Peter Thomas

-----------[000468][next][prev][last][first]----------------------------------------------------
Date:      20 Dec 1994 15:40:26 -0500
From:      sunil@prospero.dev.cdx.mot.com (Sunil Menon)
To:        comp.protocols.tcp-ip
Subject:   Assymetric Bandwidth


If the Bandwidth in the two directions is Assymetric i.e is unequal, does it cause any problems in TCP/IP? Especially at the TCP layer. Does anybody have any info, experience, suggestions, pointers etc?

Thanks in advance.

Sunil.

-----------[000469][next][prev][last][first]----------------------------------------------------
Date:      Tue, 20 Dec 1994 07:31:03 GMT
From:      liberty@netcom.com (Liberty Info Network)
To:        comp.protocols.tcp-ip
Subject:   can nslookup but not ping ?

I've got our Primary DNS setup on our Unixware server.  It seems to work
fine on that system.  But, in testing from one of our Sun systems (running
Solaris) I created the resolv.conf but when I try and ping the nameserver
it says host unknown but if I do an nslookup it returns the correct address.

Am I missing something here.  I've not worked with Solaris before.  Perhaps
there is something else I need to do.

Any help would be greatly appreciated!

Thanks.

Michael Jay


-----------[000470][next][prev][last][first]----------------------------------------------------
Date:      20 Dec 1994 16:54:56 -0500
From:      barmar@nic.near.net (Barry Margolin)
To:        comp.infosystems.www.users,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.infosystems.www.providers,comp.os.ms-windows.networking.tcp-ip,comp.os.ms-windows.networking.misc
Subject:   Re: WWW Server/Client without INTERNET??

In article <3d4llj$4tu@btmplq.god.bel.alcatel.be> pnor@se.bel.alcatel.be writes:
>I think what I want is to spoof the server part of a URL. In other words, I
>want to say "Hey, browser. When you see http://expensive.xxx.com/myfile.html,
>replace it with http://localhost/myfile.html, then go get the file". As an
>alternative, a local server which answers to the name of expensive.xxx.com
>would be fine.

Use relative URLs to link between documents that live on the same server.
So the URLs should look like "http:/myfile.html".  The client will then
merge in the hostname field that was used to access the referencing
document to form the actual URL that is used.  So if you got to the
refernecing document via "http://expensive.xxx.com/home.html", the link
will go to "http://expensive.xxx.com/myfile.html", but if you got there via
"http://localhost/home.html" it will go to "http://localhost/myfile.html".
-- 

Barry Margolin
BBN Internet Services Corp.
barmar@near.net

-----------[000471][next][prev][last][first]----------------------------------------------------
Date:      20 Dec 1994 10:46:18 GMT
From:      eng10412@leonis.nus.sg (LEW CHAI SECK)
To:        comp.protocols.tcp-ip
Subject:   help in writing multiple client server

I need to write a simple server programme that can be used to serve 
mulitple client but not sure about how to go about doing it....

Is there any programme available that I can use to follow or is that any 
good books on this subject...

Any help will be greatly appreciated....

--
==============================================================
Lew Chai Seck
National Unversity of Singapore
Email : eng10412@leonis.nus.sg

-----------[000472][next][prev][last][first]----------------------------------------------------
Date:      20 Dec 1994 10:44:04 +0100
From:      teisa@beatles.cselt.stet.it (Claudio Teisa)
To:        comp.protocols.tcp-ip
Subject:   Internet routing protocols

Hi,
 looking at the RFCs and other Internet documentation, several protocols related
to the routing, seems to exist:

local routing:	ARP,RARP,INARP,NARP,...
IGPs:		RIP,RIP-2,OSPF,DUAL IS-IS,...
EGPs:		EGP,EGP-2,BGP,BGP-2,BGP-3,BGP-4,...
others:		DHCP,...

 But: what is their actual deployment in Internet?

 Which are obsolete, actual or future? 
 Which are already implemented in commercial products?

 I have seen some interesting messages on DHCP and RIP-2 on this conference
and I wonder if someone could be so kind to complete the scenario with some
information (just few words) on the other protocols (even if I have forgotten 
them in the list above).

 Thanks,
                                          C. Teisa
 
 teisa@beatles.cselt.stet.it


-----------[000473][next][prev][last][first]----------------------------------------------------
Date:      20 Dec 1994 13:00:08 GMT
From:      rdmastro@jpmorgan.com (Richard DelMastro)
To:        comp.protocols.tcp-ip
Subject:   Re: Help! binding socket to address with bind

In article q0d@news.eecs.uic.edu, anewman@ernie.eecs.uic.edu (Aaron Newman) writes:
> ...  But after I
>exit the app and restart, Bind dies with 'address in use' error code.
> ...

Aaron -

Do a setsockopt (SO_REUSEADDR)

Rich


-----------[000474][next][prev][last][first]----------------------------------------------------
Date:      Tue, 20 DEC 94 21:40:26 -0500
From:      ngai@delphi.com
To:        comp.protocols.tcp-ip
Subject:   Network design and requirements analysis


Due to the rapid advances in telecom and network technologies, today's
MIS, telecom and network managers face increasing challenge in selecting
the right kind of network technology and equipment for their specific
application and when they are planning their enterprise network.

If you are working on local area network or wide area network technologies
such as T1 & T3, ISDN, ATM, SONET, SMDS, FDDI, X.25 and Frame Relay, PYN
Associates can assist you in the following areas: telecom network and system
requirements analysis, telecom and data network design, network planning and
and management, product system level design, network equipment evaluation/
acquisition.

Hence if you are a MIS, network ot telecom manager planning an enterprise
network, upgrading existing telecom network to accomodate increasing voice/
data traffic, acquiring new telecom/data communication equipment and service,
or a product manager developing a new telecom system, why not contact PYN
Associates for more information.

For free information via Email, send request to
NGAI@delphi.com
or contact
PYN Associates
Telecom Consultant
109 Lindsey Court, Franklin Park, NJ 08823.
Tel: 908-940-1363  Fax: 908-940-1364


-----------[000475][next][prev][last][first]----------------------------------------------------
Date:      Tue, 20 Dec 1994 14:03:29 GMT
From:      sar@plc.com (Steve Rago)
To:        comp.protocols.tcp-ip
Subject:   Re: t_look() internals

In article <3d4mi4$888@transfer.stratus.com> ganguly@sw.stratus.com (Deb Dutta Ganguly) writes:
>I wish to implement a function similar to t_look() for a API I am
>developing. Can someone tell me the function(s) used by t_look() ?
>
>Thanks
>debdutta ganguly

Take a peek at I_PEEK in streamio(7).

Steve Rago
sar@plc.com

-----------[000476][next][prev][last][first]----------------------------------------------------
Date:      20 Dec 1994 14:05:55 GMT
From:      etxmesa@eos.ericsson.se (Michael Salmon)
To:        comp.protocols.tcp-ip
Subject:   Re: Internet routing protocols

In article <3d6914$9s2@beatles.cselt.stet.it>, teisa@beatles.cselt.stet.it (Claudio Teisa) writes:
|> Hi,
|>  looking at the RFCs and other Internet documentation, several protocols related
|> to the routing, seems to exist:
|> 
|> local routing:	ARP,RARP,INARP,NARP,...
|> IGPs:		RIP,RIP-2,OSPF,DUAL IS-IS,...
|> EGPs:		EGP,EGP-2,BGP,BGP-2,BGP-3,BGP-4,...
|> others:		DHCP,...
|> 
|>  But: what is their actual deployment in Internet?
|> 
|>  Which are obsolete, actual or future? 
|>  Which are already implemented in commercial products?

I think that you need to take a look at STD 1 aka RFC 1720.

-- 

Michael Salmon

#include	<standard.disclaimer>
#include	<witty.saying>
#include	<fancy.pseudo.graphics>

Ericsson Telecom AB
Stockholm

-----------[000477][next][prev][last][first]----------------------------------------------------
Date:      Tue, 20 Dec 1994 14:48:28 GMT
From:      rsgawera@ss11.wg.icl.co.uk (Raj Gawera)
To:        comp.protocols.tcp-ip,comp.protocols.vxworks
Subject:   rdate

Hi,
	Where can I get hold of the source for the rdate command ?
	
		Raj

Raj Gawera
ICL CSD
email rsgawera@wg.icl.co.uk



-----------[000478][next][prev][last][first]----------------------------------------------------
Date:      20 Dec 1994 14:49:12 GMT
From:      ehssand@aom.ericsson.se (Ulrik Sandstr|m)
To:        comp.protocols.tcp-ip,comp.unix.solaris
Subject:   dup2() before connect() won't work on SunOS5.3

Hi,

I have problems connecting a socket after calling dup2(). Here is 
the scenario:

1. An open connection is closed by the server side.

2. The client creates a new socket.

3. The new socket is duplicated onto the old socket. This is done
   to preserve the old socket number.

4. The new socket is closed by the client.

5. The old socket is used to make a new connection to the server
   again. Same port is used.

This works fine on SunOS4.1.3 and HPUX09.03, but on SunOS5.3 the 
connect() call fails and returns EAFNOSUPPORT.

Code example:

	...
	new_sock = socket(AF_INET, SOCK_STREAM, 0);
	dup2(new_sock, old_sock);
	close(new_sock);
	connect(old_sock, (struct sockaddr*)address,
		sizeof(struct sockaddr)); // Returns EAFNOSUPPORT
	...

I've also noticed that an attempt to send() on an disconnected
socket will result in EAFNOSUPPORT when trying to connect() it
later on.

I'm very grateful for any explanations, solutions or suggestions.

/Ulrik

---
--------------------------------------------------------------------
Ulrik Sandstrom       Ericsson Hewlett Packard Telecommunications AB
email ehssand@aom.ericsson.se   phone +46(0)31-672488   ECN 865 2488


-----------[000479][next][prev][last][first]----------------------------------------------------
Date:      20 Dec 1994 23:49:55 -0500
From:      aray@pipeline.com (Arjun Ray)
To:        comp.protocols.tcp-ip,comp.unix.solaris
Subject:   Re: dup2() before connect() won't work on SunOS5.3

In article <3d6qt8$hg9@erinews.ericsson.se>,
Ulrik Sandstr|m <ehssand@aom.ericsson.se> wrote:
>Hi,
>
>I have problems connecting a socket after calling dup2(). Here is 
>the scenario:
>
>1. An open connection is closed by the server side.
>
>2. The client creates a new socket.
>

I assume the client is aware that the server has closed the previous
connection -- i.e. that the old socket desriptor is no longer usable for I/O.

>3. The new socket is duplicated onto the old socket. This is done
>   to preserve the old socket number.

Your code accomplishes this by dup2(). But dup2() doesn't need the old
socket/descriptor to be open.
Your problem may be due to the fact that the descriptor is still (technically)
open, but in fact unusable. So, instead of relying on the kernel (actually,
the STREAMS system, in this case the sockmod module pushed onto the /dev/tcp
driver) to do everything right, it could be a good idea to _force_ it to do
everything correctly. (See below.)

>
>4. The new socket is closed by the client.
>
>5. The old socket is used to make a new connection to the server
>   again. Same port is used.
>
>This works fine on SunOS4.1.3 and HPUX09.03, but on SunOS5.3 the 
>connect() call fails and returns EAFNOSUPPORT.
>
>Code example:
>
>	...
>	new_sock = socket(AF_INET, SOCK_STREAM, 0);
>	dup2(new_sock, old_sock);
>	close(new_sock);
>	connect(old_sock, (struct sockaddr*)address,
>		sizeof(struct sockaddr)); // Returns EAFNOSUPPORT
>	...
>
I suggest a paranoid approach using the following sequence of calls:

	shutdown( old_sock, 2 ) ; /* --[ forces buffer flushes ]--*/
	close( old_sock ) ; /* --[ make sure the old STREAM is dismantled ]-- */
	new_sock = socket( AF_INET, SOCK_STREAM, 0 ) ;
	connect( old_sock, (struct sockaddr*) address, sizeof(struct sockaddr) ) ;
	dup2( new_sock, old_sock ) ;
	close( new_sock ) ;


>I've also noticed that an attempt to send() on an disconnected
>socket will result in EAFNOSUPPORT when trying to connect() it
>later on.
>
>I'm very grateful for any explanations, solutions or suggestions.
>
>/Ulrik
>
>---
>--------------------------------------------------------------------
>Ulrik Sandstrom       Ericsson Hewlett Packard Telecommunications AB
>email ehssand@aom.ericsson.se   phone +46(0)31-672488   ECN 865 2488
>

Hope this helps.

Arjun Ray
The Pipeline Network Inc.


-----------[000480][next][prev][last][first]----------------------------------------------------
Date:      Tue, 20 Dec 1994 18:15:07 GMT
From:      mrisch@computerlaw.com (Michael Risch)
To:        comp.protocols.tcp-ip
Subject:   [Q] Bindist download errors

Crosspost of article

In article <1994Dec20.001353.14802@nntpxfer.psi.com>, 
mrisch@computerlaw.com says...
>
>I am trying to FTP the bindist files for my FreeBSD2.0 installation,
>and my computer keeps hanging.  
>
>Thus, i downloaded the files to a non-BSD machine on my net, and tried
>to FTP them from there.  Alas, more hangs.  A trace on the FTP server 
>shows TCP CHKSUM errors.  
>
>Any ideas?
>
>Thanks in advance
>
>Michael Risch.


-----------[000481][next][prev][last][first]----------------------------------------------------
Date:      20 Dec 1994 18:17:28 GMT
From:      netmar@cybernetics.net (Cengiz Akinli)
To:        comp.infosystems.www.users,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.infosystems.www.providers,comp.os.ms-windows.networking.tcp-
Subject:   Re: WWW Server/Client without INTERNET??

>Been following this thread for a while now. Maybe somebody can answer this.
>
>Scenario: I want to rent space on a Web server. So I need to develop Web pages.
>          So I need to test Web pages. Lets develop the pages locally, on my
>          stand-alone Windows PC. Lets test 'em there too. Got no TCP/IP,
>          got no network card either (got a modem though). I know what the
>          ultimate URL of the pages will be. But for now, wanna develop pages
>          locally.
>
>Question: How do I develop pages locally, then transfer them to a server,
>          *WITHOUT* having to change every godamned URL?
>
>I think what I want is to spoof the server part of a URL. In other words, I
>want to say "Hey, browser. When you see http://expensive.xxx.com/myfile.html,
>replace it with http://localhost/myfile.html, then go get the file". As an
>alternative, a local server which answers to the name of expensive.xxx.com
>would be fine.
>
>Is it possible? How close can I get? What do others do?
>
>-- 
>Regards
>
>Philip Norris               "An electric twitch ... sometimes I swear
>pnor@se.bel.alcatel.be       this body's got a mind of its own" - Chris Rea

There's nothing really tricky about it.  Choose from among the other 
recommendation that have been given so far in the thread (nullsock, Trumpet
disable, etc.) to get your browser running locally.  Personally, I had trouble
with nullsock but didn't stick with it very long since I had Trumpet lying 
around.

Insofar as writing your HTML is concerned, you should ALWAYS, ALWAYS, ALWAYS,
ALWAYS, ALWAYS (maybe a couple more) use relative paths.  For instance,
you know how in DOS if your in like the C:\MYFILES\THEIRS\BOB\HOMEWORK\SCIENCE
directory if you want to go to the BIOLOGY directory in that directory you
don't type 'CD C:\MYFILES\THEIRS\BOB\HOMEWORK\SCIENCE\BIOLOGY' you just type
'CD BIOLOGY' since you're already in the directory tree.

It's the same thing in constructing web pages.  If you have page A and page B
in the same directory on the same server you don't write the link to B
in A as 'http://sameserver.blah/directory/pageb' you just use 'pageb.'

The browser remember what machine and what directory the current document is
in so you can use relative paths.  You always should, otherwise moving data
from one directory to another can be a nightmare.


Good luck!



Seasons Greetings,
Cengiz Akinli
 __                                                                      __ 
<  >--------------------------------------------------------------------<  >
 || Cengiz Akinli                |  Every great achievement was once     ||
 || Netmar, Inc.                 |  considered impossible                ||
 || Voice - (919) 309-2459	 |					 ||
 || Voice Mail - (919) 254-3561  |					 ||
 || cengiz@cybernetics.net       |                                       ||
 || netmar@cyberentics.net	 |                          -Unknown     ||
<__>--------------------------------------------------------------------<__>

-----------[000482][next][prev][last][first]----------------------------------------------------
Date:      20 Dec 1994 21:34:43 GMT
From:      paulc@vnet.ibm.com (Paul Chmielewski)
To:        comp.protocols.tcp-ip
Subject:   Re: AS/400 TCP/IP Limits

In article <3d5p2u$2fd@relay.tor.hookup.net>, peterjt@hookup.net (Peter Thomas) writes:
|>Is this limit still in effect on the current release (V3R5), and if so, would anyone know if IBM has any plans
|> to increase this?
|> 
|> Peter Thomas


Peter,

I assume you mean V3R0.5 since there is no V3R5.  The limit *is* still in effect on V3R0.5.

The latest release is V3R1.0 which has removed the limitation.  In fact, AS/400 TCP/IP was almost completely rewritten for V3R1.0 with lots of other goodies added.

-Paul Chmielewski

-----------[000483][next][prev][last][first]----------------------------------------------------
Date:      20 Dec 1994 21:48:03 GMT
From:      freeman@mr.net (Alex Li)
To:        comp.protocols.tcp-ip
Subject:   HELP!  PROTMAN.DOS & NETBIND.EXE for DEC EtherWork3

I'm desparately looking for PROTMAN.DOS, PROTMAN.EXE, NETBIND.EXE
and UNBIND.EXE for the DEC EtherWorks3 card.  I've emailed DEC twice 
and haven't hear from them yet.

Thanks for any help!

Alex Li

-----------[000484][next][prev][last][first]----------------------------------------------------
Date:      20 Dec 1994 21:51:01 GMT
From:      freeman@mr.net (Alex Li)
To:        comp.protocols.tcp-ip
Subject:   Promiscuous mode & DEC EtherWorks 3 card?

Anyone knows if DEC EtherWorks 3 card supports PROMISCUOUS mode?
It's not documented in any of the docs. that I can find.

Thanks for your help.

Alex Li

-----------[000485][next][prev][last][first]----------------------------------------------------
Date:      20 Dec 1994 22:06:52 GMT
From:      paulc@vnet.ibm.com (Paul Chmielewski)
To:        comp.protocols.tcp-ip
Subject:   Re: Mosaic server for an AS/400


We currently have an AS/400 connected to the Internet that is running
a WWW server (ie. a server that speaks http which is what I assume you
mean by a "Mosaic" server).  The software has not been released yet,
but we're working on it.  The same AS/400 is also running the gopher
server.

I would expect an announcement to appear on that server once the code
is released.  The server is at http://as400.rochester.ibm.com/.

Your best option (right now) for connecting the AS/400 is to use a router or gateway.  Routers can be bought that provide firewall capabilities, if that is of any interest.

If you would like more info on this, you might email to me directly 
(paulc@vnet.ibm.com) since I don't regularly monitor this bboard.

-Paul Chmielewski

-----------[000486][next][prev][last][first]----------------------------------------------------
Date:      20 Dec 1994 23:34:51 GMT
From:      besancon@excalibur.ens.fr (BESANCON Thierry)
To:        comp.protocols.tcp-ip
Subject:   Re: IPng/IPv6 Documentation Pointer Request


Give it a try at :

        http://web.cnam.fr/Network/IPng/

It is written in French but lots of keywords/references are in English.
You might find what you want.

Thanks for the CNAM for providing this useful information.

        Thierry
--

		Thierry

-----------[000487][next][prev][last][first]----------------------------------------------------
Date:      20 Dec 1994 23:56:47 GMT
From:      balenson@tis.com (David M. Balenson)
To:        alt.security,alt.security.pgp,alt.security.ripem,comp.protocols.tcp-ip,comp.security.misc,sci.crypt
Subject:   IEEE PASC Study Group on Encryption APIs

IEEE PASC Study Group on Encryption APIs

The IEEE PASC (Portable Applications Standards Committee) group will be
hosting a study group on API's for encryption and other cryptographic
services as an extension to the POSIX standards.  The second meeting on
this will be held at the Bonaventure Hotel, Ft. Lauderdale, FL, Jan
16-19, 1995 in conjunction with the quarterly PASC meetings.  There is
a meeting fee of $100/day or $350 for the week that includes lunches.
The contact for this work is David Balenson who can be reached at
+1.301.854.6889 or balenson@tis.com.  A general meeting announcement is
available on request from Todd Campbell at NAPS International, who can
be reached at +1.612.888.0074 or tc@bungia.mn.org

The study group will investigate the feasibility of developing IEEE and
ISO standards for application program interfaces to encryption and
other cryptographic mechanisms, including data encryption,
authentication, data integrity, digital signatures, key management, and
timestamping.  The PASC study group would like to invite broad
participation from interested parties consisting of private
individuals, industry, government, users and producers.

------------------------------------------------------------------------
David M. Balenson
Trusted Information Systems, 3060 Washington Rd., Glenwood, MD 21738 USA
balenson@tis.com; tel 301.854.6889; fax 301.854.5363

-- 
David M Balenson

-----------[000488][next][prev][last][first]----------------------------------------------------
Date:      Tue, 20 Dec 1994 11:18:29 +0800
From:      peter.lewis@info.curtin.edu.au (Peter N Lewis)
To:        comp.protocols.tcp-ip
Subject:   Re: ftp and the SYST command (HELP)

In article <19941218.144254.08@comptech.demon.co.uk>,
adam@comptech.demon.co.uk wrote:

>In article <3cuf4i$vtg@styx.uwa.edu.au>,
>                Paul Wyatt Wagland <penguin@ucc.gu.uwa.edu.au> wrote:
 
>>Automated clients do a SYST, get OS/2, and then expect the LIST format to
>>be in OS/2 format, (this is a real drag BTW), and usually refuse to show
>>the output.

Paul, use OS/2 and tell the client writers to get a clue.  The SYST
command is not a specification for the output of the LIST command, it's
sto specify the base system.  It's unreliable at indicating any useful
feature as far as I've ever seen.  If you return UNIX, be prepared for ftp
clients to send UNIX ascii files (linefeed terminated) in binary mode
(another risky behaviour).

>In the case of FTP, it would perhaps help if a standard was specified for
>specific listing formats, and while typical unix output contains a lot of
>garbage that most people arnt interested in, it is the most common, and
>IHMO, should be adopted by all. Either that, or a new list command is needed
>that indicates type (dir/file/link), size, date and name in a strict format.

Agreed.  I'll post (again) the current working draft of both the FTP LIST
output format (which Paul complies with), and the new listing command
specification, which have both been debated endlessly on this group.
   Peter.
-- 
Peter N Lewis <peter.lewis@info.curtin.edu.au> - Macintosh TCP fingerpainter
It is Organisation, not Organization.  It's not our fault you folks can't spell.

-----------[000489][next][prev][last][first]----------------------------------------------------
Date:      Tue, 20 Dec 1994 11:19:32 +0800
From:      peter.lewis@info.curtin.edu.au (Peter N Lewis)
To:        comp.protocols.tcp-ip
Subject:   Re: ftp and the SYST command (HELP) - FTP LIST Specification

FTP LIST Specification - Peter N Lewis, Feb 1994.

Target Audience: 
FTP server authors and FTP client authors who wish to parse the LIST
command to display information in a local (non-english or graphical)
manner.

Premise:
There is no way to change existing servers in a reasonable amount of time.

For a long time the FTP LIST output has been unspecified and yet there are
more and more programs each year that rely on this format in one way or
another.  To try to combat this problem I am proposing a specification of
a LIST format that is easily decoded and very close to the current common
format.

Please note that I am offering up this specification in the hope of
increasing compatibility amongst servers and clients but no one is under
any obligation to pay any attention to this document.

It is necessary to distinguish between the output format and the method of
decoding the format since we wish to allow all decoders to interpret as
many of the current versions of the LIST format, while trying to reduce
the variation in the output.

The output should be something like this:

-rw-------  1 peter         848 Dec 14 11:22 00README.txt
or
- whatever you feel like 848 Dec 14 11:22 00README.txt

This specification only defines the first character, and the information
(size/date/name) at the end of the line.  The infoirmation between is at
the discretion of the server author, although you need to be careful to
avoid displaying something that will match the date format.

The first character should be one of [-dl].  "-" means a file, "d" means a
directory and "l" means a link.  In many cases it may be preferable to
resolve the link and display it simply as a file or directory, since the
client otherwise has no way to tell which it is except by trying to change
directory into it (or by using the -p switch, see later).  Currently the
only way I've been able to find to determine if a link is a file or
directory is to check if it has an extension (ie terminates in a dot
followed by 1 to 3 characters) (this is obviously less than an ideal
solution)

The rest of the line should be found by matching
<space><3-letter-month-code><space><any><digit><space><digit|space><digit>
(regexp: / ([A-Za-z][A-Za-z][A-Za-z]) .[0-9] [0-9 ][0-9]/ where $1 is a
valid english month) and using the earliest such match (since that string
may be part of the file name.  The month name match should be case
insensitive, and should match one of the twelve English three-letter month
abbreviations: Jan, Feb, Mar, Apr, May, Jun, Jul, Aug, Sep, Oct, Nov, Dec.
Servers should always use the English version.  Clients should of course
display the date in the local language and format where possible.  Some
servers may currently use French month names, so you might like to also
accept them (though it would be better to have the servers change to
English).

Once you have found that, you can determine the file size by looking at
the number immediately before that match location.  This number may be
missing or invalid for directories, links, or other "file types" (and thus
should be ignored except for files).

The date format is either:

  MMM DD hh:mm
OR
  MMM DD  YYYY
OR
  MMM DD YYYY 

In the first case, the year is taken to be the most recent past occurence
of the date.  Servers should choose to switch over well before this
becomes close to allow for variations in the local time (not to mention
user confusion).

And the file name follows the date and continues to the end of the line.
If the file name is a link, it may include a pointer to the original, in
which case it is in the form "name -> link".  This is really very bad,
since that is also a perfectly valid file name under unix and other
systems (especially the Macintosh).  Servers should not do this where
possible (resolve the link, correct the filesize and display as a file or
directory is in general a better plan).

The information between the first character and the size (for files) or
date (for other types) should be ignored.  Servers may put any system
dependent information here (although as noted about, you should avoid
displaying anything that looks like a date).

When decoding, it is important to note that many implementations include a
line at the start like "total <number>".  Clients should ignore any lines
that don't match the described format.  If no matches are found, and more
than a few lines are read, then the server is probably non-conformant, so
you might consider using the NLST command.  Also, listings may include the
special files "." and "..".  I have no idea what you do with them or how
you tell if they are special.  Servers should not display them where
possible.

Servers should try to support the important unix "ls" switches:

-p - add / on the end of directories, especially important for links.
-F - add /, *, @, = to the end of various files.
-l - in the long format described above (ie LIST should ignore it)
-R - recursive listing (some mirroring software relies on this)

Site maintainers should consider the problem of determining the type of a
link and try to maintain the original name or at least the original
extensions (this will also help users as well as client software).

Caveates:

A better sollution would be to implement a structured list command, but
this is not practical (even if this were done, it would be years before a
reasonable percentage of servers were converted, and thus this document
remains necessary).

Although it looks unix-centric, and english-centric, the intention is the
exact opposite.  By standardizing the list display, we allow clients to
display dates in the local language, and listings in a local and/or
graphical format.

BNF:

<list> ::= <line> <list>
         | <comment> <list>
         | <null>

<line> ::= <type-char> <stuff> <size> SPC <date> SPC <longname> <crlf>

<type-char> ::= "-" | "d" | "l"

<stuff> ::= <ascii-char> <stuff>
          | <null>

<size> ::= <digit> <size>
         | <null>

<date> ::= <month> SPC <day> SPC <time>
         | <month> SPC <day> SPC SPC <year>
         | <month> SPC <day> SPC <year> SPC

<month> ::= "Jan" | "Feb" | "Mar" | "Apr" | "May" | "Jun" | "Jul"
          | "Aug" | "Sep" | "Oct" | "Nov" | "Dec"

<day> ::= <digit> <digit>
        |   SPC   <digit>

<time> ::= <digit> <digit> ":" <digit> <digit>
         |   SPC   <digit> ":" <digit> <digit>

<year> ::= <digit> <digit> <digit> <digit>

<longname> ::= <name> SPC "->" SPC <name>
             | <name>

<name> ::= <ascii-char> <name>
         | <ascii-char>

<comment> ::= <stuff> <crlf>

<crlf> ::= CR LF

<null> ::=

Notes:
Month names are case insensitive.

<ascii-char> is any character from 1 to 255 of any ISO 8859-1, excluding
CR and LF.

Names in capital letters and without angle bracket ("<>") quotes are names
of ASCII control characters.

The ambiguity between <comment> and <line> must always be resolved in
favor of <line>.

There is also an ambiguity between the first and second forms of
<longname>.  Resolution of this ambiguity is up to the programmer, but
always prefering the first form is likely to give reasonable results.  It
may be possible to use the FTP NLST command to obtain disambiguating
information.

More ambiguity: Should <stuff> ever contain a legal <date>, all bets are off!

Author:

Peter N Lewis <peter.lewis@info.curtin.edu.au>
10 Earlston way,
Booragoon, 6154, WA,
AUSTRALIA

Contributors:

Quinn <quinn@cs.uwa.edu.au>
James W. Matthews <James.W.Matthews@Dartmouth.EDU>
Stephen Trier<trier@ins.cwru.edu>

Obviously, any errors in this document are my own!
-- 
Peter N Lewis <peter.lewis@info.curtin.edu.au> - Macintosh TCP fingerpainter
FTP my programs from ftp://ftp.share.com/ or ftp://ftp.amug.org/

-----------[000490][next][prev][last][first]----------------------------------------------------
Date:      Tue, 20 Dec 1994 11:20:10 +0800
From:      peter.lewis@info.curtin.edu.au (Peter N Lewis)
To:        comp.protocols.tcp-ip
Subject:   Re: ftp and the SYST command (HELP) - FTP Structured List Command

FTP Structured List command - Peter N Lewis, Mar 1994.

This document defines two proposed FTP commands to work with structured
lists in an attempt to make a machine-readable LIST format.  This is
necessary to give people a viable alternative to attemptint to parse the
LIST command (a method for which is documented elsewhere).

The two commands are

XMRA <attribute-list>
XMRL <path>
[Should these be names differently?  SLAT dn SLST?]
[Should there be a command to return the attributes supported]

The first specifies the attributes the client wishes the server to list
for each file.  The latter is the equivalent of the LIST/NLST command, and
asks the server to display the specified files/directories in the format
documented here with attributes as listed in the XMRA command.

<attribute-list> is simply a comma seperated list of attribute names.  An
attribute name is a sequence of ascii seven bit characters from the 32 to
126 excluding comma and space.  Current attribute names should include
only characters from the set [A-Z,a-z,0-9,_,-,+,.] starting with a letter
to allow for future expansion.  Attribute names are case insensitive. 
This document will attempt to describe all the standard attributes. 
System specific attributes should be in the format
<system-name>-<attribute> (where <system-name> is the System Name as
defined in the Assigned Numbers document).  Attributes starting with X
(including both Xattribute and UNIX-Xattribute) are reserved for
experimental or local attributes.  A server can respond to a system
specific attribute even if it is not actually on the system.  In general,
generic attributes should be used whereever possible, and clients MUST be
able to back off to the defined minimal set of attributes.

The XMRA command should reply with either:
200 Everything's cool
502 Command not implemented
504 <attribute-list> command not implemented for that parameter.
(<attribute-list> is a subset of the parameters to the XMRA command, and
may be either a single attribute or all attributes that the server does
not understand)

The XMRL command works in the same way as the LIST and NLST command (and
SHOULD list the same files, with the exception that the special files like
 and .. MUST NOT be listed).  The returned list should consist of one
line per file/directory which is a comma seperated list.  Characters in
each entry may be escaped using the MIME = method (ie, = followed by two
hex character is translated into that character).  Characters SHOULD come
from the ISO-8859-1 character set (which SHOULD be the same characterset
you use in the filename parameters (LIST, STOR, etc)).  Character 0-31,
127, 255, space, comma (,) and equals (=) MUST be escaped, other
characters SHOULD NOT be escaped.  The comma seperated entries in the list
are the requested attributes (in the same order) as specified in the XMRA
command (which MUST preceed it).

The XMRL command should reply with any response the LIST command can give.

Mandatory Attributes:

[ These need a lot of work!]

NAME: File name
TYPE: "file" or "dir" or ???
LINK: 0 or 1 ???
[How about LINK-TO: "item link points to" or empty string]
SIZE: Approximate size of the file in bytes, or for directories, either
empty string or a count of contained files and directories
PERM: [g-][p-][c-][d-][r-] (case insensitive) where dash (-) means "can't"
and letter means "can", the characters are (in this order):
g - Can get (RETR) the file (or files inside this directory for a directory)
  p - Can put (STOR) over this file (or in this directory for a directory)
  c - Can cd (CWD) into this directory
  d - Can delete (DELE/RMD) this file or directory
  r - Can rename (RNFR/RNTO) this file or directory
DATE: Modification date or closest approximation (create date?). 
Formatted as 19940227114609 (or what is an ISO date format???)

Other Generic Attributes:
MOD_DATE
CREATE_DATE
ACCESS_DATE
BACKUP_DATE

System Specific Attributes:

Macintosh:
MACOS-FTYPE: File type (4-letter code)
MACOS-FCREATOR: File creator
MACOS-FLAGS: Finder flags
MACOS-DIRID: DirID (empty string for files)
MACOS-DLEN: Data fork length
MACOS-RLEN: Resource fork length
MACOS-PARID: Parent DirID

Unix:
UNIX-MODE: 9-letter mode
UNIX-FTYPE: 1-letter mode?
UNIX-LINKS: Link count
UNIX-UID: number
UNIX-GID: number
UNIX-UNAME: name
UNIX-GNAME: name

VMS:

Authors:
Peter N Lewis <peter.lewis@info.curtin.edu.au>
Robert J Minich <minich@a.cs.okstate.edu>

Contributers:
Steve Fosdick <stevef@aom.bt.co.uk>
Pete Resnick <resnick@cogsci.uiuc.edu>
Guy Harris <guy@netapp.com>
Mike Gleason <mgleason@cse.unl.edu>
-- 
The probability of me making it to the WWDC in 1995 is currently: 52%

-----------[000491][next][prev][last][first]----------------------------------------------------
Date:      Wed, 21 Dec 94 08:14:56 PDT
From:      adar0@routers.com
To:        comp.sys.sun.admin,comp.protocols.tcp-ip
Subject:   Re: [Q]: What is MX entry in dns.


> Could anyone please explain what is the meaning of MX entry in
> the in.named configuration file.
> 
> For example, in my domain's server named configuration file
> 
> goofy	IN	A	199.1.1.10
> 	IN	MX	10 sub-domain.domain.com.
> 

The MX record is a "Mail eXchange" pointer that essentially is returned
to other machine's name resolvers to tell other machines where SMTP email
messages should be sent to for the domain or host that it is applied to.

In your example, you cut out the stuff above "goofy IN ...", however there
is likely a "$ORIGIN" statement somewhere above in your example.  So if
it says "$ORIGIN domain.com.", the above example then means deliver all
mail destined for anywhere in the "domain.com" to the machine named
"sub-domain.domain.com".

The MX record can describe "mail receiving machines" for full domains,
sub-domains, or individual hosts.  An individual machine might look
something like:

$ORIGIN subdomain.domain.com.
wizard          IN      A       198.247.251.66
ace             IN      A       198.247.251.67
wizard          IN      MX      100 ace.subdomain.domain.com
		IN	MX	50 wizard.subdomain.domain.com

Which means; deliver all mail for wizard.subdomain.domain.com to the
wizard machine if it is up (preference of 50), or if it is not up then
deliver to ace (preference of 100).  The lower the preference number,
the more likely mail will be delivered to that machine if it is 
available and accessible by the sending machine.

Most records in the in.named configuration files have some dependancies
upon other records that appear before them in the file.  So when asking
questions it is sometimes easier to answer them if you include stuff
above the record in question.

Hope that helps...
Rich Adamson
adar0@routers.com




-----------[000492][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 02:12:44 GMT
From:      tli@cisco.com (Tony Li)
To:        comp.protocols.tcp-ip
Subject:   Re: Internet routing protocols

In article <3d6914$9s2@beatles.cselt.stet.it> teisa@beatles.cselt.stet.it
(Claudio Teisa) writes: 

     looking at the RFCs and other Internet documentation, several
    protocols related to the routing, seems to exist:
    
    local routing:	ARP,RARP,INARP,NARP,...
    IGPs:		RIP,RIP-2,OSPF,DUAL IS-IS,...
    EGPs:		EGP,EGP-2,BGP,BGP-2,BGP-3,BGP-4,...
    others:		DHCP,...
    
     But: what is their actual deployment in Internet?
    
I would suspect that almost all are deployed somewhere in the Internet.
It's a big place.  Of those that are common, there's ARP, RIP, OSPF,
Integrated IS-IS, BGP-4, and IGRP (non-standard).

     Which are obsolete, actual or future? 

I would argue that INARP, NARP, RIP, RIP-2, EGP, EGP-2, BGP, BGP-2, and
BGP-3 are all obsolete.  This does NOT mean that people aren't using
them...  ;-)

     Which are already implemented in commercial products?
    
I know of ARP, RARP, RIP, OSPF, Integrated ISIS, EGP-2, BGP-2 thru 4, and
DHCP.

Tony

-----------[000493][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 02:15:28 GMT
From:      mgleason@cse.unl.edu (Mike Gleason)
To:        comp.protocols.tcp-ip
Subject:   Re: ftp and the SYST command (HELP) - FTP Structured List Command

peter.lewis@info.curtin.edu.au (Peter N Lewis) writes:

|XMRA <attribute-list>
|XMRL <path>
 
|[Should these be names differently?  SLAT dn SLST?]

I am partial to SLAT/SLST for arbitrary reasons....

|[Should there be a command to return the attributes supported]

You could have XMRA return the attributes in use as it's response code.

 > XMRA NAME TYPE ACCESS_DATE MOD_DATE
 < 200 XMRA command okay.  Using: NAME TYPE MOD_DATE

(Note ACCESS_DATE was not returned).

Or:

 > XMRA NAME TYPE ACCESS_DATE MOD_DATE
 < 200 XMRA command okay for: 1 1 0 1

(Have 1's correspond to supported attrs, 0's for unsupported.)


|The first specifies the attributes the client wishes the server to list
|for each file.

This would be terrific, if we could get it supported!

|<attribute-list> is simply a comma seperated list of attribute names.

For ease of parsing, I would prefer space-delimited.  Then I could use
a function like C's sscanf() to strip out what I want without looking for
commas.

|SIZE: Approximate size of the file in bytes, or for directories, either
|empty string or a count of contained files and directories

I've been thinking about this, I think we should have a separate attribute
for the item count in the directories.  If I'm not mistaken, MacOS for example,
can't tell you this count without counting each item in the directory manually.

Let's say I am in a directory /archive, with subdirectories windoze, mac,
amiga, and unix. Each of those subdirectories has 500 files in it. If I want
to do a 'dir' from /archive, I might just want a list of directories without
caring what's in them.  For example, a user might 'cd /archive' and do a
directory listing, and recognize instantly that she wants to go to the mac
directory.  Thrashing the disk just to count all the files in those huge
subdirectories would be a waste of time in cases like that.

Suggestion:  Add 'ITEM_COUNT' which would make sense when an item is a
directory.  This could then be used only when needed.

|DATE: Modification date or closest approximation (create date?). 
|Formatted as 19940227114609 (or what is an ISO date format???)

Just a note:  I think it's extremely important that all dates be in UTC
time.  I've been trying to get use MDTM to get this in UTC, but I've found
that many just use the site's local time.  This is useless if your goal is
to mirror items on remote hosts.

Another thing.  Regular expressions.  Most unix FTP servers allow you to
use a wildcard to limit the number of matches.  I'd like to be able to do
that still, yet have a way to turn off the special meaning of those wildcard
characters too.

|The XMRL command works in the same way as the LIST and NLST command (and
|SHOULD list the same files, with the exception that the special files like
| and .. MUST NOT be listed).

Should there be a PARENT_PATH attribute for directories?

--
===== Mike Gleason <mgleason@cse.unl.edu> ================= Go Huskers! =====
Current version of NcFTP is 1.8.8, and is available from cse.unl.edu, in the
/pub/mgleason/ncftp directory.  Public permissions are turned on daily at 5pm
CST;  To decode the gzip'd tar file, do "gunzip -c ncftp.tgz | tar xvof -". 

-----------[000494][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 94 06:16:03 GMT
From:      ssatchell@BIX.com (ssatchell on BIX)
To:        alt.security,alt.security.pgp,alt.security.ripem,comp.protocols.tcp-ip,comp.security.misc,sci.crypt
Subject:   Re: IEEE PASC Study Group on Encryption APIs

balenson@tis.com (David M. Balenson) writes:

>IEEE PASC Study Group on Encryption APIs
 
>The IEEE PASC (Portable Applications Standards Committee) group will be
>hosting a study group on API's for encryption and other cryptographic
>services as an extension to the POSIX standards.  The second meeting on
>this will be held at the Bonaventure Hotel, Ft. Lauderdale, FL, Jan
>16-19, 1995 in conjunction with the quarterly PASC meetings.  There is
>a meeting fee of $100/day or $350 for the week that includes lunches.
>The contact for this work is David Balenson who can be reached at
>+1.301.854.6889 or balenson@tis.com.  A general meeting announcement is
>available on request from Todd Campbell at NAPS International, who can
>be reached at +1.612.888.0074 or tc@bungia.mn.org
 
>The study group will investigate the feasibility of developing IEEE and
>ISO standards for application program interfaces to encryption and
>other cryptographic mechanisms, including data encryption,
>authentication, data integrity, digital signatures, key management, and
>timestamping.  The PASC study group would like to invite broad
>participation from interested parties consisting of private
>individuals, industry, government, users and producers.

If the PASC study group *really* wants to invite broad participation of
private individuals, they would create a newsgroup in the IEEE hierarchy
and invite people to post papers there.  (To reduce the flood of junk,
the newsgroup could be moderated.)

That $350 for the week doesn't include the $135/day for the hotel room,
the plane ticket which can cost upwards of $1300 depending on where you
are coming from, and if it works like most of these meetings they will
break up into discussion groups, only one of which you can participate
in.

Last time I looked at attending a standards meeting in Ft. Laudardale,
it would cost me upwards of $2100 -- and this looks to be the same scale
of cost.

I challange the PASC group to hold an electronic meeting as I described,
and present a summary of the traffic to the meeting in January.

-----------[000495][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 15:21:07 -0500
From:      barmar@nic.near.net (Barry Margolin)
To:        comp.protocols.tcp-ip
Subject:   Re: Source Routing?

In article <ricard.36.006B524C@axis.se> ricard@axis.se (Ricard Wolf) writes:
>Is IP source routining really used in practice? I've seen it in the IP RFC, 
>and some implementations of 'ping' seem to be able to set up various source 
>routing fields, but I've never seen it being used in the real world.

  It's mostly intended for debugging.  For instance, if there's a
problem with a routing protocol, source routing can be used to force use of
alternate routes, in order to login to one of the misbehaving routers and
reconfigure it.
-- 

Barry Margolin
BBN Internet Services Corp.
barmar@near.net

-----------[000496][next][prev][last][first]----------------------------------------------------
Date:      Wed, 21 Dec 94 15:31:48 PDT
From:      Steve Hathaway <steve@oem.state.or.us>
To:        comp.unix.unixware,comp.protocols.tcp-ip lfa@olie.wvitcoe.wvnet.edu
Subject:   Re: [Q]Slip Service


In article <19941215.113819.357232.NETNEWS@WVNVM.WVNET.EDU>, 
<lfa@olie.wvitcoe.wvnet.edu> writes:
> 
> I am looking for advice, pointers, experiences, etc in setting up a slip
> server.  There may be 400+ potential users with a maximum of 15 users at
> a time.
> 
> TAI
> Larry
> 
====================
AT&T Unix SVR4.0
  I have had no luck getting SLIP to work over a dial-in/dial-out line.
  The implementation of SLIP works over a dedicated RS-232 line.

I do not know if Unixware supports dial-in/dial-out SLIP.

There are some terminal server products that do handle dial-in SLIP,
but I have not evaluated these products or their capabilities.

- Steve Hathaway <steve@oem.state.or.us>
  Oregon State Police // Emergency Management Division


-----------[000497][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 17:01:28 -0500
From:      barmar@nic.near.net (Barry Margolin)
To:        comp.protocols.tcp-ip
Subject:   Re: getting ip_addr from FD

In article <3d9pl8$sfo@ra.nrl.navy.mil> danmcd@sundance.itd.nrl.navy.mil (Dan McDonald) writes:
>In article <3d9on9$1a28@tequesta.gate.net>,
>Michael S. Scheidell <scheidel@gate.net> wrote:
>>Is there an EASY way to get the remote host IP address from STDIN FD?
>>is getipaddr(STDIN,&in_addr):
>
>How about getpeername()?  I am assuming you're on a BSD-ish machine that
>uses sockets.  I don't know winsock or any other API well enough to answer
>your question.

This will work fine if stdin is a socket (e.g. if you use this in a server
that was started by inetd).  If you're trying to get the remote host of a
telnet or rlogin login session, stdin is a pty connected to the local
telnetd or rlogind process, not a socket to the remote host, so this won't
help you.  The only easy way to get the remote host is the entry in
/etc/utmp, but if the name is longer than 16 characters it will be
truncated.  If that's not adequate you'll have to write a setgid program
that rummages through kernel memory tracing all the connections.

-- 

Barry Margolin
BBN Internet Services Corp.
barmar@near.net

-----------[000498][next][prev][last][first]----------------------------------------------------
Date:      Wed, 21 Dec 1994 09:34:35 GMT
From:      moshel@tasu60.nsc.com (Moshe Linzer)
To:        comp.protocols.nfs,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.sys.novell
Subject:   Wollongong NFS and Novell

I am trying to install Wollongong's TCP/IP ODI driver on a PC together with
IPXODI.  The thing seems to work, but when I bring up NFS, I get all sorts 
of strange errors when trying to access the Novell drive, and the NFS is
VERY slow.  Anyone do this successfully?  Am I missing some essential
NET.CFG parameters?  Any help would be appreciated.

Moshe

--
                                                             (O-O)
---------------------------------------------------------oOO--(_)--OOo------
Moshe Linzer				Phone:	(972) 9-594-247
Unix Systems Manager			Fax:	(972) 9-558-322
National Semiconductor, Israel		E-mail:	moshel@taux01.nsc.com
----------------------------------------------------------------------------

-----------[000499][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 11:07:39 GMT
From:      jim@cs.strath.ac.uk (Jim Reid)
To:        comp.protocols.tcp-ip
Subject:   Re: any RIP-2 sources

In article <davidp.256.00A1F438@qpsx.oz.au> davidp@qpsx.oz.au (David Pascoe) writes:

   I have gotten hold of the ucsd hamradio copy of RIPv2 for NOS
   (whatever that is) and was wondering if there are any other RIP
   implementations available out there.

gated - the all-singing, all-dancing UNIX routing software - supports
RIP V2 and a host of routing protocols.

-----------[000500][next][prev][last][first]----------------------------------------------------
Date:      Wed, 21 Dec 1994 16:28:17
From:      BFLYNN@email.unc.edu (Brian Flynn)
To:        comp.sys.novell,comp.sys.protocols.tcpip,,alt.winsock
Subject:   Re: Packet frame interference ?

In article <3d47n7$6r@hecate.umd.edu> ceham@w3eax.umd.edu (Maurice De Vidts NE3S) writes:

>I have setup Trumpet winsock using packet drivers ( -n )
>on a Netware 3.11 over our company LAN.  It seems to work
>but our network administrator has prohibited me from using it
>since he claims it crashes the server"
 
>I know I may be using a different frame type, but it seems
>a little far fetched that different ethernet frames will
>cause such a crisis.

Someone more familar with the Trumpet stack will have to handle that part
(using packet driver ??).  Different frame types won't matter a bit.  There
could be a potential problem with the AMOUNT of traffic you could generate,
but that would show up with any frame type.

>I would like some techincal pointers to understand
>if I am truly creating a problem.  The novell winsock
>driver he has installed does not work with any
>of the winsock telnets I have found on the net.

The his drivers are  not set up right.  I'm running novell winsock right now,
running any winsock program I want.

>I would appreciate a clarification on this issue !

Clear enough?  Seriously, ask your adminstrator to fix your problems.  If he
can't do it, go over his head.  IMHO, if he doesn't know how to make his
winsock run, and doesn't know where to ask, he really shouldn't be taking care
of a LAN.

Oh, and Merry Christmas...

Brian Flynn
CNE/CNA
School of Nursing, UNC-CH
Go 'Heels
*******************************************************************************
Disclaimer:  I don't tell SON how to teach nursing and they don't tell me what
to post.



-----------[000501][next][prev][last][first]----------------------------------------------------
Date:      Wed, 21 Dec 1994 11:31:02 +0000 (GMT)
From:      general admin login <sysmgr@vacgen.fisons.co.uk>
To:        comp.protocols.tcp-ip,comp.unix.questions,comp.unix.large,comp.unix.sys5.r4
Subject:   Printing to JetDirect from drs6000

We have a HP plotter with a JetDirect card on the same ethernet/tcp/ip 
network as our drs6000. I would like to spool print requests for the 
plotter on the drs6000 using the lp print service. 
We are still running drsnx v6 and icl say this cannot be done without 
upgrading to drsnx v7. Anyone got any ideas?

	Thanks 
	

---------------------------------------
Jim Cozens

Fisons Instruments Vacuum Generators

EMail
      jcozens@vacgen.fisons.co.uk 
or
      sysmgr@vacgen.fisons.co.uk

Tel.  01424 851291
Fax.  01424 851489



-----------[000502][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 12:30:18 GMT
From:      carlson@Xylogics.COM (James Carlson)
To:        comp.protocols.tcp-ip
Subject:   Re: Internet routing protocols

In article <3d82us$4e0@cronkite.cisco.com>, tli@cisco.com (Tony Li) writes:
|> In article <3d6914$9s2@beatles.cselt.stet.it> teisa@beatles.cselt.stet.it
|> (Claudio Teisa) writes: 
|> 
|>      looking at the RFCs and other Internet documentation, several
|>     protocols related to the routing, seems to exist:
|>     
|>     local routing:	ARP,RARP,INARP,NARP,...
|>     IGPs:		RIP,RIP-2,OSPF,DUAL IS-IS,...
|>     EGPs:		EGP,EGP-2,BGP,BGP-2,BGP-3,BGP-4,...
|>     others:		DHCP,...

You left out ICMP, which has messages which fall somewhere between local
routing and IGPs.  (Redirect and router discovery protocols.)

[...]
|> I would argue that INARP, NARP, RIP, RIP-2, EGP, EGP-2, BGP, BGP-2, and
|> BGP-3 are all obsolete.  This does NOT mean that people aren't using
|> them...  ;-)

Hmpf.  Just because you prefer LS to DV algorithms doesn't mean that the
latter are obsolete!

|>      Which are already implemented in commercial products?
|>     
|> I know of ARP, RARP, RIP, OSPF, Integrated ISIS, EGP-2, BGP-2 thru 4, and
|> DHCP.

... and both ICMP and RIP-2.

---
James Carlson <carlson@xylogics.com>            Tel:  +1 617 272 8140
Annex Software Support / Xylogics, Inc.               +1 800 225 3317
53 Third Avenue / Burlington MA  01803-4491     Fax:  +1 617 272 2618

-----------[000503][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 10:55:45 +0100
From:      casper@fwi.uva.nl (Casper H.S. Dik)
To:        comp.protocols.tcp-ip,comp.unix.solaris
Subject:   Re: dup2() before connect() won't work on SunOS5.3

aray@pipeline.com (Arjun Ray) writes:

>Your code accomplishes this by dup2(). But dup2() doesn't need the old
>socket/descriptor to be open.
>Your problem may be due to the fact that the descriptor is still (technically)
>open, but in fact unusable. So, instead of relying on the kernel (actually,
>the STREAMS system, in this case the sockmod module pushed onto the /dev/tcp
>driver) to do everything right, it could be a good idea to _force_ it to do
>everything correctly. (See below.)

The basic problem is that the socket library caches per fd information.
When dupping over an descriptor that is already in use as a socket, the
system may get confused.  That is a bug in the socket library.

Unfortunately, close may not fix this.

Casper

-----------[000504][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 13:23:13 GMT
From:      jrmt@uk.gdscorp.com (Jon Thackray)
To:        comp.protocols.tcp-ip
Subject:   Re: IPng/IPv6 Documentation Pointer Request

BESANCON Thierry writes:
>It is written in French but lots of keywords/references are in English.
>You might find what you want.

There's also an excellent guide at:

   http://playground.sun.com/pub/ipng/html/ipng-main.html

Well worth checking out.

Regards,
Jon.
-- 
// Jon Thackray, Cambridge, England.   ><>
//        Work:  jrmt@uk.gdscorp.com      +44 (01223) 371918
// Rest & Play:  jrmt@froggy.demon.co.uk  +44 (01223) 502156

-----------[000505][next][prev][last][first]----------------------------------------------------
Date:      Wed, 21 Dec 94 20:28:37 PDT
From:      PP001287@interramp.com
To:        comp.protocols.tcp-ip
Subject:   TCP-IP for DOS???


Does anyone know of some decent TCP-IP software that DOESN't run under Windows? 

I have to believe other people have had similar problems I have experienced 
with Windows hanging up (Never giving the same error message), not to mention 
the agony of running Mosiac under WIN32.

The people responsible for Chamelion should be shot.

-----------[000506][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 13:44:34 GMT
From:      hethmon@apac.ag.utk.edu (Paul Hethmon)
To:        comp.protocols.tcp-ip
Subject:   Re: ftp and the SYST command (HELP) - FTP Structured List Command

I'd like to make some general comments on this whole thing.

Right now, I'm in the process of writing a FTP server and I've
run up against these problems again and again. At the moment,
it seems the only thing to do is either contact every author
of an automated client (not very likely) or try and support
every new command someone has come up with to try and
deal with the problem.

The thing is is that introduces many more problems when
the FTP standard is "extended" unofficially. Not every
client and server will support the extensions or support them
in the correct way.

It seems to me that the time has come to rewrite the
FTP standard to reflect the usage of FTP today. I've conversed
with a few people via email over doing just this, though I 
haven't pursued it further. It seems that Peter Lewis' group
has made good progress towards extending the standard
in a very useful way. I think that an official RFC needs to be
done to integrate all of the FTP standard into one document.
Currently we have 959, 1123, and 1579 that all have a bearing
on it (and I've probably missed some).

Is anyone else ready to pursue an official RFC?

Paul Hethmon
Programmer/Analyst & IBM Certified OS/2 Engineer
Agricultural Policy Analysis Center
The University of Tennessee at Knoxville
hethmon@apac.ag.utk.edu  ==  615-974-3666



-----------[000507][next][prev][last][first]----------------------------------------------------
Date:      22 Dec 1994 04:16:44 -0800
From:      cgi@crl.com (Paul Smith)
To:        comp.protocols.tcp-ip
Subject:   Re: Summary: How to detect dead socket under HP-UX 9.03 (and more generally)?

W. Richard Stevens (rstevens@noao.edu) wrote:
[snip]
: Kind of.  You really need to understand what's going on below you.  The
: first write succeeds (assuming it's a small write, << SO_SNDBUF) because
: all that write() normally does for a TCP socket is copy the data into the
: socket buffer and return.  (TCP will decide when to send a segment, based
: on a number of factors.)  TCP sends the segment, and the other ends responds
: with an RST some number of milliseconds later.  TCP remembers that it received
: the RST assuming the application doesn't already have an I/O operation pending
: on the socket.  Then you come along after the RST has been received and do
: the 2nd write().  Since TCP has received an RST on the connection, bang,
: you get SIGPIPE.  The rule is: if you write() to a connection that has
: received an RST, you get SIGPIPE.  If you read() from a connection that has

Does send() return with errno != EINTR in this scenario?? with it set to
some error that indicates a broke circuit?

...


-----------[000508][next][prev][last][first]----------------------------------------------------
Date:      22 Dec 1994 04:37:44 -0800
From:      cgi@crl.com (Paul Smith)
To:        comp.protocols.tcp-ip
Subject:   Re: Source code from Stevens' book

Ed Martini (emartini@netcom.com) wrote:
: I'm looking for an FTP site for the source code from Comer/Stevens
: Internetworking with TCP/IP Volume III.  Any pointers would be appreciated.

Any major ftp site, under the /archive/published/books/.. directory.

Just greped my internet host dir listings and found:
ftp.uu.net
dutepp0.et.tudlft.nl


: Ed Martini
: Senior Software Engineer
: Digital Video Systems
: Sunnyvale, CA

-----------[000509][next][prev][last][first]----------------------------------------------------
Date:      22 Dec 1994 04:44:36 -0800
From:      cgi@crl.com (Paul Smith)
To:        comp.protocols.tcp-ip
Subject:   Re: Kermit with MS TCP/IP-32?

Ray Harder (rharder@ctp.org) wrote:
: Is it possible to use Kermit with Windows For Workgroups 3.11 and
: TCP/IP-32 from Microsoft? I installed Kermit 3.13 on a laptop with a
: PCMCIA card for 10BaseT Ethernet access. If I load Kermit under DOS (6.2)
: it works beautifully. If I load windows first and then load Kermit it
: works fine also -until I try to access the server using TCP/IP. At the
: Kermit prompt, I type "Set Port TCP/IP 163.xxx.xxx.xxx" (With the actual
: IP address of course) and then on the next line I type "Connect."
 
: Sometimes the connection is made for a few seconds. I have even been able
: to login once or twice, but always within a screen or two, it crashes
: --taking windows down with it. I cannot access it from Kermit within
: Windows when using TCP/IP, but from DOS it works fine. There seems to be
: a conflict between the built-in TCP and the 32 bit version. Adjusting the
: usual settings does not seem to have any effect. In the Docs for Kermit
: it mentions that it can be used in a couple of others including Beam and
: Whiteside. Can Kermit be modified to use an existing TCP/IP stack?

The problem is that kermit is built with a built in WatTCP stack that
attaches directly to the NDIS pkt driver.  If your winsock stack is 
running, that NDIS coinnection is not avaible.  You and some others are
looking for a winsock version of kermit, not the default WatTCP.

: Alternately, is there a way to get VT220 (Or better V320) emulation over
: a TCP/IP connection from Windows using Public Domain software? Is there a
: version of Telnet that allows V220 or 320 for a Windows environment?

WinQVT is the best package I've heard of.  It's on all the winsock
app ftp sites have it.  The older (may be current) version I have is
qvtws396.zip.



: ***********************************************************************
: ****
: * Raymond G. Harder                  "Can't walk today, I don't feel
: well."*
: * Educational Technology Consultant  "Why don't you sit out in the sun?" 
:  *
: * 909-983-4713                       "What? People will think I'm lazy!" 
:  *
: * rharder@ctp.org                         -- My 92 year old grandmother  
:  *
: ***********************************************************************
: ****

-----------[000510][next][prev][last][first]----------------------------------------------------
Date:      Wed, 21 Dec 1994 16:01:33 GMT
From:      gdonoso@tolten.puc.cl (Gerardo Donoso Contreras)
To:        comp.protocols.tcp-ip
Subject:   Assigning IP multicast address

Does anybody know how can i assign an IP multicast address to my host,
to joint a multicast hostgroup????


Thank you

-----------[000511][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 17:29:20 GMT
From:      sbrog@drexler.nwest.mccaw.com (Steve Brog)
To:        comp.protocols.tcp-ip
Subject:   Re: Summary: How to detect dead socket under HP-UX 9.03 (and more generally)?

In article <3d5ab2$efl@orca1.vic.design.telecom.com.au>  
jamesc@osaiphp1.osa.csi.itg.telecom.com.au (James Crawford) writes:
> I have a slightly different question.  How can I detect that the
> socket connection has been closed (or better still shutdown) from
> the other end when writing to the socket.
> 
> It seems that I have no problem detecting this when doing a read,
> but when I write to the socket, the write will work fine, even if
> the other end has shutdown or closed their end of the conversation.
> 
> I also am running HPUX 9.03, but would prefer a generic solution
> if there is one.
> 
> Thanks for any advice.
> 
> James Crawford.

Setup a signal handler for SIGPIPE, this will be invoked when you write on  
the socket and the other side has disconnected.  You can set a flag in the  
signal handler and check on the status of the flag after the write.   This  
works in some cases, but I have found that in many cases two writes are  
actually needed.  The first write will fail (the message gets lost), but  
appears to work with a good return code. However, the second write will  
trigger the SIGPIPE.  In addition, you can use the SO_KEEPALIVE option to  
setsockopt() to force TCP to occasionally test the connection, this will  
sometimes allow the SIGPIPE to be returned on the first write to the  
socket as TCP has already figured out that it is dead.  Since my data is  
important to me, I end up sending back an application level "ACK" after  
each message (or message group) to confirm the other side received it when  
writing applications that transfer data from one node to another.  This is  
generally not needed if you are writing simple servers that return  
responses where the response is an implied acknowledgement.

You might also want to explore the TCP_NODELAY option of setsockopt() as  
well as it forces TCP to send short messages out promptly and not buffer  
them up internally until they are large enough to be sent efficiently.

Steve Brog

-----------[000512][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 17:30:17 GMT
From:      scheidel@gate.net (Michael S. Scheidell)
To:        comp.protocols.tcp-ip
Subject:   getting ip_addr from FD

Is there an EASY way to get the remote host IP address from STDIN FD?
is getipaddr(STDIN,&in_addr):

if not, is there a HARD way?


--
Michael S. Scheidell                    Florida Datamation, Inc. 
scheidell@fdma.com                      (407) 241-2966
Distributors of QNX Real Time OS        http://www.gate.net/~scheidel/
NOTICE: Opinions expressed here are not the opinions of employees!

-----------[000513][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 17:46:16 GMT
From:      danmcd@sundance.itd.nrl.navy.mil (Dan McDonald)
To:        comp.protocols.tcp-ip
Subject:   Re: getting ip_addr from FD

In article <3d9on9$1a28@tequesta.gate.net>,
Michael S. Scheidell <scheidel@gate.net> wrote:
>Is there an EASY way to get the remote host IP address from STDIN FD?
>is getipaddr(STDIN,&in_addr):

How about getpeername()?  I am assuming you're on a BSD-ish machine that
uses sockets.  I don't know winsock or any other API well enough to answer
your question.

The man page for getpeername() explains things very nicely.
-- 
Daniel L. McDonald | Mail:  danmcd@itd.nrl.navy.mil -------------------------+
Computer Scientist | WWW:   http://wintermute.itd.nrl.navy.mil/danmcd.html   |
Naval Research Lab | Phone: (202) 404-7122        #include <disclaimer.h>    |
Washington, DC     | "Rise from the ashes, A blaze of everyday glory" - Rush +

-----------[000514][next][prev][last][first]----------------------------------------------------
Date:      22 Dec 1994 02:07:04 -0500
From:      aray@pipeline.com (Arjun Ray)
To:        comp.protocols.tcp-ip,comp.unix.solaris
Subject:   Re: dup2() before connect() won't work on SunOS5.3

In article <3d8c5j$80q@pipe2.pipeline.com>,
Arjun Ray <aray@pipeline.com> wrote:
>In article <3d6qt8$hg9@erinews.ericsson.se>,
>Ulrik Sandstr|m <ehssand@aom.ericsson.se> wrote:
>>Hi,
>>
>>I have problems connecting a socket after calling dup2(). Here is 
>>the scenario:
>>Code example:
>>
>>	...
>>	new_sock = socket(AF_INET, SOCK_STREAM, 0);
>>	dup2(new_sock, old_sock);
>>	close(new_sock);
>>	connect(old_sock, (struct sockaddr*)address,
>>		sizeof(struct sockaddr)); // Returns EAFNOSUPPORT
>>	...
>>
>I suggest a paranoid approach using the following sequence of calls:
>
>	shutdown( old_sock, 2 ) ; /* --[ forces buffer flushes ]--*/
>	close( old_sock ) ; /* --[ make sure the old STREAM is dismantled ]-- */
>	new_sock = socket( AF_INET, SOCK_STREAM, 0 ) ;
>	connect( old_sock, (struct sockaddr*) address, sizeof(struct sockaddr) ) ;
             ^^^^^^^^

Sorry! That should have been new_sock.

>	dup2( new_sock, old_sock ) ;
>	close( new_sock ) ;
>

Arjun Ray
The Pipeline Network Inc.



-----------[000515][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 17:58:55 GMT
From:      scheidel@gate.net (Michael S. Scheidell)
To:        comp.protocols.tcp-ip
Subject:   Re: getting ip_addr from FD

Dan McDonald (danmcd@sundance.itd.nrl.navy.mil) wrote:
: In article <3d9on9$1a28@tequesta.gate.net>,
: Michael S. Scheidell <scheidel@gate.net> wrote:
: >Is there an EASY way to get the remote host IP address from STDIN FD?
: >is getipaddr(STDIN,&in_addr):
 
: How about getpeername()?  I am assuming you're on a BSD-ish machine that
: uses sockets.  I don't know winsock or any other API well enough to answer
: your question.

I tried that first, ie 
rc = getpeername(0,(struct  sockaddr*)&his_name,&addrlen);

get back ENOSYS.
(ps, this is AFTER telnet/rlogin in, ie from user program)

--
Michael S. Scheidell                    Florida Datamation, Inc. 
scheidell@fdma.com                      (407) 241-2966
Distributors of QNX Real Time OS        http://www.gate.net/~scheidel/
NOTICE: Opinions expressed here are not the opinions of employees!

-----------[000516][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 18:56:12 GMT
From:      Rob Whittlesea <systems@freepressgrp.co.uk>
To:        comp.protocols.tcp-ip
Subject:   server nominated addressing

	At the moment I have a private addressing scheme of IP on a MAC based
network. And would like to attach this network to the INTERNET to allow
publication of a Newspaper ( Is that the right word for it on the
Internet ??? ) To do this I need a IP address resolver or translator. I
was wondering if anyone knew of such a piece of software or
hardware...... Could anybody please contact me at the email address below
if they have or know of such kit.....

	Thanks


	Gez 

	systems@freepressgrp.co.uk

-----------[000517][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 19:11:00 GMT
From:      carlson@Xylogics.COM (James Carlson)
To:        comp.protocols.tcp-ip
Subject:   Re: getting ip_addr from FD

In article <3d9qcv$1og9@tequesta.gate.net>, scheidel@gate.net (Michael S. Scheidell) writes:
|> Dan McDonald (danmcd@sundance.itd.nrl.navy.mil) wrote:
|> : In article <3d9on9$1a28@tequesta.gate.net>,
|> : Michael S. Scheidell <scheidel@gate.net> wrote:
|> : >Is there an EASY way to get the remote host IP address from STDIN FD?
|> : >is getipaddr(STDIN,&in_addr):
 
|> : How about getpeername()?  I am assuming you're on a BSD-ish machine that
|> : uses sockets.  I don't know winsock or any other API well enough to answer
|> : your question.
|> 
|> I tried that first, ie 
|> rc = getpeername(0,(struct  sockaddr*)&his_name,&addrlen);
|> 
|> get back ENOSYS.
|> (ps, this is AFTER telnet/rlogin in, ie from user program)

No, without grovelling in kernel structures, you're not going to be able
to get the peer IP address.

The problem is that your stdin/stdout/stderr are attached to the slave
side of a pseudo-terminal.  The master side is attached to an rlogind or
telnetd daemon, which is reading the data you write and forwarding it
to a socket.  The relationship between the pseudo-terminal master/slave
pair in use and the socket is known only to the daemon and to the kernel
itself.  Unless the daemon has written this relationship down somewhere
(AIX, in particular, writes the peer name into utmp), there's no obvious
way to get at it.

---
James Carlson <carlson@xylogics.com>            Tel:  +1 617 272 8140
Annex Software Support / Xylogics, Inc.               +1 800 225 3317
53 Third Avenue / Burlington MA  01803-4491     Fax:  +1 617 272 2618

-----------[000518][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 19:38:44 GMT
From:      rstevens@noao.edu (W. Richard Stevens)
To:        comp.protocols.tcp-ip
Subject:   Re: Summary: How to detect dead socket under HP-UX 9.03 (and more generally)?

> Setup a signal handler for SIGPIPE, this will be invoked when you write on  
> the socket and the other side has disconnected.  You can set a flag in the  
> signal handler and check on the status of the flag after the write.   This  
> works in some cases, but I have found that in many cases two writes are  
> actually needed.  The first write will fail (the message gets lost), but  
> appears to work with a good return code. However, the second write will  
> trigger the SIGPIPE.

Kind of.  You really need to understand what's going on below you.  The
first write succeeds (assuming it's a small write, << SO_SNDBUF) because
all that write() normally does for a TCP socket is copy the data into the
socket buffer and return.  (TCP will decide when to send a segment, based
on a number of factors.)  TCP sends the segment, and the other ends responds
with an RST some number of milliseconds later.  TCP remembers that it received
the RST assuming the application doesn't already have an I/O operation pending
on the socket.  Then you come along after the RST has been received and do
the 2nd write().  Since TCP has received an RST on the connection, bang,
you get SIGPIPE.  The rule is: if you write() to a connection that has
received an RST, you get SIGPIPE.  If you read() from a connection that has
received an RST, you'll get an error with errno set to ECONNRESET.

> You might also want to explore the TCP_NODELAY option of setsockopt() as  
> well as it forces TCP to send short messages out promptly and not buffer  
> them up internally until they are large enough to be sent efficiently.

I suggest understanding more about this option (aka, the Nagle algorithm)
before blindly turning it on (i.e., disableing the algorithm).

	Rich Stevens

-----------[000519][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 19:52:13 GMT
From:      scheidel@gate.net (Michael S. Scheidell)
To:        comp.protocols.tcp-ip
Subject:   Re: getting ip_addr from FD

Stephane Bortzmeyer (bortz@sheckley.cnam.fr) wrote:
: scheidel@gate.net (Michael S. Scheidell) writes:
 
: >I tried that first, ie 
: >rc = getpeername(0,(struct  sockaddr*)&his_name,&addrlen);
 
: >get back ENOSYS.
: >(ps, this is AFTER telnet/rlogin in, ie from user program)
 
: It cannot work: after a telnet (or rlogin), 0, the standard input, denotes
: the pseudo-device which is created by the telnet daemon, not the socket.
: Try to launch your program from rsh or inetd and it'll work.


OK, then refrased question:
after telnet (or rlogin) is there an EASY or hard way to find the host 
address of the remote host FROM  user PROGRAM, with the FD of stdin?
or with FD of stdin, get the socket associated with pseudo-device..



--
Michael S. Scheidell                    Florida Datamation, Inc. 
scheidell@fdma.com                      (407) 241-2966
Distributors of QNX Real Time OS        http://www.gate.net/~scheidel/
NOTICE: Opinions expressed here are not the opinions of employees!

-----------[000520][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 20:17:05 GMT
From:      cezar@PROBLEM_WITH_INEWS_DOMAIN_FILE (Cezar Cichocki)
To:        comp.protocols.tcp-ip
Subject:   Re: Urgent: SMTP Server for DOS

Steve Holden (Steve@scotia.co.uk) wrote:
: In article <3ctpe4$l7t@eis.calstate.edu>
:            shunt@eis.calstate.edu "Steve J. Hunt" writes:
 
: > juanm@ac.upc.es (Juan Manuel Bernabeu) writes:
: > 
: > > I'm looking desperatly for a DOS-TSR based SMTP server program.
: > > We have no NetWare server, so we can't use pegasus/mercury. :(
: > How about running Linux on a PC?  That would do what you want.  And 
: > Linux is pretty cheap.  Pegasus doesn't require a Netware server 
: > anymore, but it can't act as an "SMTP server" anyway.
: > 
: And it's perfect for the DOS environment ... ;-)
 
: > I don't think there are any DOS or Windows based sendmail type 
: > programs.
 
: The KA9Q variant I obtained from ftp.demon.co.uk as a part of their
: DIS216C package includes DOS-compatible SMTP send and receive code.
: I believe this is shareware, and I don't think it varies too radically
: from the standard KA9Q code. So you should be able to get source.
 
: Steve Holden     Scotia Electronic Publishing  +---------------------------+

Just try popmail. It's good enough for SMTP I think...

Cezar
cezar@freud.psych.uw.edu.pl


-----------[000521][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 20:40:15 GMT
From:      carlson@Xylogics.COM (James Carlson)
To:        comp.protocols.tcp-ip
Subject:   Re: getting ip_addr from FD

In article <3da11d$2333@tequesta.gate.net>, scheidel@gate.net (Michael S. Scheidell) writes:
|> Stephane Bortzmeyer (bortz@sheckley.cnam.fr) wrote:
|> : scheidel@gate.net (Michael S. Scheidell) writes:
 
|> : >I tried that first, ie 
|> : >rc = getpeername(0,(struct  sockaddr*)&his_name,&addrlen);
 
|> : >get back ENOSYS.
|> : >(ps, this is AFTER telnet/rlogin in, ie from user program)
 
|> : It cannot work: after a telnet (or rlogin), 0, the standard input, denotes
|> : the pseudo-device which is created by the telnet daemon, not the socket.
|> : Try to launch your program from rsh or inetd and it'll work.
|> 
|> 
|> OK, then refrased question:
|> after telnet (or rlogin) is there an EASY or hard way to find the host 
|> address of the remote host FROM  user PROGRAM, with the FD of stdin?
|> or with FD of stdin, get the socket associated with pseudo-device..

Ok, then a rephrased answer:

No, there is no EASY way to do this.

There's a pair of utilities called ofiles and pff in Dan Bernstein's
kstuff package which can be used to do this, but, again, this requires
grovelling around in kernel structures and thus isn't terribly portable.

A better way to do it would be to write your own telnetd or rlogind and
make it set the remote address and port as environment variables to the
spawned shell.

---
James Carlson <carlson@xylogics.com>            Tel:  +1 617 272 8140
Annex Software Support / Xylogics, Inc.               +1 800 225 3317
53 Third Avenue / Burlington MA  01803-4491     Fax:  +1 617 272 2618

-----------[000522][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 19:32:18 +0100
From:      bortz@sheckley.cnam.fr (Stephane Bortzmeyer)
To:        comp.protocols.tcp-ip
Subject:   Re: getting ip_addr from FD

scheidel@gate.net (Michael S. Scheidell) writes:

>I tried that first, ie 
>rc = getpeername(0,(struct  sockaddr*)&his_name,&addrlen);
 
>get back ENOSYS.
>(ps, this is AFTER telnet/rlogin in, ie from user program)

It cannot work: after a telnet (or rlogin), 0, the standard input, denotes
the pseudo-device which is created by the telnet daemon, not the socket.
Try to launch your program from rsh or inetd and it'll work.

Stephane Bortzmeyer           Conservatoire National des Arts et Metiers	
bortzmeyer@cnam.fr            Laboratoire d'Informatique
                              292, rue Saint-Martin			
tel: +33 (1) 40 27 27 31      75141 Paris Cedex 03
fax: +33 (1) 40 27 27 72      France	

"C'est la nuit qu'il est beau de croire a la lumiere." E. Rostand

http://web.cnam.fr/personnes/bortzmeyer/home_page.dom

					
	


-----------[000523][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 21:19:38 GMT
From:      bob@comlab.gtri.gatech.edu (Bob Baggerman)
To:        comp.protocols.tcp-ip,comp.unix.admin
Subject:   Multiple IP addresses for one network port

I've heard a number of people mention that some flavors of Unix allow
multiple IP addresses to be associated with one physical network port.
I understand Linux and BSD/OS have this capability.  A chap recently
mentioned there is a way to make SunOS 4.1.x and Solaris 2.x perform
this nifty feat but so far he won't part with the details.  I'm 
particularly interested in doing this on a SunOS 4.1.3 systems.  Any
ideas or helpful pointers would be appreciated.

bob

-- 
Bob Baggerman                    ! bob.baggerman@gtri.gatech.edu
GTRI/ITTL/CND                    ! bob@comlab.gtri.gatech.edu
Georgia Tech Research Institute  ! 404-894-7100 or 404-894-3525
Atlanta, GA  30332  USA          !

-----------[000524][next][prev][last][first]----------------------------------------------------
Date:      Thu, 22 Dec 1994 11:04:13 -0800
From:      amizuno@jessica.stanford.edu (Kuniaki Mizuno)
To:        comp.protocols.tcp-ip
Subject:   What is "TCP Slow-Start" ?

Hi Everyone ,

 I have interest about "TCP Slow-Start". 
Please tell me some URLs, papers,documents etc.

Actually, I need them soon. So I will be happy if you
tell them by mail. Give me Cristmas presents.

Regard,

-- 
Kuniaki Mizuno
NetOne Systems
E-mail: amizuno@jessica.stanford.edu

-----------[000525][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 21:44:26 GMT
From:      tli@cisco.com (Tony Li)
To:        comp.protocols.tcp-ip
Subject:   Re: Internet routing protocols

In article <3d974q$5gd@newhub.xylogics.com> carlson@xylogics.com writes:
    |> I would argue that INARP, NARP, RIP, RIP-2, EGP, EGP-2, BGP, BGP-2, and
    |> BGP-3 are all obsolete.  This does NOT mean that people aren't using
    |> them...  ;-)
    
    Hmpf.  Just because you prefer LS to DV algorithms doesn't mean that the
    latter are obsolete!
    
Hmpf yourself.  I prefer DV to LS, thank you very much.  Nevertheless,
there are certain DV protocols (and path vector and ...) that are simply
not up to snuff.

Tony

-----------[000526][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 21:45:26 GMT
From:      maartenh@stack.urc.tue.nl (Maarten van den Hoek)
To:        comp.protocols.tcp-ip
Subject:   Some help wanted on protocol

Hi all,

If I'am in the wrong newsgroup, please direct me to the right one.

I am writing a network simulator executing protocol 5 from Tanenbaum's
Computer Networks. I have the following problem: 

host1 is communicating with host2, host1 sends the last package, host 2
receives it and sends and ACK. The ACK gets lost (cause of noise). Now comes
the problem, host 2 is ready and is now free to communicate with another
host, say host3. Host1 did not get the ACK from host2 and sends the last
package again. Host2 sees this package is not from host3 and throws it away.

See the problem? What is the common way to solve this? 
Some solutions:
 1) Host2 could wait some time before communicating again.
 2) Host1 could give up after sending the lastpackage say 5 times. 




Greetings,
Maarten van den Hoek

---
maartenh@stack.urc.tue.nl
M.J.v.d.Hoek@stud.tue.nl
vdhoek@eb.ele.tue.nl


-----------[000527][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 22:22:56 GMT
From:      fromhell@u.washington.edu (Danny Armstrong)
To:        comp.protocols.tcp-ip
Subject:   TCP/IP programming info

Hey, I'm new to this group and I'm interested in where I can find 
programming notes on TCP/IP. Source, lib docs, libs, etc. Also, is there 
a newsgroup that concentrates on network programming, or does this one? 
I'm going to do something that involves networking sims, but I don't have 
any experience with networking, and anyone who knows the pitfalls... I 
have an idea what I'm going to run into, but I don't know the best way to 
deal with it- for instance, synchronizing separate platforms- how often 
does this need to happen? Any response would be appreciated. Thanx.

-----------[000528][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 94 22:35:49 GMT
From:      ddl@harvard.edu (Dan Lanciani)
To:        comp.protocols.tcp-ip,comp.unix.solaris
Subject:   Re: dup2() before connect() won't work on SunOS5.3

In article <3d8c5j$80q@pipe2.pipeline.com>, aray@pipeline.com (Arjun Ray) writes:

[...]
| I suggest a paranoid approach using the following sequence of calls:
| 
| 	shutdown( old_sock, 2 ) ; /* --[ forces buffer flushes ]--*/
| 	close( old_sock ) ; /* --[ make sure the old STREAM is dismantled ]-- */
| 	new_sock = socket( AF_INET, SOCK_STREAM, 0 ) ;
| 	connect( old_sock, (struct sockaddr*) address, sizeof(struct sockaddr));
                 ^^^^^^^^
I assume you meant new_sock here.

| 	dup2( new_sock, old_sock ) ;
| 	close( new_sock ) ;

You really don't want to do it quite this way.  Remember, new_sock could
be equal to old_sock (since you close old_sock first).  Even if dup2()
gets this right (i.e., does nothing)--and it might not in some
implementations--you will then close your one socket.  At the very
least you would have to surround the dup2/close with an
if(new_sock != old_sock) to avoid this problem.

				Dan Lanciani
				ddl@harvard.*

-----------[000529][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 22:50:13 GMT
From:      sd03@roger.gte.com (Shuang Deng)
To:        comp.protocols.tcp-ip
Subject:   Contents of IGMP for Multicast Control?

RFC-1054 describes the format of IGMP packet for reporting
a host's multicast memebership to multicast routers as 
consisting of version=1, type=report, and the address of
the group the host belongs to.

I understand this can be used to notify the router
to join a multicast group. 

But I cannot figure out the packet content for leaving
a group. Could some kind, wise soul tell me what the 
content of the IGMP packet for leaving a particular 
group is?

Thanks.



-----------[000530][next][prev][last][first]----------------------------------------------------
Date:      Wed, 21 Dec 1994 22:52:40 GMT
From:      emartini@netcom.com (Ed Martini)
To:        comp.protocols.tcp-ip
Subject:   Source code from Stevens' book

I'm looking for an FTP site for the source code from Comer/Stevens
Internetworking with TCP/IP Volume III.  Any pointers would be appreciated.

Ed Martini
Senior Software Engineer
Digital Video Systems
Sunnyvale, CA

-----------[000531][next][prev][last][first]----------------------------------------------------
Date:      Wed, 21 Dec 1994 23:16:40 GMT
From:      sunilg@forge.Tandem.COM (gaitonde_sunil)
To:        comp.protocols.tcp-ip
Subject:   Path MTU RFC Implementation

I am looking for a free implementation of the PATH MTU RFC (1191).
The 4.4 BSD Lite code I have access to only has the router part (i.e.
the correct PATH MTU specified in the "datagram too big" message.).

Can anyone point me to the right ftp site (etc.) where I can find
this code?

Please use my e-mail address, if possible, instead of posting here.

Sunil


-----------[000532][next][prev][last][first]----------------------------------------------------
Date:      21 Dec 1994 23:21:01 GMT
From:      ramseyea@drake.cig.mot.com (Erica A. Ramsey)
To:        comp.protocols.tcp-ip
Subject:   Telnetd protocol


Hello, i'm looking for the telnetd protocol. I want to write a program to 
communicate with a terminal server and send commands. The 
Current, I use telnet program to communicate with the terminal
server, but I need to embard to connection to the terminal
server within a C program. Therefore, I think that it will be
best if I used the telnetd protocol.

I will take anything I can get, ftp address etc...

thanks,
erica

-----------[000533][next][prev][last][first]----------------------------------------------------
Date:      Thu, 22 Dec 1994 08:28:28 -0500
From:      yue@cc.swarthmore.edu (Frank Yue)
To:        comp.protocols.tcp-ip
Subject:   Gatorstars and IP allocations (How do _you_ do it?)

Long subject and maybe long discussion..

Background:
Swarthmore College is a Macintosh dominated network with localtalk routers
(Gatorstar-GX) connecting the dorms to our ethernet backbone.  The
Gatorstars have the ability to defend a block of IP addresses and
dynamically allocated them to the localtalk nodes.  They also have the
ability to defend a block of static addresses for localtalk nodes.  This
combination of dynamic and static addresses has a maximum limit of 64.

Current situation:
Presently, we have the Gatorstars defending 64 dynamic addresses.  The
problem is that there are some students that are setting their MacTCP
control panel to take one of the addresses and use it as a static one
(Manual vs. server in the control panel) without our approval.  They are
doing this because they are running  WWW servers, FTP servers or who knows
what.  We do not want to totally prevent the students from
researching/working with these options. We also do not want to disable our
network with the high traffic some of these servers can produce.  Lastly,
we do not want these students to be taking unauthorized 'static' addresses
from this dynamic pool because if it is done improperly (taking one of the
lower addresses which the Gatorstar tries to dynamically assign) it can
disable the Gatorstar and that segment of the network.

Questions:
How have other institutions dealt with the above, or a similar situation? 
What did you do to alleviate this possible snafu?

How have other institutions dealt with the unauthorized use of IP
addresses?  How do you protect your network from such abuse?

If we establish a small pool of static addresses on each Gatorstar, what
criteria should be used in the distribution of the addresses to the
students?

Points to note:
It is too expensive at this current time to upgrade all the dorms to ethernet.
We have 69 Gatorstars.  Replacing all the hardware is prohibitively expensive.
Currently, only one student has permission to use a static address.
Even though we hand out locked versions of MacTCP, it is readily available
unlocked.

Please send responses to this group and email me a copy so I can compile
information which I can repost at a later time.  Thanks for your time and
effort in advance.
-- 
Frank Yue
Swarthmore College Computing Center
yue@cc.swarthmore.edu

"I didn't do it."  -Bart Simpson et. al.

-----------[000534][next][prev][last][first]----------------------------------------------------
Date:      Thu, 22 Dec 1994 10:32:55 -0500
From:      wtk@panix.com (William Kelley)
To:        comp.protocols.tcp-ip,comp.sys.mac.comm
Subject:   Re: IP over localtalk/AppleTalk locally, which MacTCP extension??

In article <ko-2212941157270001@komac.knoware.nl>, ko@komac.knoware.nl
(kS) wrote:

> Appletalk extension for MacTCP wanted
> ( NOT appletalk over Internet)

  [stuff deleted]

I'm sure this is all doable without any extra extension, since it is
possible to do it on one machine. I run a WWW server and a WWW client on
one Mac and "connect" to the server by giving an IP number to the client.
It seems perfectly obvious that I could have the server on one machine and
the client on another, connect them via LocalTalk or Ethernet, set IP
numbers and then connect, but it doesn't seem to work easily.

Let me know what you find out, please.

Thanks,
William

-- 
William Kelley
email: wtk@panix.com

-----------[000535][next][prev][last][first]----------------------------------------------------
Date:      Wed, 21 Dec 1994 14:36:22 +0700 (GMT+0700)
From:      Juliet Lim <juliet@emailhost>
To:        comp.protocols.tcp-ip
Subject:   netware as bootp


Hello!

I am looking for a utility to make a netware server as bootp server.  any 
help would be appreciated.  please e-mail to 
harianto@emailhost.ait.ac.th.  please, please send reply by e-mail as i 
am not a regular reader of this group.  thank you

juliet


-----------[000536][next][prev][last][first]----------------------------------------------------
Date:      22 Dec 1994 02:41:20 GMT
From:      djhan@eews23.kaist.ac.kr ( Doo-Jin Han )
To:        comp.protocols.tcp-ip
Subject:   NDIS driver-driver (help!!!)

I am a student, and i have worked PC communications.

I can get NDIS driver source, but I can not get working it. 

I need NDIS driver driver that can make working NDIS driver.

Without sorts of NDIS driver, it must make working NDIS driver.

If anyone has NDIS driver driver source, please send me it.

Help me everyone ,

Thanks you.

-----------[000537][next][prev][last][first]----------------------------------------------------
Date:      Wed, 21 Dec 94 14:21:57 MYT
From:      fhlee@csam.com.MY (Lee Fook Heng)
To:        comp.protocols.tcp-ip,comp.sys.sun.admin
Subject:   [Q]: What is MX entry in dns.

Hi,

Could anyone please explain what is the meaning of MX entry in
the in.named configuration file.

For example, in my domain's server named configuration file

.....
.....
.....
goofy	IN	A	199.1.1.10
	IN	MX	10 sub-domain.domain.com.


I was told that it is use for mail delivery purposes. 

Thanks for any response.

-- 
   Lee, Fook Heng
   System Engineer
   CSA Malaysia

   Tel  : +603-7587878 (ext. 320)
   Fax  : +603-7587382
   email: fhlee@csam.com.my


-----------[000538][next][prev][last][first]----------------------------------------------------
Date:      Thu, 22 Dec 1994 09:43:48 UNDEFINED
From:      mackinto@biologysx.lan.nrc.ca (David Mackintosh)
To:        comp.sys.novell,comp.sys.protocols.tcpip,,alt.winsock
Subject:   Re: Packet frame interference ?

In article <3d47n7$6r@hecate.umd.edu> ceham@w3eax.umd.edu (Maurice De Vidts NE3S) writes:
>I have setup Trumpet winsock using packet drivers ( -n )
>on a Netware 3.11 over our company LAN.  It seems to work
>but our network administrator has prohibited me from using it
>since he claims it crashes the server" 

The problem, at least here in my shop, is that TCP apps are very agressive, 
causing the server and other terminals to watchdog out because of excessive 
collisions during watchdog attempts.  It is less of a problem on ethernet -- 
but we have several times brought an ARCnet network completely to its knees.  
We don't have a fix.  Oh, we have also noticed the same problem using the 
Novell Winsock -- more on that later.

>I know I may be using a different frame type, but it seems
>a little far fetched that different ethernet frames will
>cause such a crisis. 

Bandwidth is bandwidth.  802.3 stations ignore the Ethernet_II frames -- but 
they still use the same wires.  Our problem is compounded by the fact that we 
have chatty Macs doing Ethertalk (Ethernet_Snap) on the same wires.

>I would like some techincal pointers to understand
>if I am truly creating a problem.  The novell winsock
>driver he has installed does not work with any
>of the winsock telnets I have found on the net. 

I know this one!  I know this one!  Your administrator has installed either 
LanWorkPlace or LanWorkGroup, right?  Either way, you are probably not loading 
the TCPIP stack.  If you are running LanWorkPlace, you will need to change 
your NET.CFG file to include a PROTOCOL TCPIP section.  Then you will have to 
run a file called TCPIP.EXE, probably located in a directory called XLN\BIN40 
(if it is on your local hard drive).  If you are running LanWorkGroup, you 
have to get someone to run a program called WGSETUP.EXE, and you have to make 
sure your user account is a member of the group LANWORKGROUP.  Your 
administrator should know what this is about, or know where to find out.  Once 
you are setup with LanWorkGroup, you will have access to the applications 
which LanWorkGroup comes with -- and once you remove your old Winsock, your 
new winsock applications should run fine.

We just installed LanWorkGroup on our server and went through all this.  Your 
administrator is probably running the LanWorkGroup, as the IP assignments are 
centrally controlled.

If you want more details, drop me a line.  I may be able to supply you with 
some documentation I wrote.



-----------[000539][next][prev][last][first]----------------------------------------------------
Date:      22 Dec 1994 05:01:43 GMT
From:      Dave Ung <ung@gcrex2.trt.allied.com>
To:        comp.protocols.tcp-ip,comp.unix.admin
Subject:   Re: Multiple IP addresses for one network port

Try PROXY ARP.  We stumbled upon it while getting PPP working between
a Sun sparc and a PC.
> I've heard a number of people mention that some flavors of Unix allow
> multiple IP addresses to be associated with one physical network port.
> I understand Linux and BSD/OS have this capability.  A chap recently
> mentioned there is a way to make SunOS 4.1.x and Solaris 2.x perform
> this nifty feat but so far he won't part with the details.  I'm 
> particularly interested in doing this on a SunOS 4.1.3 systems.  Any
> ideas or helpful pointers would be appreciated.
> 
> bob
> 
> -- 
> Bob Baggerman                    ! bob.baggerman@gtri.gatech.edu
> GTRI/ITTL/CND                    ! bob@comlab.gtri.gatech.edu
> Georgia Tech Research Institute  ! 404-894-7100 or 404-894-3525
> Atlanta, GA  30332  USA          !


-----------[000540][next][prev][last][first]----------------------------------------------------
Date:      22 Dec 1994 19:52:33 -0800
From:      aather@pyrtech.mis.pyramid.com (Amer Ather)
To:        comp.protocols.tcp-ip
Subject:   How transport provider handles out_of_band data


Problem description:

We have Sybase application Power Builder that runs over Lan Workplace for
Dos 4.2. Sybase SQL server 4.9 runs on Pyramid platform(UNIX SVR4).

The situation is as follow:
A query that returns for instance 1500 rows from server if aborted by
using dbcancel generates out of band message. Since out of band message is
handled by Transport provider first, therefore server should discard any data
recieve after out of band message related to query.

When we started looking at this problem what we found that Lan Workplace for
dos doesn't support out of band data, and that causing transport provider on
PC ignores out of band data request, and treats it as normal band data.
Which causes server on other side send whole result back to PC.

When we contacted Novell they recommended to apply patch lwp42T to fix this
problem. Now we see less data come back from server after using dbcancel.
We compared results with other Operating systems such as window NT and SUN
Sparc, and found that still out of band message from LAN WORKPLACE is not
handled as fast as it handles on NT and SUN when using same server. Having
late acknowledgment of out of band data causes unnecesary traffic on network,
and slower response from later queries.

We checked sniffer output and it does show that out of band message was sent
a head of normal messages, but still PC recieves over 60-70k of data from
server.

I like to know why server returns more data after dbcancel for PC than
NT or Sparc. Is there any kernel parameter available that can reslove this
problem.

Appreciate for any kind of assistance.

Amer..

-----------[000541][next][prev][last][first]----------------------------------------------------
Date:      Thu, 22 Dec 1994 06:39:58 GMT
From:      vjs@calcite.rhyolite.com (Vernon Schryver)
To:        comp.protocols.tcp-ip
Subject:   Re: getting ip_addr from FD

In article <3da3rf$72l@newhub.xylogics.com> carlson@xylogics.com writes:
>In article <3da11d$2333@tequesta.gate.net>, scheidel@gate.net (Michael S. Scheidell) writes:
 
> ...
>|> after telnet (or rlogin) is there an EASY or hard way to find the host 
>|> address of the remote host FROM  user PROGRAM, with the FD of stdin?
>|> or with FD of stdin, get the socket associated with pseudo-device..
>
>Ok, then a rephrased answer:
>
>No, there is no EASY way to do this.
> ...

It all depends on the system and exactly what is required (i.e. whether
FD 0 is necessarily relevant).  Rlogind, telnetd, and rshd in some
commercial operating UNIX systems using 4.*BSD TCP/IP put the remote
system name (or IP address if the name is unknown) into environment
variables with names like "REMOTEHOST".  One in particular does that
because in the distant past it was blessed with an SVR0 style utmp that
did not have a slot for the remote info.

(Yes, it is not entirely secure except in some cases such as the first
shell spawned by the network daemon.)


Vernon Schryver    vjs@rhyolite.com

-----------[000542][next][prev][last][first]----------------------------------------------------
Date:      Thu, 22 Dec 1994 08:03:25 GMT
From:      jsasaki@orion.kawacc.mocoil.co.jp (Josiah Sasaki)
To:        comp.protocols.tcp-ip
Subject:   [Q] free inplementation of DHCP


Does anyone know the free inplementation of DHCP?


--------------------------------------------------
Josiah Sasaki
Systems Planning Group
Information Systems Department
Mitsubishi Oil Co.,Ltd.

E-mail: jsasaki@kawacc.mocoil.co.jp
--
-------------------------------------------------------
佐々木 由也

川崎市川崎区扇町12ー1 三菱石油(株)情報システム部
tel(044)344-7819
fax(044)344-7857
E-mail:jsasaki@kawacc.mocoil.co.jp

-----------[000543][next][prev][last][first]----------------------------------------------------
Date:      Thu, 22 Dec 1994 08:32:04 +0000
From:      Andrew@cucumber.demon.co.uk (Andrew Gabriel)
To:        comp.protocols.tcp-ip
Subject:   Re: Assymetric Bandwidth

In article <1994Dec20.153656@prospero.dev.cdx.mot.com>
           sunil@prospero.dev.cdx.mot.com "Sunil Menon" writes:

> 
> If the Bandwidth in the two directions is Assymetric i.e is unequal, does it
>  cause any problems in TCP/IP? Especially at the TCP layer. Does anybody have
>  any info, experience, suggestions, pointers etc?
> 

No experience, but I was recently informed (on uk.telecom) that Internet
access over cable TV is usually very asymetric, with a few channels being
combined to give ethernet type throughput for receive (which can therefore
be broadcast over a sizable area), but transmit using a modem over standard
telephone twisted pair. (Transmit and Receive as viewed by the cable
subscriber).
-- 
Andrew Gabriel                        Home: Andrew@cucumber.demon.co.uk
Consultant Software Engineer          Work: Andrew.Gabriel@gpt.co.uk

-----------[000544][next][prev][last][first]----------------------------------------------------
Date:      Thu, 22 Dec 1994 10:57:27 GMT
From:      ko@komac.knoware.nl (kS)
To:        comp.protocols.tcp-ip,comp.sys.mac.comm
Subject:   IP over localtalk/AppleTalk locally, which MacTCP extension??

Appletalk extension for MacTCP wanted
( NOT appletalk over Internet)

Please read carefully

Setup;

Powerbook , MacIIfx, Printer , all 3 connected by local talk network hardware.
TO BE CLEAR NO modem No ethernet No Internet Just these 3 pieces off hardware.

How can I tell MacTCP to use Appletalk/localtalk to let the 2 Mac's
communicate IP (beit tcp or udp and more) , IP means InternetProtocole NOT
that I am connected to the Internet. Selecting localtalk doesn't work

Each Mac has it's own fixed IP address (set manually under MacTCP cntr
panel more... button)


I am not looking for Gator-something they are for routing to the internet
network nor for AIR (Apple Internet Routing) BTW they all have dynamic IP
adressing and IP packets are send over localtalk besides Appletalk packets


Then I could do MacPING to the other Mac BTW Why can't I IP ping to myself?
Then I could start a telnet connection ; one telnet app listining one
telnet app requesting connection. (Versaterm PRO can, though it uses it's
own trick and exactly that trick I want to use UNDER MacTCP)

I have this other App that only uses IP (thus MacTCP) it has no other
connection-tools.  I want to run this Application on my 2 macs




Technicaly IMO this extension should do;

1. Put up an IPttAppletalk service (tt = tunnel through)
   pack and unpack IPpackets in/out Appletalkpackets 
2. Scan for other IPttAppletalk services
3. When started broadcast it's IP address and Appletalk address
4. Keep a table off IPaddresses and Appletalk adresses

etc.

Someone with the (near) right answer please email me

Regards Ko

-- 
V2S
Holland

-----------[000545][next][prev][last][first]----------------------------------------------------
Date:      22 Dec 1994 11:22:25 GMT
From:      Ray Harder <rharder@ctp.org>
To:        comp.protocols.tcp-ip
Subject:   Kermit with MS TCP/IP-32?

Is it possible to use Kermit with Windows For Workgroups 3.11 and
TCP/IP-32 from Microsoft? I installed Kermit 3.13 on a laptop with a
PCMCIA card for 10BaseT Ethernet access. If I load Kermit under DOS (6.2)
it works beautifully. If I load windows first and then load Kermit it
works fine also -until I try to access the server using TCP/IP. At the
Kermit prompt, I type "Set Port TCP/IP 163.xxx.xxx.xxx" (With the actual
IP address of course) and then on the next line I type "Connect."

Sometimes the connection is made for a few seconds. I have even been able
to login once or twice, but always within a screen or two, it crashes
--taking windows down with it. I cannot access it from Kermit within
Windows when using TCP/IP, but from DOS it works fine. There seems to be
a conflict between the built-in TCP and the 32 bit version. Adjusting the
usual settings does not seem to have any effect. In the Docs for Kermit
it mentions that it can be used in a couple of others including Beam and
Whiteside. Can Kermit be modified to use an existing TCP/IP stack?

Alternately, is there a way to get VT220 (Or better V320) emulation over
a TCP/IP connection from Windows using Public Domain software? Is there a
version of Telnet that allows V220 or 320 for a Windows environment?

***********************************************************************
****
* Raymond G. Harder                  "Can't walk today, I don't feel
well."*
* Educational Technology Consultant  "Why don't you sit out in the sun?" 
 *
* 909-983-4713                       "What? People will think I'm lazy!" 
 *
* rharder@ctp.org                         -- My 92 year old grandmother  
 *
***********************************************************************
****

-----------[000546][next][prev][last][first]----------------------------------------------------
Date:      22 Dec 1994 22:19:09 -0600
From:      vkhare@tiger.lsu.edu (Vikram Kumar Khare)
To:        comp.protocols.tcp-ip
Subject:   fsp anyone?



	Need some fsp.  Anyone got a ftp site?



-----------[000547][next][prev][last][first]----------------------------------------------------
Date:      Thu, 22 Dec 1994 17:22:40 UNDEFINED
From:      timmi@ftp.com (Timothy G. Reynolds)
To:        comp.protocols.tcp-ip
Subject:   Re: What is "TCP Slow-Start" ?

In article <amizuno-221294110413@aki-powerbook.stanford.edu> amizuno@jessica.stanford.edu (Kuniaki Mizuno) writes:
>From: amizuno@jessica.stanford.edu (Kuniaki Mizuno)
>Subject: What is "TCP Slow-Start" ?
>Date: Thu, 22 Dec 1994 11:04:13 -0800
 
>Hi Everyone ,
 
> I have interest about "TCP Slow-Start". 
>Please tell me some URLs, papers,documents etc.
 
>Actually, I need them soon. So I will be happy if you
>tell them by mail. Give me Cristmas presents.

According to _Internetworking with TCP/IP_ by Douglas Comer:

Slow-start (soft-start) Recovery: Whenever starting traffic on a new 
connection or increasing traffic after a period of congestion, start the 
congestion window at the size of a single segment and increase the congestions 
window by one segment each time an acknowledgement arrives.

You should read RFC896: Congestion Control in IP/TCP Networks by John Nagle.

Timothy G. Reynolds					FTP Software, Inc.
Technical Services					2 High St.
timmi+email@ftp.com					N. Andover, MA
(800) 382-4FTP							01845

	    "What did Cinderella do after she became queen?
	      What does 'Happily Ever After' really mean?"

-----------[000548][next][prev][last][first]----------------------------------------------------
Date:      22 Dec 1994 14:31:10 GMT
From:      spidey@rtp.vnet.ibm.com
To:        comp.protocols.tcp-ip
Subject:   Re: help in writing multiple client server

In <3d6clq$a4b@nuscc.nus.sg>, eng10412@leonis.nus.sg (LEW CHAI SECK) writes:
>I need to write a simple server programme that can be used to serve 
>mulitple client but not sure about how to go about doing it....
>
>Is there any programme available that I can use to follow or is that any 
>good books on this subject...
>
>Any help will be greatly appreciated....
>
>--
>==============================================================
>Lew Chai Seck
>National Unversity of Singapore
>Email : eng10412@leonis.nus.sg

Take a look at "Power Programming With RPC"
it is an OReilly book Author is John Bloomer.

Jim
--- When the snake and the mongoose meet the mongoose prevails
---     For it is the jaws with which we speak 
---     not the venom with which we say it
---     that determines our strength!




-----------[000549][next][prev][last][first]----------------------------------------------------
Date:      22 Dec 1994 14:39:58 GMT
From:      mpeter@ee.ethz.ch (Martin Peter)
To:        comp.protocols.tcp-ip
Subject:   communication with a terminalserver

Hello,
I have a terminal server connected to an tcp/ip ethernet and I would
like to communicate to a device connected to a serial port of the
terminal server. 
I can connect the terminal server over 'telnet', but I see no easy way
to do that from a programm. 
Do I really have to implement something like a telnet client, so that
I can read and write characters to the serial device ?

Thanks for any help

--
Martin Peter, ETH Zuerich             e-mail: mpeter@ee.ethz.ch
Computing Support Group                   tel:    41 1 632 5286
ISG Gloriastrasse 35                      fax:    41 1 632 1194
8092 Zurich/ Switzerland

-----------[000550][next][prev][last][first]----------------------------------------------------
Date:      Thu, 22 Dec 1994 15:47:45 +0000
From:      peter@bhcnt.demon.co.uk (Peter Gross)
To:        comp.protocols.tcp-ip
Subject:   Re: Network design and requirements analysis

ngai@delphi.com wrote:

<dirty orrible advert deleted>

Please refer back to your nettiquette guide !!!!

-- 
The ideas, suggestions, jokes, cautions, complaints, comments, scripts, insults,
critiscms, observations, diagrams, sarcastic remarks, stories, programs, and
thoughts shown in the text above are mine and mine alone, and are NOT those of
my employer or anyone else for that matter (unless I say otherwise ;-)

-----------[000551][next][prev][last][first]----------------------------------------------------
Date:      22 Dec 1994 18:14:00 GMT
From:      hutchin@sunset.cs.concordia.ca (Fraser Hutchinson)
To:        comp.unix.unixware,comp.protocols.tcp-ip
Subject:   Re: [Q]Slip Service

In article <3dadsn$da3@jadzia.CSOS.ORST.EDU> Steve Hathaway <steve@oem.state.or.us> writes:

   In article <19941215.113819.357232.NETNEWS@WVNVM.WVNET.EDU>, 
   <lfa@olie.wvitcoe.wvnet.edu> writes:
   > 
   > I am looking for advice, pointers, experiences, etc in setting up a slip
   > server.  There may be 400+ potential users with a maximum of 15 users at
   > a time.
   > 
   > TAI
   > Larry
   > 
 ====================
>   AT&T Unix SVR4.0
>     I have had no luck getting SLIP to work over a dial-in/dial-out line.
>     The implementation of SLIP works over a dedicated RS-232 line.
 
>   I do not know if Unixware supports dial-in/dial-out SLIP.

No, it doesn't. SLIP works fine on my Unixware box as my connection
to my Internet Provider, but SLIP-on-demand is not supported.

However, PPP handles this, if you do not need dynamic address allocation.
If you do, there is a product I have heard good things about called 
Morningstart PPP that provides dynamic address allocation.

Fraser
--
-------------------------------------------------------------------------
Fraser Hutchinson                                hutchin@cs.concordia.ca
#include <std/disclaimer.h>                    Godot: Sorry I'm late.....
-------------------------------------------------------------------------

-----------[000552][next][prev][last][first]----------------------------------------------------
Date:      Thu, 22 Dec 1994 19:43:35 GMT
From:      jhpb@sarto.gaithersburg.md.us (Joseph H. Buehler)
To:        comp.protocols.tcp-ip,comp.unix.questions,comp.unix.large,comp.unix.sys5.r4
Subject:   Re: Printing to JetDirect from drs6000

>>>>> "gal" == general admin login <sysmgr@vacgen.fisons.co.uk> writes:

    gal> We have a HP plotter with a JetDirect card on the same
    gal> ethernet/tcp/ip network as our drs6000. I would like to spool
    gal> print requests for the plotter on the drs6000 using the lp
    gal> print service.  We are still running drsnx v6 and icl say
    gal> this cannot be done without upgrading to drsnx v7. Anyone got
    gal> any ideas?

Write a PERL script to connect to the proper port on the plotter and
write the job to the plotter.  I don't know if the JetDirect card
requires the lpd protocol; if it has a "raw stream of bytes" mode, the
PERL script will probably do quite nicely.  You have to integrate the
script into the lp system too, of course, but that's usually not too
difficult.

Joe Buehler

-----------[000553][next][prev][last][first]----------------------------------------------------
Date:      Thu, 22 Dec 1994 19:51:45 GMT
From:      hcohen@world.std.com (Howard D Cohen)
To:        comp.protocols.tcp-ip
Subject:   get duplicate pings back


I have a Sun Classic running Solaris 2.3.  It has two ethernet connections, 
one attached to our internal 192... net, the other attached to our internal
204... net.  204.176.164.1 is our router to an Internet server. The routing 
is configured as follows(routed is not running):

#netstat -nr
Routing Table:
  Destination           Gateway           Flags  Ref   Use   Interface
-------------------- -------------------- ----- ----- ------ ---------
127.0.0.1            127.0.0.1             UH       0   7710  lo0
192.9.200.0          192.9.200.215         U        2   1007  le1
204.176.164.0        204.176.164.2         U        3 165313  le0
224.0.0.0            204.176.164.2         U        3      0  le0
default              204.176.164.1         UG       0  10038  

When I ping internal systems on 192... or 204... or an external system, I get
duplicate packets back:

#ping -s sun1
64 bytes from sun1 (192.9.200.1): icmp_seq=0. time=4. ms
64 bytes from sun1 (192.9.200.1): icmp_seq=0. time=10. ms
64 bytes from sun1 (192.9.200.1): icmp_seq=1. time=2. ms
64 bytes from sun1 (192.9.200.1): icmp_seq=1. time=3. ms
64 bytes from sun1 (192.9.200.1): icmp_seq=2. time=2. ms
64 bytes from sun1 (192.9.200.1): icmp_seq=2. time=4. ms
64 bytes from sun1 (192.9.200.1): icmp_seq=3. time=2. ms
64 bytes from sun1 (192.9.200.1): icmp_seq=3. time=3. ms
----sun1 PING Statistics----
4 packets transmitted, 8 packets received, 2.00 times amplification
round-trip (ms)  min/avg/max = 2/3/10

#ping -s otto
64 bytes from otto (204.176.164.141): icmp_seq=0. time=4. ms
64 bytes from otto (204.176.164.141): icmp_seq=0. time=35. ms
64 bytes from otto (204.176.164.141): icmp_seq=1. time=1. ms
64 bytes from otto (204.176.164.141): icmp_seq=1. time=5. ms
64 bytes from otto (204.176.164.141): icmp_seq=2. time=1. ms
64 bytes from otto (204.176.164.141): icmp_seq=2. time=4. ms
64 bytes from otto (204.176.164.141): icmp_seq=3. time=1. ms
64 bytes from otto (204.176.164.141): icmp_seq=3. time=3. ms
...
----otto PING Statistics----
8 packets transmitted, 16 packets received, 2.00 times amplification
round-trip (ms)  min/avg/max = 1/4/35

#ping -s world.std.com
64 bytes from world.std.com (192.74.137.5): icmp_seq=0. time=15. ms
64 bytes from world.std.com (192.74.137.5): icmp_seq=0. time=1194. ms
64 bytes from world.std.com (192.74.137.5): icmp_seq=1. time=198. ms
64 bytes from world.std.com (192.74.137.5): icmp_seq=1. time=200. ms
64 bytes from world.std.com (192.74.137.5): icmp_seq=2. time=8. ms
64 bytes from world.std.com (192.74.137.5): icmp_seq=2. time=10. ms
----world.std.com PING Statistics----
3 packets transmitted, 6 packets received, 2.00 times amplification
round-trip (ms)  min/avg/max = 8/270/1194

When I disconnect the 192... circuit from my system, I get no duplicate
responses, but cannot connect to the 198... network.  The same goes for 
disconnecting from the 198... circuit.  

The 192... network, the 198... network, and DECNET are all being transmitted
on the same ethernet wire.

What did I do wrong?  Why do I get the duplicate responses.

Thank you.

Howard Cohen 
CDA Division, Analogic Corp   8 Centennial Drive   Peabody, MA 01960
phone: 508-977-3000, fax: 508-977-6813   email: hcohen@analogic.com

-----------[000554][next][prev][last][first]----------------------------------------------------
Date:      22 Dec 1994 19:58:51 GMT
From:      Mike Chan <mchan@mail.edtel.alta.net>
To:        comp.protocols.tcp-ip
Subject:   HELP!! Ip Subnet

Hi Folks,	

I have some difficulties to subnet my network. Say if I have a class C network	
with ip address200.200.100.0.  If I use subnet Mask 255.255.255.252. The
mask bits are 6 and I can have 62 subnet, but, I can only have 2 hosts,
even though I have 2 bits to play with hosts. Instead of 4, I can only 
have 2 hosts.  I assume bits 00 and 11 would not be availlabl.e Can 
anyone tell me why ? Is it for broadcast or multicast ? Because of subnetting,
I may lose 2 x 63 = 124 ip addresses.  Is there any better solution besides
switch to 5 bit masking ?

Best Reggards, & Merry Xmas

Mike Chan
E-mail id: mchan1@agt.alta.net


-----------[000555][next][prev][last][first]----------------------------------------------------
Date:      Thu, 22 Dec 1994 20:04:05 GMT
From:      gfischer@gfischer.asd.tse.ca (Grant Fischer)
To:        comp.protocols.tcp-ip
Subject:   Should ICMP Echo generate ICMP Redirect?

I have a host that uses a default router, relying on ICMP redirects
from that router to learn of better routes to certain nets.


So, like this (where R1 is the default route):

        H1
        |
    |-------------------|
           |         |
           R1       R2
           |         |
       |------|   |-----|
                      |
                      H2


When H1 sends an ICMP Echo to H1, and H1 does not know
of the route to H2 yet, H1 sends the Echo Reply to R1
(it's default route).

Now the question: what should happen next?
A) R1 passes the Echo Reply on to R2.
B) R1 generates an ICMP Redirect

Is it some combination of A & B? I was expecting A+B, 
based on RFC 792 & 1122, but I wanted to ask the opinions of others.
One of our routers seems to do A, and another seems to do
neither.

I'd appreciate e-mail replies; posts might get aged off before
I can get to them over Christmas.


Grant Fischer   (gfischer@tse.com)
Technology and Network Services,  The Toronto Stock Exchange

-----------[000556][next][prev][last][first]----------------------------------------------------
Date:      Thu, 22 Dec 1994 20:24:07 GMT
From:      Frederic Montoya <Frederic.Montoya@loria.fr>
To:        comp.parallel,comp.protocols.tcp-ip
Subject:   Re: Massively parallel programming for the Internet

In article <loss.433.000AAA53@husky.bloomu.edu>, loss@husky.bloomu.edu (Doug Loss) writes:
|> Approved: rmuise@dragon.acadiau.ca
|> Lines: 22
|> Xref: loria.fr comp.parallel:10026 comp.protocols.tcp-ip:27920
|> 
|> In article <thinmanD0toH9.B94@netcom.com> thinman@netcom.com (Technically Sweet) writes:
|> 
|> >Are there any projects attempting to build a massively parallel
|> >programming environment using spare cycles on the Internet?
|> >Something where you put a really simple interpreter on the
|> >end machines, mid-level cluster managers spaced periodically,
|> >and end-user control stations?
|> 

This was the big dream .... recycling idle cycle....In the same way of
thinking why not using all scientific calculator sleeping in draws :)
I think all these attempts has demonstrated their limits lately.

|> This sounds reminiscent of Linda, a distributed parallel computing system I 
|> read about a while ago.  I'm sure there's material about it somewhere on the 

I think you mean piranhas which was built on top of LINDA

Frederic.

-- 

--

Frederic Montoya

INRIA Lorraine                       Voice : +33 83 59 30 84
LORIA                                Fax   : +33 83 27 83 19
615, rue du Jardin Botanique         E-Mail: montoya@loria.fr
F-54602 Villers les Nancy Cedex      W3    : http://www.loria.fr/~montoya
France


-----------[000557][next][prev][last][first]----------------------------------------------------
Date:      Thu, 22 Dec 1994 20:26:06 GMT
From:      jhgreen@eecs.nwu.edu (Jethro H. Greene)
To:        comp.parallel,comp.lang.misc,comp.protocols.tcp-ip
Subject:   Re: Massively parallel programming for the Internet

In article <thinmanD0toH9.B94@netcom.com>,
Technically Sweet <thinman@netcom.com> wrote:
>Are there any projects attempting to build a massively parallel
>programming environment using spare cycles on the Internet?
>Something where you put a really simple interpreter on the
>end machines, mid-level cluster managers spaced periodically,
>and end-user control stations?
>
>This was inspired by the project to crack large prime numbers by
>mailing large lists of numbers around to volunteers running the
>sieve program.  Such a thing could be entirely automated.
>David Gelernter's "Mirror Worlds" is clear articulation of this vision.
>
>It's apparent to me that the Internet is going to have to dramatically
>increase the sophistication of its internal "metabolic" processes;
>parallel computations which monitor traffic and alter routing tables
>will have to be done at some point.  An Internet Nervous System if you will.
>
>Social consequences: well, at some point, you may be required to
>run the interpreter as part of your Internet Tax.
>
>Back to tech talk: is anybody using multisets as a basis for
>parallel programming?  Multisets looked like a possibly pleasant
>paradigm for programming the distributed interpretive system.
>
>-- 
>
>Lance Norskog
>thinman@netcom.com
>Artisputtingtogether. Art  s th ow n  aw y.
>

Look at legion at http://cs.virginia.edu

-- 
Jethro H. Greene                                      jhgreen@eecs.nwu.edu
Computational Electromagnetics Laboratory             office: 708 491-8887
Northwestern University                               fax:    708 467-3217


-----------[000558][next][prev][last][first]----------------------------------------------------
Date:      22 Dec 1994 22:07:30 GMT
From:      freeman@mr.net (Alex Li)
To:        comp.protocols.tcp-ip
Subject:   HELP! NDIS drivers and 2 DEC cards in one box

Hi,

Has anyone here configured 2 DEC EtherWorks 3 cards using the NDIS2.01 
driver (EtherWorks driver v3.00) in one PC?  My current config. will load 
PROTMAN.DOS and EWRK3.DOS ok but it sees only ONE card.  I followed
the sample(protocol.ini) in the readme file but it doesn't help.

Thanks for any help.  Merry Christmas!

Alex Li

-----------[000559][next][prev][last][first]----------------------------------------------------
Date:      Fri, 23 Dec 1994 00:26:59 GMT
From:      perec@freenet.fsu.edu (Pere Camps)
To:        comp.protocols.tcp-ip
Subject:   Re: Urgent: SMTP Server for DOS

In article <3ctpe4$l7t@eis.calstate.edu>, shunt@eis.calstate.edu (Steve J. Hunt) says:

>> I'm looking desperatly for a DOS-TSR based SMTP server program.
>> 
>> We have no NetWare server, so we can't use pegasus/mercury. :(

First of all, I'm the same guy who originally posted the message... :)

>How about running Linux on a PC?  That would do what you want.  And 
>Linux is pretty cheap.

    We have considered this... however the PC is running a BBS and it
would be very complicated to have run over Linux.

>  Pegasus doesn't require a Netware server 
>anymore, but it can't act as an "SMTP server" anyway.

    So no luck...

>I don't think there are any DOS or Windows based sendmail type 
>programs.

	That's a pity... do you know anything about PC/TCP software?
It might work...

	Salutacions,
	Pere.

-----------[000560][next][prev][last][first]----------------------------------------------------
Date:      23 Dec 1994 01:46:43 -0000
From:      Simon <sb91@ecs.soton.ac.uk>
To:        comp.protocols.tcp-ip
Subject:   Re: HELP! My first TCP/IP server program doesn't respond :^(

In <3cu4vg$ols@hprcl192.mayfield.hp.com> colin@mayfield.hp.com (Colin Wynd) writes:

>: It's basically just a daytime server called "TCPdaytimed" that uses
>: argv[1] to determine what port to listen to.  I added the line:
>:  
>:         masterq   256/tcp

I've not seen the original post yet (our newsfeed has been a little flacky here
recently), but are you running as root ? Any port number between 1 and 1023 is
reserved and you have to be running as root to bind() to them. Try with a
different port number which is higher than 1023. Also, add error checking to
your program's code, ie :

#include <errno.h>

  if (bind(....)<0) {
    perror("bind() failed");
    exit(1); }

--
Simon
EMail: simes@tcp.co.uk

-----------[000561][next][prev][last][first]----------------------------------------------------
Date:      23 Dec 94 10:21:37 EST
From:      jkerr@wvnvms.wvnet.edu (Jason Kerr)
To:        comp.databases.oracle,comp.client-server,comp.protocols.tcp-ip
Subject:   Help with testing methodology

We are going to be doing some testing over the next few weeks and I would like
to be able to gather some quantitative evidence to support a theory.

Our testing is going to involve the transmition of snapshot refreshed tables
generated by Oracle v7.1.3 using SQL*NET v1.2.  The requests for the data will
be made from a remote site using tcp-ip protcol accross a T1 network.  Being a
DBA and not extremely versed in network diagnotic techniques, could someone
offer a suguestion as to way to quantify the following:

1.  number of packets per request
2.  size of transmitted packets
3.  percent of bandwidth utilization by request
4.  elapsed time between request and receipt of data

Has anyone done this sort of testing?  What kind of results have you observed?

Also, if this goes well, we would like to reproduce the same data on requests to
remote sites generated directly from SQL Forms.

Any input would be greatly welcomed.

Thanks and have a happy holiday.



-----------[000562][next][prev][last][first]----------------------------------------------------
Date:      Fri, 23 Dec 1994 06:04:56 GMT
From:      act9m@viper.cs.Virginia.EDU (Alan Chih-Chang Tai)
To:        comp.protocols.tcp-ip
Subject:   Timer Source

I'm trying to implement a timer for a new protocol on top of UDP;
unfortunately, I have not been able to find any sources to go by.  I
am envisioning a process that constantly calls gettimeofday, but I
know that ain't the solution.  The timer has to have at least 10 ms
accuracy.  Can someone point me to any resources?  Thanks.


Alan
--
Alan Tai (act9m@virginia.edu) | Computer Science Dept. |   home: 804-977-7854
Graduate Teaching Assistant   | University of Virginia | office: 804-982-2294

-----------[000563][next][prev][last][first]----------------------------------------------------
Date:      23 Dec 1994 08:06:48 GMT
From:      kfc@wimsey.com (Ken Cunningham)
To:        comp.protocols.tcp-ip,comp.sys.mac.comm
Subject:   Re: IP over localtalk/AppleTalk locally, which MacTCP extension??

In article <wtk-2212941032550001@wtk.dialup.access.net>, wtk@panix.com
(William Kelley) wrote:

> In article <ko-2212941157270001@komac.knoware.nl>, ko@komac.knoware.nl
> (kS) wrote:
> 
> > Appletalk extension for MacTCP wanted
> 
>   [stuff deleted]
> 
> I'm sure this is all doable without any extra extension.

I have three macs, all connected by localtalk.

MacTCP on each one, each one given a manual address.

Selected "LocalTalk" in each MacTCP configuration.

It all works fine. I can 'CU-SeeMe" between any two macs, run MacHTTP on
one and log into it from any other, and I tried "MailShare" and it works
fine too.

If you set up your hosts file correctly, you can use one as a DNR as well.

What kind of a problem are you having with this?

-----------[000564][next][prev][last][first]----------------------------------------------------
Date:      23 Dec 1994 08:15:39 GMT
From:      Kari.Hurtta@Helsinki.FI (Kari E. Hurtta)
To:        comp.mail.sendmail,comp.protocols.tcp-ip
Subject:   Re: Ping/Mail Bounces Here; But Not Elsewhere

In article <17093F4F9.JLITTLEF@UGA.CC.UGA.EDU> JLITTLEF@UGA.CC.UGA.EDU (jlittlef@uga.cc.uga.edu) wrote in comp.mail.sendmail:
<...>
» that we can't.  I have discovered that I can never ping a site that I
» cannot send mail to.  The ping command line just hangs there until I do a
» control-c to terminate it.  Does anybody on Peachnet have a sendmail.cf
» or another config file that I can show our sysad?  Any examples (doesn't
» have to be peachnet) are most definitely appreciated.  By no means am I
» a net guru but I am trying to assist our sysad in resolving this issue.
» If I had to guess, it just seems to me that the nameresolver is either not
» getting something from us or we aren't getting the right stuff from it
» since we can neither mail nor ping.

Perhaps network/router problem. If ping hangs then preblem have not do with
sendmail. Try traceroute to offended site.

--
- K E H                                      /  Elämä on monimutkaista
  Kari.Hurtta@Helsinki.FI

-----------[000565][next][prev][last][first]----------------------------------------------------
Date:      23 Dec 94 14:41:56
From:      glen@noon.midnight.com (Glen B. Glater)
To:        comp.protocols.tcp-ip
Subject:   Re: FTPD always uses port 20 on active open.  Why?

In article <3df12q$363@unix.sri.com> ric@coronacorp.com (Ric Steinberger) writes:

 >        Can anyone help with this?  I need to write an application
 >somewhat similar to ftp/ftpd.  The similarity is that there will
 >be one socket connection for (ASCII) control information and a second socket
 >(created after a connection is established on the first) for binary
 >data.
 >
 >        I have beed reading up on ftp/ftpd in Richard Steven's TCP/IP
 >Illustrated, Volume 1.  He explains that the client sends the server an IP
 >address/port combination and enters a passive listen on that port.  Then the
 >server is expected to do an active open on that IP/port pair.  It seems that
 >the server always uses the same port (20).  So I have two questions:
 >
 >1) Why always bind to port 20 instead of letting the system give you a port?
 >
 >2) I'm a little unclear on how the server can (in it's child processes)
 >   establish active opens on the same port (20 in this case).
 >   Doesn't this create a problem when ftp clients try to talk to
 >   "their" ftp server if "all" the ftp server child process have
 >   (what seems to me) the same endpoint address (IP/port pair)?

Reading in Steven's book (page 15), you are pointed to RFC 1340 (which 
has been obsoleted by RFC 1700) to determine the assignments of Internet
numbers.  Looking under the "well known ports" section, you find:

ftp-data         20/tcp    File Transfer [Default Data]
ftp-data         20/udp    File Transfer [Default Data]
ftp              21/tcp    File Transfer [Control]
ftp              21/udp    File Transfer [Control]

So, port 20 is assigned to FTP for opening data connections.  That's why
it always uses it.

If you have any questions, feel free to email me at glen@midnight.com.

--glen

--


...............................................................................
  Glen B. Glater    : Midnight Networks Inc. 200 Fifth Avenue Waltham MA 02154 
  glen@midnight.com : Vox 617/890-1001 Fax 0028  The Best in Network Software


-----------[000566][next][prev][last][first]----------------------------------------------------
Date:      Fri, 23 Dec 1994 10:26:07 GMT
From:      rossw@march.co.uk (Ross Wakelin)
To:        comp.protocols.tcp-ip,comp.unix.questions,comp.unix.large,comp.unix.sys5.r4
Subject:   Re: Printing to JetDirect from drs6000

general admin login (sysmgr@vacgen.fisons.co.uk) wrote:
: We have a HP plotter with a JetDirect card on the same ethernet/tcp/ip 
: network as our drs6000. I would like to spool print requests for the 
: plotter on the drs6000 using the lp print service. 
: We are still running drsnx v6 and icl say this cannot be done without 
: upgrading to drsnx v7. Anyone got any ideas?
 
: 	Thanks 
: 	
 
: ---------------------------------------
: Jim Cozens
 
: Fisons Instruments Vacuum Generators


what version of the Jetdirect card is it?  Does it support lpd, or is one
of the old ones with the direct port access.  Either way, it can be
done (cos we have done it for a client, using V6).  Contact me direct
and I will bring you up to date (you may need a fof from ICL)

Ross
 
-- 
Ross Wakelin                              r.wakelin@march.co.uk
Open Systems Director     
March Systems Consultancy Ltd 		+44 1734 304 224
or rossw@manuka.demon.co.uk at home	PGP signature available

-----------[000567][next][prev][last][first]----------------------------------------------------
Date:      Fri, 23 Dec 1994 12:01:17 GMT
From:      ko@komac.knoware.nl (kS)
To:        comp.protocols.tcp-ip,comp.sys.mac.comm
Subject:   Re: IP over localtalk/AppleTalk locally, which MacTCP extension??

Q: CAN I CONNECT USING tcp/ip TO ANOTHER Mac  WITHOUT INTERNET ACCESS?
A: programs that use tcp/ip or udp/ip don't need nescessarily access to
the internet . For instance 2 Versaterm PRO telnet users can connect via
modem or directly via  serial ports by using their respective ip client
software ie slip /ppp, no server involved here. In case of 2 Mac's in a
localtalk network you can also use  localtalk under the MacTCP panel. If
you have no IP number just invent one  for each participant .



Problem solved;

No extension needed

(..Connecting 2 Mac's over localtalk to videophone my wife upstairs..)

Finally someone helped me out! Thanks Robin

rdhw@cam.ac.uk (Robin Walker) wrote
>> >It would be helpful if you could say what is going wrong.
>>
>> Macping cannot find the other mac in the localtalk network
>>
>> >It should work.

YES YES IT WORKS

History (as you'v heard more) :)

I was misled by wrong and old information Hmmm....  

1. I didn't know system 7.5 has  "ip encapsulation in Appletalk" build in
2. I tested MacPING and thought the other side automatically responded to
a ping, wrong you need a PINGER on each of the 2 systems. I stopped from
there on experimenting further
3. I thought you could only do dynamic IP addressing , wrong

Well I know now :)

Talk works, Ping works , CU-SeeMe works (what is she DOING there???)


Now I have to inform other people ...

Regards & CU 

In article <ko-2212941157270001@komac.knoware.nl>, ko@komac.knoware.nl
(kS) wrote:

> Appletalk extension for MacTCP wanted
> ( NOT appletalk over Internet)
> 
> Please read carefully
> 
> Setup;
> 
> Powerbook , MacIIfx, Printer , all 3 connected by local talk network hardware.
> TO BE CLEAR NO modem No ethernet No Internet Just these 3 pieces off hardware.
> 
> How can I tell MacTCP to use Appletalk/localtalk to let the 2 Mac's
> communicate IP (beit tcp or udp and more) , IP means InternetProtocole NOT
> that I am connected to the Internet. Selecting localtalk doesn't work
> 
> Each Mac has it's own fixed IP address (set manually under MacTCP cntr
> panel more... button)
> 
> 
> I am not looking for Gator-something they are for routing to the internet
> network nor for AIR (Apple Internet Routing) BTW they all have dynamic IP
> adressing and IP packets are send over localtalk besides Appletalk packets
> 
> 
> Then I could do MacPING to the other Mac BTW Why can't I IP ping to myself?
> Then I could start a telnet connection ; one telnet app listining one
> telnet app requesting connection. (Versaterm PRO can, though it uses it's
> own trick and exactly that trick I want to use UNDER MacTCP)
> 
> I have this other App that only uses IP (thus MacTCP) it has no other
> connection-tools.  I want to run this Application on my 2 macs
> 
> 
> 
> 
> Technicaly IMO this extension should do;
> 
> 1. Put up an IPttAppletalk service (tt = tunnel through)
>    pack and unpack IPpackets in/out Appletalkpackets 
> 2. Scan for other IPttAppletalk services
> 3. When started broadcast it's IP address and Appletalk address
> 4. Keep a table off IPaddresses and Appletalk adresses
> 
> etc.
> 
> Someone with the (near) right answer please email me
> 
> Regards Ko
> 
> -- 
> V2S
> Holland
 
-- 
V2S
Holland

-----------[000568][next][prev][last][first]----------------------------------------------------
Date:      Fri, 23 Dec 1994 17:55:46
From:      m_hazel@enet.net (Mike Hazel)
To:        comp.protocols.tcp-ip
Subject:   Re: HELP!! Ip Subnet

In article <3dclpr$qff@tigger.edtel.alta.net> Mike Chan <mchan@mail.edtel.alta.net> writes:
>From: Mike Chan <mchan@mail.edtel.alta.net>
>Subject: HELP!! Ip Subnet
>Date: 22 Dec 1994 19:58:51 GMT
 
>Hi Folks,       
 
>I have some difficulties to subnet my network. Say if I have a class C network  
>with ip address200.200.100.0.  If I use subnet Mask 255.255.255.252. The
>mask bits are 6 and I can have 62 subnet, but, I can only have 2 hosts,
>even though I have 2 bits to play with hosts. Instead of 4, I can only 
>have 2 hosts.  I assume bits 00 and 11 would not be availlabl.e Can 
>anyone tell me why ? Is it for broadcast or multicast ? Because of subnetting,
>I may lose 2 x 63 = 124 ip addresses.  Is there any better solution besides
>switch to 5 bit masking ?
 
>Best Reggards, & Merry Xmas
 
>Mike Chan
>E-mail id: mchan1@agt.alta.net

It is for Broadcast and Network, a host of 0 would be the network address, a 
host of 11 would be the broadcast address for the subnet. As to losing hosts, 
any time you subnet a 'C' you will lose some hosts.

Mike

-----------[000569][next][prev][last][first]----------------------------------------------------
Date:      23 Dec 1994 13:38:29 GMT
From:      duffy@cais.cais.com (Duffy Men)
To:        comp.protocols.tcp-ip
Subject:   HELP!! need information about "netscape"

Can anyone tell me what is "netscape".  Can it work with UNIX system or 
not.  Thank you.

-----------[000570][next][prev][last][first]----------------------------------------------------
Date:      23 Dec 1994 14:39:50 GMT
From:      baker@stat3.mgh.harvard.edu
To:        comp.protocols.misc,comp.protocols.nfs,comp.protocols.tcp-ip
Subject:   Need guide to setting up anonoymous ftp site


We seek to set up a sort-of anonymous ftp site here.  The specifications would
be the following:

	* Ability for people to ftp to our address and be able to get access
	  via a password that they would have to be told beforehand.  This
	  is because there are only about 20 users who we want to let in
	  to ftp files in and out so we want to tell them the password ahead
	  of time so they'll know.  Security is of a mild but at least
	  fairly serious concern.

	* Once inside, the user would have access to a certain part of 
	  the tree's directories but not all.  Specifically, they cannot
	  read anything above the root that they log into (these are SUN
	  SparcStations, incidentally).  For example, they could be logged
	  into /public/home/ftp or whatever, but cannot get into /public/home
	  or just /public or plain old / itself.   And they cannot move
	  across in the tree (to parallel directories) either.

        * Is there a daemon that can run that will accomplish this?  And
	  where can I get it?  


	So this is not a true anonymous ftp but I think it can be done.
      ANYONE with explicit instructions or even just tips on how to
      set this up please e-mail to me ASAP.  Or if there is an ftp site
      at which I cna get info about this kind of ftp site construction
      then please point me to it.   And if there's another kind of 
      newsgroup that deals with this kind of information that might also
      be relevant then I'd be grateful if someone could point me to it.
      Thanks a million (10^6) to anyone who can help in any way no
      matter how small.   E-mail responses please.  Don't post unless
      you want to.  Thanks again.

-Fred Baker	baker@stat3.mgh.harvard.edu
12-23-94
********

			HAPPY HOLIDAYS TO EVERYONE!
			* * * * * * * * * * * * * *


-----------[000571][next][prev][last][first]----------------------------------------------------
Date:      23 Dec 1994 23:11:34 -0500
From:      barmar@nic.near.net (Barry Margolin)
To:        comp.protocols.tcp-ip
Subject:   Re: FTPD always uses port 20 on active open.  Why?

In article <3df12q$363@unix.sri.com> ric@coronacorp.com (Ric Steinberger) writes:
>1) Why always bind to port 20 instead of letting the system give you a port?

Because that's what the FTP protocol says to do.  RFC 959, section 5.2,
paragraph 2 says:

      The server shall initiate the data connection from his own default
      data port (L-1) using the specified user data port.

L is the well-known FTP port, 21, so L-1 is 20.

I'm not really sure why the protocol specifies a particular source port for
data connections.  My best guess is that it allows the client to specify a
remote host and port in his passive open of the data port, so that only the
server can connect to it.  It could also be some kind of remnant of
emulation of the NCP version of the FTP protocol.

>2) I'm a little unclear on how the server can (in it's child processes)
>   establish active opens on the same port (20 in this case).
>   Doesn't this create a problem when ftp clients try to talk to
>   "their" ftp server if "all" the ftp server child process have
>   (what seems to me) the same endpoint address (IP/port pair)?

The port numbers at the two ends should be different.  The port number at
the client end should be the one specified in the PORT command.  If the
PORT command isn't used, it should be the same client port as the control
connection is using.  Thus, if the FTP connection is from client C to
server S, the control connection would be the tuple <C, U, S, 21>, where U
is the arbitrary local port that C's TCP assigns; the default data
connection would be <C, U, S, 20>, and if the client selects a non-default
data port P with the PORT command, it will be <C, P, S, 20>.  You can see
that there's no problem even if C and S are the same system.
-- 

Barry Margolin
BBN Internet Services Corp.
barmar@near.net

-----------[000572][next][prev][last][first]----------------------------------------------------
Date:      23 Dec 1994 13:27:13 +0100
From:      robjan@rabo.nl (Rob Janssen)
To:        comp.protocols.tcp-ip
Subject:   Re: Contents of IGMP for Multicast Control?

In <3dabf5$bmq@ceylon.gte.com> sd03@roger.gte.com (Shuang Deng) writes:

>RFC-1054 describes the format of IGMP packet for reporting
>a host's multicast memebership to multicast routers as 
>consisting of version=1, type=report, and the address of
>the group the host belongs to.
 
>I understand this can be used to notify the router
>to join a multicast group. 
 
>But I cannot figure out the packet content for leaving
>a group. Could some kind, wise soul tell me what the 
>content of the IGMP packet for leaving a particular 
>group is?

There is none.  Just don't report anymore you want to be member of the group.
After some timeout you will be considered to be left.

Rob

-----------[000573][next][prev][last][first]----------------------------------------------------
Date:      Fri, 23 Dec 1994 15:16:08 GMT
From:      bettez@phoque.info.uqam.ca (Jean-Sebastien Bettez)
To:        comp.protocols.tcp-ip,comp.sys.mac.comm
Subject:   Re: IP over localtalk/AppleTalk locally, which MacTCP extension??

Yellow,

You could use localtalk with MacTCP but I think you'll need to have 
class A adresses.


_____________________________________________________________________________

bettez@phoque.info.uqam.ca        
Jean-Sebastien Bettez            

-----------[000574][next][prev][last][first]----------------------------------------------------
Date:      23 Dec 1994 17:23:38 GMT
From:      ric@coronacorp.com (Ric Steinberger)
To:        comp.protocols.tcp-ip
Subject:   FTPD always uses port 20 on active open.  Why?

        Can anyone help with this?  I need to write an application
somewhat similar to ftp/ftpd.  The similarity is that there will
be one socket connection for (ASCII) control information and a second socket
(created after a connection is established on the first) for binary
data.

        I have beed reading up on ftp/ftpd in Richard Steven's TCP/IP
Illustrated, Volume 1.  He explains that the client sends the server an IP
address/port combination and enters a passive listen on that port.  Then the
server is expected to do an active open on that IP/port pair.  It seems that
the server always uses the same port (20).  So I have two questions:

1) Why always bind to port 20 instead of letting the system give you a port?

2) I'm a little unclear on how the server can (in it's child processes)
   establish active opens on the same port (20 in this case).
   Doesn't this create a problem when ftp clients try to talk to
   "their" ftp server if "all" the ftp server child process have
   (what seems to me) the same endpoint address (IP/port pair)?


Thanks in advance to all who reply!

Please send replies to the address below (or post them).

Ric Steinberger
Software Engineer
ric@coronacorp.com

"For De Mille, young fur-henchmen can't be rowing!" Thomas Pynchon
                                                    Gravity's Rainbow


-----------[000575][next][prev][last][first]----------------------------------------------------
Date:      23 Dec 1994 16:18:52 +0100
From:      robjan@rabo.nl (Rob Janssen)
To:        comp.protocols.tcp-ip
Subject:   Re: Telnetd protocol

In <3dad8t$6rv@delphinium.cig.mot.com> ramseyea@drake.cig.mot.com (Erica A. Ramsey) writes:


>Hello, i'm looking for the telnetd protocol. I want to write a program to 
>communicate with a terminal server and send commands. The 
>Current, I use telnet program to communicate with the terminal
>server, but I need to embard to connection to the terminal
>server within a C program. Therefore, I think that it will be
>best if I used the telnetd protocol.
 
>I will take anything I can get, ftp address etc...

There are several RFCs that discuss the telnet protocol.  Have a look
at the rfc-index and work from the cross references you find in there.

Rob

-----------[000576][next][prev][last][first]----------------------------------------------------
Date:      24 Dec 1994 02:55:37 -0500
From:      aray@pipeline.com (Arjun Ray)
To:        comp.protocols.tcp-ip
Subject:   Re: HELP! My first TCP/IP server program doesn't respond :^(

In article <MORGAN.94Dec16193502@world.std.com>,
William M Stair <morgan@world.std.com> wrote:
>It's basically just a daytime server called "TCPdaytimed" that uses
>argv[1] to determine what port to listen to.  I added the line:
> 
>        masterq   256/tcp
> 
>to /etc/services and everything in the server seems to run just fine
>(if it's root) until it gets to the "accept" call.  Then it just sits
>there blocked for ever.
> 
>It's clear that it's resolving the machine name, and what port # to
>connect to just fine, so it's not an obvious installation problem, and
>I know it's NORMAL for "accept" to block, but when I to CONNECT to
>that port with a "telnet localhost 256" the server continues to sit
>there, and my telnet session reports:
> 
>        % telnet localhost 256
>        Trying 127.0.0.1 ...
>        telnet: connect: Connection refused
>also
>        % telnet localhost masterq
>        Trying 127.0.0.1 ...
>        telnet: connect: Connection refused
> 

"Connection refused" means that there is no process, as far as the kernel
knows, listening on port 256. The problem is probably in your IpcPassiveTCP
routine. To isolate it, you may have to split your server code into its
two logical components.

1. The TCPdaytimed() routine. Write a main() wrapper around it, and install
this program in /etc/inetd.conf. In other words, let inetd handle the
socket monitoring. If this works, the problem is almost definitely in
the details of the IPCPassiveTCP() routine.

2. The socket establishment stuff.
[snip]
>  struct sockaddr_in fsin;      /* the from address of a client */
>  char *service = "daytime";    /* service name or port number */
                    ^^^^^^^
shouldn't this be "masterq" ?

[snip]
>        /**************************************************************/
>        /* IpcPassiveTCP basically does:                              */
>        /*                                                            */
>        /*   struct sockaddr_in sin = (bunch of stuff);               */
                                        ^^^^^^^^^^^^^^
Check for the hton..() conversions, is host IN_ADDR_ANY? and so on.
>        /*                                                            */
>        /*   s = socket(PF_INET, SOCK_STREAM, TCP);                   */
>        /*   bind (s, (struct sockaddr *)sin, sizeof(sin));           */
>        /*   listen (s, QLEN);                                        */
>        /*   return s;                                                */
>        /*                                                            */
>        /* According to all the debugging, each of these steps        */
>        /* APPARENTLY works okay.                                     */

I assume you're checking all return values. Do you also have the routine,
at least for debugging, print out the contents of sockaddr_in _after_
the apparently succesful bind()? i.e. what port did it _actually_ bind?
for what IP address? Take No Prisoners!

Hope this helps.

Happy Holidays!

Arjun Ray
The Pipeline Network Inc.

-----------[000577][next][prev][last][first]----------------------------------------------------
Date:      Fri, 23 Dec 1994 20:00:08 GMT
From:      CHENP@ACCTON.COM.TW
To:        comp.protocols.tcp-ip
Subject:   Any Novell transparent BRIDGE.NLM available?

   Does there anyone know if there is any Novell loadable transparent
bridge module of Novell 3.11 and up version?
   I have a tri-segment ethernet-hub and install 3 cards on a Novell server,
intending to connect these 3 ports to this hub seperately and have the
bridging ability.
   Any suggestion or information will be very much appreciated!

-------------------------------------
Chen Pohung   (chenp@accton.com.tw)
Advisory software engineer



-----------[000578][next][prev][last][first]----------------------------------------------------
Date:      Fri, 23 Dec 1994 20:30:35 GMT
From:      aa257@cfn.cs.dal.ca (Tarek Loubani)
To:        comp.protocols.tcp-ip
Subject:   Netscape

I was trying to program Netscape, and I am having trouble with it. It 
seems that when I try to run it, it just hangs up the modem. I suspect 
that this is because of my trumpet connection. I have never tried to 
install internet works, and mosaic, with no luck in either. I suspect the 
problem is the exact same thing. I understand that this might not be the 
right conference, but I thought it was appropriate enough. Please try to 
help me because I have heard so many things about these programs, but 
think that they are a lot of trouble to set up. I am really sorry for 
rambling :). NEway, help is appreciated. Thanx in advance for any 
information.

--
 _____ 
(__ __)
  | | arek		My views are my own. Don't bitch anybody else about
  | |  _		them. Thanx. 
  |_| | | oubani
      | |___
      |_____)

-----------[000579][next][prev][last][first]----------------------------------------------------
Date:      Fri, 23 Dec 1994 20:35:31 GMT
From:      Edward Segall <segall+@cs.cmu.edu>
To:        comp.parallel,comp.lang.misc,comp.protocols.tcp-ip
Subject:   Re: Massively parallel programming for the Internet

In article <thinmanD0toH9.B94@netcom.com>,
Technically Sweet <thinman@netcom.com> wrote:
>Are there any projects attempting to build a massively parallel
>programming environment using spare cycles on the Internet?
...

>David Gelernter's "Mirror Worlds" is clear articulation of this vision.
...

>Back to tech talk: is anybody using multisets as a basis for
>parallel programming?  Multisets looked like a possibly pleasant
>paradigm for programming the distributed interpretive system.

This is an intersting juxtaposition, considering that Linda, Gelernter's
creation, in fact implements distributed multisets.  I haven't seen
"Mirror Worlds" - is this why you asked? In any case, the answer is
"Linda".

--Edward Segall


-----------[000580][next][prev][last][first]----------------------------------------------------
Date:      Fri, 23 Dec 1994 21:27:23 GMT
From:      dougs@kalama.doe.Hawaii.Edu (Doug Stanfield)
To:        comp.protocols.tcp-ip
Subject:   Re: Tel no. or e-mail for Ftp SW

Kent_Chao (kchao@boi.hp.com) wrote:
: Does anyone know Ftp Software, Inc.'s telephone number and/or 
 e-mail address.
: The one I have (617-2460900) is not in service any more.
 
: thx.

Try to gopher or mosaic to ftp://ftp.ftp.com
You should find the number there.
--

Doug Stanfield          "The significant problems we face cannot be solved
Oceanic Cable            at the same level of thinking we were at when we
Project Engineer         created them."  - Albert Einstein 
dougs@kalama.doe.hawaii.edu


-----------[000581][next][prev][last][first]----------------------------------------------------
Date:      23 Dec 1994 22:12:50 GMT
From:      shafto@aristotle.ils.nwu.edu (Eric Shafto)
To:        comp.protocols.tcp-ip
Subject:   PCRoute troubleshooting Q.

I am installing a PCRoute router on an old AT.  Neither port will
return a ping.  At least one of the ports (the one on my side of
the net) is seeing broadcast packets, though, according to pctst.
I can't tell about the other port, since no one else on that network
has *their* routers working either.  :-)

What could make a port see broadcast packets but not respond to a
ping?

I've checked that the addresses and netmasks are right about 30 or
40 times, and I've even tried pinging the wrong one, just in case,
all to no avail.

In the troubleshooting guide that came with PCRoute, it says to try
pctst, and what to do if it won't see broadcast packets.  But it 
doesn't say anything about what to do if it *does* see them.  Any
ideas?



-----------[000582][next][prev][last][first]----------------------------------------------------
Date:      Sat, 24 Dec 1994 01:56:19 +0000
From:      Steve@scotia.co.uk (Steve Holden)
To:        comp.protocols.tcp-ip
Subject:   Re: HELP!! Ip Subnet

In article <3dclpr$qff@tigger.edtel.alta.net>
           mchan@mail.edtel.alta.net "Mike Chan" writes:

> [With Class C 6-bit subnets]  I assume bits 00 and 11 would not be availlabl.e Can 
> anyone tell me why ? Is it for broadcast or multicast ? Because of subnetting,
> I may lose 2 x 63 = 124 ip addresses.  Is there any better solution besides
> switch to 5 bit masking ?
> 
The lowest address in a network or subnet (all-zeroes host number) is by
convention used to refer to the network itself.  The highest (all-ones
host number) is the broadcast address.  Thus 5-bit subnetting appears to
be your only option [unless, that is, someone knows better ...]
-- 
Steve Holden     Scotia Electronic Publishing  +---------------------------+
Tel: +44 436 678962            29 John Street  |    Tools, Training and    |
Fax: +44 436 677814 (bureau)      Helensburgh  | Technology for Successful |
Internet: steve@scotia.co.uk         Scotland  |   Electronic Publishing   |
Compuserve: 100343,2205               G84 8XL  +---------------------------+

-----------[000583][next][prev][last][first]----------------------------------------------------
Date:      Sat, 24 Dec 1994 02:02:02 +0000
From:      Steve@scotia.co.uk (Steve Holden)
To:        comp.protocols.nfs,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc
Subject:   Re: License Manager software?

In article <1994Dec21.093640.24296@taux01.nsc.com>
           moshel@tasu60.nsc.com "Moshe Linzer" writes:

> Is anyone aware of software that will track PC software licenses on an NFS
> server?  The server can be SunOS or a PC running NFSD (If it matters, we are 
> using Wollongong's NFS).  I have seen writeups on a few packages, but they
> seem to work only with PC NOS's like Novell or LAN Manager.  I am also 
> interested in UNIX-PC remote control software that runs over TCP/IP and NFS.
> Any help would be appreciated.
> 
> Moshe
> 
I know Frame Technology were seeking interest in their licensing
scheme, which works very well on Sun networks.  Try e-mailing them
at comments@frame.com.  Other vendors are also active in this area.
-- 
Steve Holden     Scotia Electronic Publishing  +---------------------------+
Tel: +44 436 678962            29 John Street  |    Tools, Training and    |
Fax: +44 436 677814 (bureau)      Helensburgh  | Technology for Successful |
Internet: steve@scotia.co.uk         Scotland  |   Electronic Publishing   |
Compuserve: 100343,2205               G84 8XL  +---------------------------+

-----------[000584][next][prev][last][first]----------------------------------------------------
Date:      Fri, 23 Dec 1994 14:43:36 +0800
From:      peter.lewis@info.curtin.edu.au (Peter Lewis)
To:        comp.protocols.tcp-ip
Subject:   Re: ftp and the SYST command (HELP) - FTP Structured List Command

In article <3d8340$lek@crcnis3.unl.edu>, mgleason@cse.unl.edu (Mike
Gleason) wrote:

> |SIZE: Approximate size of the file in bytes, or for directories, either
> |empty string or a count of contained files and directories
> 
> I've been thinking about this, I think we should have a separate attribute
> for the item count in the directories.  If I'm not mistaken, MacOS for
 example,
> can't tell you this count without counting each item in the directory
manually.

I'd assume that the server would return an empty string if it was hard to
count the number of files in a directory, and return the count if it was
easy.  On the Mac, it's easy, there is a count in the directory entry for
the directory.  If you add an extra attribute, then either the clients
dont send it (and the information is always missing), or the clients send
it, in which case the servers have to waste all that effort calculating it
even if it's hard.

> |The XMRL command works in the same way as the LIST and NLST command (and
> |SHOULD list the same files, with the exception that the special files like
> | and .. MUST NOT be listed).
> 
> Should there be a PARENT_PATH attribute for directories?

There is already a CDUP command, which the client can use.

BTW, all of thise doesn't really help all that much, since we still have
the problems involved in all the different ways of specifying a path (VMS,
Unix, Tandem, DOS, etc).  This is actually a much more difficult problem
than the directory listing.
   Peter.

-- 
To get random signatures put text files into a folder called ³Random Signatures² into your Preferences folder.

-----------[000585][next][prev][last][first]----------------------------------------------------
Date:      23 Dec 1994 14:46:00 +0800
From:      dinesh@bass.com.my (Dinesh Arnold Nair)
To:        comp.sys.sun.admin,comp.protocols.tcp-ip
Subject:   Re: [Q]: What is MX entry in dns.

This article by adar0@routers.com caught Dinesh's eye:
>
>> For example, in my domain's server named configuration file
>> 
>> goofy	IN	A	199.1.1.10
>> 	IN	MX	10 sub-domain.domain.com.
>> 
>In your example, you cut out the stuff above "goofy IN ...", however there
>is likely a "$ORIGIN" statement somewhere above in your example.  So if
>it says "$ORIGIN domain.com.", the above example then means deliver all
>mail destined for anywhere in the "domain.com" to the machine named
>"sub-domain.domain.com".

Sorry Rich, not quite correct. Based on the eg above *and* assuming that
there is an $ORIGIN domain.com, the lines mean that goofy.domain.com's IP
address is 199.1.1.10 and all mail for goofy.domain.com is to be delivered
to the host sub-domain.domain.com. If you wanted all mail addressed to
domain.com. to be delivered, the equivalent MX record is

*	IN	MX	10 sub-domain.domain.com.

Bear in mind that, the above would not cover hosts which have an A record,
only hosts without A records. Hosts with A records, must have a MX record
if mail is to be delivered elsewhere.

Later.

Regards,                           /\_/\   "All dogs go to heaven."
Dinesh.Nair@bass.com.my            (0 0)
+==========================----oOO--(_)--OOo----============================+
| for a in past present future; do                                          |
|   for b in clients employers associates relatives neighbours pets; do     |
|   echo "The opinions here in no way reflect the opinions of my $a $b."    |
| done; done                                                                |
+===========================================================================+



-----------[000586][next][prev][last][first]----------------------------------------------------
Date:      Fri, 23 Dec 1994 14:50:06 +0800
From:      peter.lewis@info.curtin.edu.au (Peter Lewis)
To:        comp.protocols.tcp-ip
Subject:   Re: ftp and the SYST command (HELP) - FTP Structured List Command

In article <3d9bg2$f5b@martha.utk.edu>, hethmon@apac.ag.utk.edu (Paul
Hethmon) wrote:

> Right now, I'm in the process of writing a FTP server and I've
> run up against these problems again and again. At the moment,
> it seems the only thing to do is either contact every author
> of an automated client (not very likely) or try and support
> every new command someone has come up with to try and
> deal with the problem.

Yeah, life sucks like that.  At least match the LIST output format
described in an earlier post.  This takes minimal effort and will at least
get you close.

> Is anyone else ready to pursue an official RFC?

It sounds like a good idea, but I don't have the time to do more than
assist/comment/etc with the document.  What I'd like to see is a
machine-independent specification of the protocol, so that instead of
having servers that each look like their own environment, we get make it
so that all servers look the same.  Eliminating a lot of the fluff in the
FTP spec that no one implements or uses would also be good (like all the
structured record stuff and the like).  A simple, machine independent
specification, that closely approximates what a unix ftp server generally
looks like and that server authors could reasonably easily adhere to would
be wonderful, both for server authors, client authors and users in
general.  Craig Richmond <craig@ecel.uwa.edu.au> actually recently
volunteered to write this RFC, so he should post giving som ideas of how
to go about it.  I'd suggest starting by collecting the relevent RFCs and
then chucking out the crap and tightening up the specification would go a
long way to fixing the problem (in about five years when people actually
implement the damn thing, in the mean time, follow the LIST output
specification...)
   Peter.

-- 
To get random signatures put text files into a folder called ³Random Signatures² into your Preferences folder.

-----------[000587][next][prev][last][first]----------------------------------------------------
Date:      23 Dec 1994 14:55:46 +0800
From:      dinesh@bass.com.my (Dinesh Arnold Nair)
To:        comp.protocols.tcp-ip
Subject:   Re: Printing via Xyplex Terminal Servers

This article by philc@hipstech.demon.co.uk caught Dinesh's eye:
>We have a requirement to print over a TCP/IP network where the remote 
>device handling the printer will be a Xyplex terminal server (with the 
>printer plugged into one of its serial ports).

Same sorta scenario we've implemented at over 10 sites here. We're using a
mix of making dedicated connections between host and termserver using
xyp_ptyd and on demand connections using xyp_filt. Both ways, the printer
is configured as local on the host, though the xyplex has modules to allow
it to act as a printer server. In this case, the printer would be local on
xyplex and remote on host.

>
>We are given to understand that telnet End Of Record will need to be 
>negotiated with the terminal server in order to do this - I would guess 
>that this is not one of the "ordinary" negotiations completed when you 
>open a Telnet connection.

It depends on your options to xyp_filt. If Telnet EOR Reflection is
enabled on the port, you *must* enable it when you call xyp_filt. If
disabled, you *must not* give the -eor option to xyp_filt. 


>Has anybody had any experience with Xyplex Terminal servers?
>Will we need to develop our own interface script/program?
>If we do, what should we look out for?

I assume you've already modified the printer interface script to pipe the
file to xyp_filt. Other than adding/removing the -eor option, there's
nothing else to do. 

Regards,                           /\_/\   "All dogs go to heaven."
Dinesh.Nair@bass.com.my            (0 0)
+==========================----oOO--(_)--OOo----============================+
| for a in past present future; do                                          |
|   for b in clients employers associates relatives neighbours pets; do     |
|   echo "The opinions here in no way reflect the opinions of my $a $b."    |
| done; done                                                                |
 +===========================================================================+

-----------[000588][next][prev][last][first]----------------------------------------------------
Date:      24 Dec 1994 15:02:37 -0500
From:      mailcom@aol.com (MailCom)
To:        comp.protocols.tcp-ip
Subject:   Re: Assymetric Bandwidth

In practice, an asymetric bandwidth link often ends up with a large
bandwidth * delay product. This is because the bandwidth part of the
equation is driven by the highest bandwidth side of the link, and the
delay is driven by the highest latency part of the link. 

Hypothetical (somewhat artificial):
One side of the link is a 10 Mbps Ethernet, the other side is a 1.5 Mbps
satellite link with 600 ms latency.  Result: Bandwidth-delay product is
approximately 10 Mbps * .600 s = 6 Mb!

With such a high bandwidth-delay product, transfer efficiency will be very
low unless a large window size can be used. The TCP Window Scale option
was developed to address this, allowing much larger window sizes. 

-----------[000589][next][prev][last][first]----------------------------------------------------
Date:      24 Dec 1994 15:07:25 GMT
From:      softlink@zeus.datasrv.co.il (SoftLink)
To:        comp.protocols.tcp-ip,comp.sys.sun.admin
Subject:   OOB messages on Solaris 2.3 and NCR SVR4

Hello,

We are currently developing a multi platform application in which we use
Out Of Band messages. In order to receive the OOBs we call the signal
function with the SIGURG parameter, and then call
ioctl(socket,SIOCSPGRP,pid) which enables signaling for our process.
This works fine on many opearting systems ( SunOS 4.1 , Linux , OSF
etc...) but not on Solaris 2.3, nor on NCR computers running the AT&T
SVR4 operating system ( system V ).

Is anyone familiar with the problem of receiving OOBs on Solaris 2.3 ?
If the call we are currently doing is wrong for Solaris, does anyone know
a better way of enabling OOBs ?

thanks,
Miki George - SoftLink LTD.


-----------[000590][next][prev][last][first]----------------------------------------------------
Date:      Sat, 24 Dec 1994 16:00:42 GMT
From:      aa257@cfn.cs.dal.ca (Tarek Loubani)
To:        comp.protocols.tcp-ip
Subject:   Netscape

I don't remember if I posted this already, so please don't be upset if I 
did. NEway, I am having trouble with mosaic, and netscape and inet works. 
It seems that none of them would work. They all have the same problem, 
They hang up the modem when they try to run. The only commonality between 
all three is the trumpet. This is a reason why I think that the programs 
aren't working. Please, any help will be appreciated. I would like to get 
this program run sometime while I'm alive.

--
 _____ 
(__ __)
  | | arek		My views are my own. Don't bitch anybody else about
  | |  _		them. Thanx. 
  |_| | | oubani
      | |___
      |_____)

-----------[000591][next][prev][last][first]----------------------------------------------------
Date:      Sun, 25 Dec 1994 00:57:55 +0000
From:      Andrew@cucumber.demon.co.uk (Andrew Gabriel)
To:        comp.protocols.tcp-ip
Cc:        peter.lewis@info.curtin.edu.au
Subject:   Re: ftp and the SYST command (HELP) - FTP Structured List Command

In article <peter.lewis-2312941450070001@zany.peter.com.au>
           peter.lewis@info.curtin.edu.au "Peter Lewis" writes:

>                                        What I'd like to see is a
> machine-independent specification of the protocol, so that instead of
> having servers that each look like their own environment, we get make it
> so that all servers look the same.

This sounds reasonable.

>                                     Eliminating a lot of the fluff in the
> FTP spec that no one implements or uses would also be good (like all the
> structured record stuff and the like).  A simple, machine independent
> specification, that closely approximates what a unix ftp server generally
> looks like and that server authors could reasonably easily adhere to would
> be wonderful, both for server authors, client authors and users in
> general.

This sounds terribly contradictory to me.
You want "machine independent ... approximates what a unix server
looks like ... eliminating structured record stuff and the like".

Well sorry, but not all filing systems look like unix. Many operating
systems have much more fully featured filing systems, and require
these parts of the ftp protocol to make full use of their filing
systems. Because unix can't have files with variable blocksizes,
or a file containing binary data and significant record boundaries,
it doesn't mean that these are not used by other operating systems
and handled by ftp.

Unix (or probably DOS) has the simplest (lowest common denominator)
filing system [of those I have personally come across] which is still
in common use. This makes it unsuitable as a model for a machine
independent protocol because it is missing features required by more
sophisticated filing systems. This issue is already addressed by the
current ftp protocol.
[Incidently, I'm not trying to knock the unix filing system here; one
of its nicest features is its simplicity.]

By all means extend the ftp protocol, but please be sure to do it
without unix blinkers on. 

Merry Christmas,
-- 
Andrew Gabriel                        Home: Andrew@cucumber.demon.co.uk
                                      Work: Andrew.Gabriel@gpt.co.uk

-----------[000592][next][prev][last][first]----------------------------------------------------
Date:      Sun, 25 Dec 1994 01:17:08 +0000
From:      Andrew@cucumber.demon.co.uk (Andrew Gabriel)
To:        comp.protocols.tcp-ip
Cc:        ric@coronacorp.com
Subject:   Re: FTPD always uses port 20 on active open.  Why?

In article <3df12q$363@unix.sri.com> ric@coronacorp.com "Ric Steinberger" writes:

> 2) I'm a little unclear on how the server can (in it's child processes)
>    establish active opens on the same port (20 in this case).
>    Doesn't this create a problem when ftp clients try to talk to
>    "their" ftp server if "all" the ftp server child process have
>    (what seems to me) the same endpoint address (IP/port pair)?

A TCP/IP connection is defined by the endpoint addresses at _both_ ends
of the connection. There is no problem with lots of separate TCP/IP
connections going to (or coming from if you prefer) the same endpoint
address, but no two of them can have the same endpoint address at the
other end of the connection.

This is no different from lots of users telnetting into a server; they
will all have the server's IP address and telnet port number for one
end of the connection, but the other ends of the connections must all
be different. (It makes no difference which end set up the connection.)

Merry Christmas,
-- 
Andrew Gabriel                        Home: Andrew@cucumber.demon.co.uk
                                      Work: Andrew.Gabriel@gpt.co.uk

-----------[000593][next][prev][last][first]----------------------------------------------------
Date:      25 Dec 1994 13:53:24 GMT
From:      wesley@hk.net (Chan Kim-Ho)
To:        comp.protocols.tcp-ip
Subject:   Re: Mosaic server for an AS/400

My company plan to use tcpip as our standard protocol to communciate with 

mainframe or mini.  We currently using IBM TCP/IP SNA gateway to 

communciate with IBM E9000 and for AS/400, we use Elite/400 with IPX.  

I read form the news group about using TCPIP to connect to AS/400. Just 

wondering how its work.  Is it like unix system, start a TCPIP service on 

AS/400, then use 5250 emulation (in LWP or Chamelon) telnet to AS/400.  Do

PC support still need to run on AS/400? How about file upload, can I use

FTP, there is a lot of parameter in PC support upload. How to do that 

through FTP?  Thanks in advance.

 

 

Regards,

Wesley Chan

 



-----------[000594][next][prev][last][first]----------------------------------------------------
Date:      Sun, 25 Dec 1994 12:38:37 +0500 (GMT+0500)
From:      ashwin@wipinfo.soft.net (Ashwin Hegde)
To:        comp.protocols.tcp-ip
Subject:   Unnumbered IP

Hi,
    Can anyone provide information about unnumbered (or numberless) IP. 
    Are there any RFC's on this topic?
    Kindly reply to  ashwin@wipinfo.soft.net
Thanx in advance
Ashwin.

-----------[000595][next][prev][last][first]----------------------------------------------------
Date:      Mon, 26 Dec 1994 11:24:58 pacific
From:      wfre@slip.net (Walt Freedman)
To:        comp.protocols.tcp-ip
Subject:   Re: Netscape

In article <D1BoH7.H79@cs.dal.ca> aa257@cfn.cs.dal.ca (Tarek Loubani) writes:

>The only commonality between 
>all three is the trumpet. This is a reason why I think that the programs 
>aren't working. Please, any help will be appreciated. I would like to get 
>this program run sometime while I'm alive.

You will probably get more help in comp.infosystems.www.users.
But, in the meantime, here is a copy of my LOGIN.CMD for trumpet for use over 
a PPP connection. (You need first to run Trumpet's "SETUP.CMD" script to set 
your userID and password -  and, of course, set IP address, gateway etc as 
decribed in the INSTALL.TXT file that comes with the Trumpet distribution).

-----------[000596][next][prev][last][first]----------------------------------------------------
Date:      26 Dec 1994 03:47:53 GMT
From:      ldeckard@databank.com (Lenny Deckard)
To:        comp.protocols.tcp-ip
Subject:   How can you automate an FTP download?

Weekly, I download a file from the IBM mainframe (TSO) to our IBM RISC 6000 
box. I am currenly doing this manually on the RISC 6000 by entering: 'ftp <the 
mainframe TSO IP address>', then the necessary commands to download the file. I 
would like to automate this procedure. The file name I download is consistent. 
Is there a way, from within a UNIX AIX script file to accomplish this. If this 
is possible I can incorporate this to be fired off by a cron job weekly. I 
can't always access the net news group as often as I would like so if can 
please email direct would be great. You help would be greatly appreciated. 
Thanks! Lenny Deckard

email: ldeckard@databank.com



-----------[000597][next][prev][last][first]----------------------------------------------------
Date:      Mon, 26 Dec 1994 15:51:22
From:      fks@ftp.com (Frances K. Selkirk)
To:        comp.protocols.tcp-ip
Subject:   Re: Urgent: SMTP Server for DOS

In article <1994Dec23.002659.15022@uab.es> perec@freenet.fsu.edu (Pere Camps) writes:

> >> I'm looking desperatly for a DOS-TSR based SMTP server program.
 
> >I don't think there are any DOS or Windows based sendmail type 
> >programs.
> 
>         That's a pity... do you know anything about PC/TCP software?
> It might work...

Well, PC/TCP (for DOS) has an SMTP server and an SMTP client, so the
pieces are there, but they are not integrated and the server is not a
TSR. Run it in DOS, and that is all that your PC is doing. 

Sorry...

--
Frances K. Selkirk					 fks@ftp.com
FTP Software, Inc   Technical Information Services   support@ftp.com
====================================================================
Support's newsletter provides technical information on our products.
     FTP to ftp.ftp.com, or check our Web server at www.ftp.com






-----------[000598][next][prev][last][first]----------------------------------------------------
Date:      Mon, 26 Dec 1994 11:43:02 GMT
From:      gani@netcom.com (Gani Cortez)
To:        comp.protocols.tcp-ip
Subject:   Re: TIA

In article <3craar$j1e@portal.gmu.edu>, cozmen@osf1.gmu.edu (Cenk Ozmen) says:
>
>
>
>        Hello, I was wondering if anyone knows where I can get "tia".  I
>use Mosaic for Windows and I need tia to run it from home.  Thanks a
>bunch.  If you could please e-mail, that would be very nice:
>
>cozmen@osf1.gmu.edu

You could ftp to marketplace.com or send a email to tia-info@marketplace.com
This is where you can get tia. However, you should check for compatibility
with you platform.

-----------[000599][next][prev][last][first]----------------------------------------------------
Date:      Mon, 26 Dec 1994 20:00:39 GMT
From:      aboba@netcom.com (Bernard Aboba)
To:        comp.protocols.tcp-ip.ibmpc,comp.protocols.tcp-ip,alt.winsock,comp.os.ms-windows.networking.tcp-ip,alt.answers,comp.answers,news.answers
Subject:   comp.protocols.tcp-ip.ibmpc Frequently Asked Questions (FAQ), part 1 of 3

Archive-name: ibmpc-tcp-ip-faq/part1

comp.protocols.tcp-ip.ibmpc:
FAQ Posting, part 1, 1/1/95


#################  Legalese #################

This document is Copyright (C) 1993,1994,1995 by Bernard Aboba, except where 
the copyright is retained by the original author(s). This document
may be distributed non-commercially, provided that it is not
modified in any way. However, no part of this publication may be 
sold or packaged with a product for sale in any form without the 
prior written permission of Bernard Aboba. 

This FAQ is presented with no warranties or guarantees of ANY KIND
including correctness or fitness for any particular purpose. The
author(s) of this document have attempted to verify correctness of the
data contained herein; however, slip-ups can and do happen.  If you use
this data, you do so at your own risk. While we make every effort to
keep this FAQ up to date, we cannot guarantee that it is, and we will
not be responsible for any damages resulting from the use of the information
or software referred to herein. 

Unless otherwise stated, the views expressed herein are my own.  Last
time I looked, I had not been appointed official spokesperson of any
of the following:

	The Planet Earth
	The U.S.Government
        The Internet
        Microsoft, IBM, Sun or Apple
	The State of California (not so good)
	The University of California, Berkeley
	The City of Berkeley (bringing you Riot of the Week(SM))
        Addison-Wesley
        Publisher's Group West
	Any major or minor breakfast cereal (not even oatmeal!)
	
This FAQ will be posted monthly. In between it will be
available as: 
ftp://ftp.netcom.com/pub/ma/mailcom/IBMTCP/ibmtcp.zip

Please note the change in the archive address!

############ This FAQ is on the Web ###############

After each posting, this FAQ is automatically
converted to HTML by Ohio State, and made available 
on the Web. This means that if you have a WWW browser, 
you can read the FAQ online, and click on links 
to download individual files. 

This is how I read the FAQ myself, and it is
highly recommended. To get at the HTML version, try:
http://www.zilker.net/users/internaut/update.html

################# Citation entry  #################

This FAQ may be cited as:

  Aboba, Bernard D.(1994) "comp.protocols.tcp-ip.ibmpc Frequently
  Asked Questions (FAQ)" Usenet news.answers, available via 
  ftp://ftp.netcom.com/pub/ma/mailcom/IBMTCP/ibmtcp.zip, 
  57 pages.


################# Change History  #################

Changes from 11/1/94 posting:
Updated commercial stack entries, cleaned up broken
links reported by readers (thanks!). New entries
sport a "*"
 
################# Related FAQs #################

There is a FAQ available on features of TCP/IP
Packages for DOS and Windows. This is available at:
ftp://ftp.cac.psu.edu/pub/dos/info/tcpip.packages. 

The Windows Sockets Faq is posted to alt.winsock, and
is available at:
ftp://SunSite.UNC.EDU/pub/micro/pc-stuff/ms-windows/winsock/FAQ

The PC-NFS FAQ is available at:
ftp://seagull.rtd.com/pub/tcpip/pcnfsfaq.zip 
ftp://ftp.york.ac.uk/pub/FAQ/pcnfs.FAQ

The SNMP FAQ is regularly posted to news:comp.protocols.snmp

The TCP/IP FAQ is posted to news:comp.protocols.tcp-ip, and is
maintained by mailto:gnn@netcom.com. 

The Windows NT FAQ is available from
Steve Scoggins, mailto:sscoggin@enet.net 

An NT Web FAQ is available from:
http://www.luc.edu/~tbaltru/faq/

The "How To Get It" FAQ on the Crynwr packet driver 
collection is irregularly posted to news:comp.protocols.tcp-ip.ibmpc
by Russ Nelson, mailto:nelson@crynwr.com. 


################# Book info  #################

A bunch of you have requested information on the
book I am working on.  Here is the basic info:

Title: The PC-Internet Connection, TCP/IP
Networking for DOS and Windows
Authors: Bernard Aboba
Pages: 600 (estimated), 8.5" by 11"
Distributor: Publisher's Group West
ISBN: 1-883979-00-5
Price: $32.95 (est), includes CD-ROM with
Chameleon Sampler software, WS-Gopher, PC-Eudora, 
Etherload, EtherDump, much more.  
Estimated release: No date set (yeah, it slipped a 
few months)

To look at sample chapters from the book, check out
the following WWW page:  
http://www.zilker.net/users/internaut/forth.html

I'm currently working on incorporating comments
from the first beta release, finishing the index
and generally cleaning it up, so the second beta
release won't be out for at least another month.   
Until the second beta is out, I will not be accepting
review requests. Sorry!

########### COOL WWW PAGES relating to TCP/IP ##########

http://www.charm.net/ppp.html (Cool home page with lots of pointers to
TCP/IP stuff)
http://www.zilker.net/users/internaut/update.html (This FAQ, in HTML)
http://www.crynwr.com/crynwr/nelson.html  (Crynwr Software Home Page)
ftp://ftp.biostat.washington.edu/ftp/pub/msdos/network.setups 

################# EXAMPLE CONFIG FILES  #################

Many thanks to Dave Fetrow (fetrow@biostat.washington.edu)
for creating an archive of setup files. The archive is 
particularly oriented toward sets of applications that 
are somewhat tricky, such as combinations involving 
different driver sets, mixtures of NetWare, TCP/IP, 
and W4WG, etc. 

Please include not only the setup and configuration files
but some directions. Comments included with the setup files
are highly desirable. The files can include your name if you
desire. 

Please mail submissions to mailto:ftp@ftp.biostat.washington.edu. 

The archive itself is located at:
ftp://ftp.biostat.washington.edu/ftp/pub/msdos/network.setups 

Late breaking development: the archive has crashed, and 
files have been lost. 



TABLE OF CONTENTS

A. Components of a TCP/IP solution

A-1. What do I need to run TCP/IP on the PC?
A-2. What are packet drivers?  Where do I get them?
A-3. What is Winsock?  Where can I get it? 
A-4. What is Trumpet Winsock? How do I get it to dial? 
A-5. What publicly distributable TCP/IP applications are there
     for DOS?  Windows?
A-6. What software is available for doing SLIP?  Compressed SLIP?
     PPP?  For DOS?  For Windows?
A-7. What about the software included with various books? 
A-8. What diagnostic utilities are available to find problems with
     my connection?  Where can I get them?
A-9. Is there a CD-ROM with the software included in this FAQ? 
A-10. Does Windows NT support SLIP? PPP? 
A-11. Where can I get Microsoft TCP/IP-32? 
A-12. How do I get my BBS to run over TCP/IP? 
A-13. Are there graphical TCP/IP servers out there?
A-14. What methods of dynamic address assignment are available?
A-15. How can I set up PPP server on a UNIX host? 
A-16. What is WinSNMP? Why doesn't my TCP/IP stack support SNMP? 
A-17. What HTTP proxies are available for use with NCSA Mosaic? 
A-18. Why doesn't my Web browser support direct WAIS queries? 
A-19. What is SOCKS? What TCP/IP stacks support it?
A-20. How can I handle authentication on my NNTP server? 
A-21. What is SLIPKnot? 
A-22. What is TWinSock? 

B. Questions about drivers

B-1. What do I need to know before setting up SLIP or PPP?
B-2. How do I configure SLIPDISK?
B-3. How do I install packet drivers for Windows applications?
B-4. When do I need to install  WINPKT? 
B-5. How to do I run both WinQVT and ODI?
B-6. Is it possible to use BOOTP over SLIP?
B-7. How do SLIP drivers work? 
B-8. When do I need to install PKTMUX?
B-9. Can NDIS be used underneath multiple protocol stacks of the same type?
B-10. Is there an NDIS over packet driver shim? 
B-11. How do I run NetBIOS over TCP/IP? 
B-12. How do I run NFS and another TCP/IP application?
B-13. How do I run Trumpet Winsock along with KA9Q or NFS? 
B-14. I am trying to run Netware and TCP/IP at the same time, using
      PDETHER. How do I do this? 
B-15. Sample Stick Diagrams
B-16. Strange and wonderful configuration files submitted by readers

C. KA9Q Questions

C-1.  What version of KA9Q should I use and where do I get it?
C-2.  What do I need to run KA9Q? Why won't it do VT-100 emulation?
C-3.  How do I configure KA9Q as a SLIP dialup connection?
C-4.  How do I configure KA9Q as a router?
C-5.  How do I get KA9Q to support BOOTP?
C-6.  How do I get KA9Q to support PPP?
C-7.  How do I get KA9Q to support SLIP dialin?
C-8.  Can I use KA9Q as a packet filter?
C-9.  Can I use KA9Q as a BOOTP server?
C-10. Where can I get a manual for KA9Q?
C-11. Is there a way to prevent KA9Q from listening to ICMP redirect
      packets? RIP packets?  
C-12. Will KA9Q route source-routed packets? If so, is there any way to
      turn off this (rather undesirable) behavior?
C-13. I'm trying to use the TextWin version of KA9Q as a SLIP router
      and it isn't working. What's wrong?

D. PCROUTE and PCBRIDGE

D-1. How do I get PCROUTE set up?
D-2. I want to use MS TCP/IP-32 to contact a host over a serial link,
     but have no SLIP or PPP driver. What do I do?
D-3. How do I get PCBRIDGE to use a SLIP or PPP driver?
D-4. Can I get PCROUTE to switch off RIP? 

E. Hints for particular packages

E-1. How do I get DesQView X to run over the network?
E-2. Why is NFS so slow compared with FTP?
E-3. Where can I get information on running NetWare and TCP/IP
      concurrently? 
E-4. What NetWare TCP/IP NLMs are out there and how do I get them
      to work? 
E-5. How do I get a telecom package supporting Int 14h redirection
      to work? 
E-6. I am having trouble running Netmanage Chameleon apps along with
     WFW TCP/IP-32. What do I do? 
E-7. How do I get Windows For Workgroups to work alongside NetWare?
E-9. How come package X doesn't support the AppleTalk packet driver?
E-10. NCSA Telnet doesn't reassemble fragments. What should I do?
E-11. I am trying to configure a Macintosh to set its parameters automatically   
      on bootup, but it isn't working. What's wrong?
E-12. I've heard that DHCP is a potential security risk. Is this true? 
E-13. What is TIA?
E-14. What PC TCP/IP implementations support recent advances?  
E-15. What network adapters have on-board SNMP agents?
E-16. What is the easiest way to get WFW and Novell Netware to coexist?
E-17. I'm trying to use packet driver software alongside WFW v3.11 and
      am having a hell of a time. What should I do?
E-18. What proxy software is available for those concerned about security?
E-19. How do I mount ftp.microsoft.com on the desktop using file manager? 
E-20. I am having trouble connecting to a Windows NT PPP server. What should
      I do? 

F. Information for developers

F-1. What publicly distributable TCP/IP stacks are there that I can
     use to develop my own applications?
F-2. Where can I get a copy of the Windows Sockets FAQ?


--------------------- FAQ Begins Here ---------------------------

A. Components of a TCP/IP solution

A-1. What do I need to run TCP/IP on the PC?

To run TCP/IP on the PC you will need:

* Appropriate hardware, such as:

    Ethernet card
    Token Ring card
    AppleTalk card
    Serial Port
   

  Any other network card with a packet driver or NDIS or ODI driver,
  (such as Arcnet), will also work.  If your card supports NetBIOS,
  this is also acceptable, since you can run a packet-driver-over-
  NetBIOS shim.

* Drivers for your hardware.

  Your card probably came with one or more of the following drivers:
 
    Network Device Interface Specification (NDIS) drivers
      [spec. by 3Com and Microsoft, used by LAN Manager, Windows
      for Workgroups, and Windows NT. LAN Manager uses NDIS 2.0,
      Windows NT uses 3.0, and WFW supports 2.0 and will support 
      3.0]
    ODI Drivers [spec. by Novell, abbreviation for Open DataLink
      Interface]
    Packet Drivers [spec. by FTP Software]
   
  TCP/IP stacks have been written for each of these driver interfaces, 
  so the important thing is whether your chosen stack is compatible 
  with the interface available for your card.
 
  A shim is software that runs on top of one set of drivers to
  provide an interface equivalent to another set. This is useful,
  for example,if you are looking to run software requiring an
  NDIS driver(such as Chameleon NFS) alongside software
  requiring a packet driver interface (such as KA9Q, Gopher, Popmail,
  NCSA Telnet, etc.), or run software intended for, say, a packet
  driver over an NDIS driver instead.
 
  Shims are available to run packet drivers over NetBIOS, ODI,
  or NDIS, in order to run software expecting a packet driver over
  NDIS, ODI, or NetBIOS instead. There are also shims to run NDIS
  over ODI (ODINSUP), and ODI over Packet Drivers (PDETHER). 


* A TCP/IP protocol stack.

  The TCP/IP protocol stack runs on top of the driver software, and
  uses it to access your hardware. If you are running a TCP/IP
  protocol stack that requires drivers that aren't available for your
  hardware, you're in trouble. Check into this before purchasing!
 

* If running Windows applications that require it, WINSOCK.DLL. 


  Windows Sockets is a sockets interface which was created as a 
  Windows DLL. Each  TCP/IP implementation requires its own version 
  of Windows Sockets. Trumpet Winsock and VxDTCP are the only
  two publicly distributable Windows Sockets implementations. 
  WINSOCK.DLL provides 16-bit support; WSOCK32.DLL provides 32-bit support. 

   
* Applications software.

  Although most of us in this newsgroup seem to spend our time
  looking for working combinations of applications,WINSOCK.DLLs,
  Windows Sockets compliant TCP/IP implementations, shims, 
  drivers, and hardware, ultimately your goal is eventually to 
  run an application successfully. If and when that happens, 
  please send me a note, so I can add it to this FAQ.


A-2. What are packet drivers?  Where do I get them?

Packet drivers provide a software interface that is independent of the  
interface card you are using, but NOT independent of the particular 
network technology. As Frances K. Selkirk (mailto:fks@vaxeline.ftp.com) notes:

"That's one reason they're easier to write than ODI drivers! If you
write a class three (802.5 Token Ring) driver, you will need to use
software that expects a class three driver, not software that expects
a class 1 (DIX ethernet) driver. There are a few drivers that fake class 1. 

I believe only class 1 and class 6 (SLIP) drivers are supported by 
freeware packages."

The chances are fair that your Ethernet card came with a packet
driver, and if so, you should try that first. If not, then you
can try one of the drivers from the Crynwr collection (formerly
called the Clarkson Drivers). See the Resource listing for info.

For 3COM drivers, try ftp://ftp.3com.com/pub For technical information,
try mailto:info@3com.com. For marketing and product info, try
mailto:leads@hq.3mail.3com.com.The packet driver specification is available
from ftp://vax.ftp.com/packet-d.ascii

The following vendors have packet drivers with source available for
their pocket lan adaptors:

D-Link - +1-714-455-1688
Solectek - +1-619-450-1220
Accton - +1-415-266-9800
Compulan - +1-408-922-6888
(soon Kodiak's Noteport - +1-408-441-6900)

You can obtain a complete library of packet drivers from many of the
Simtel20 mirror sites, including:
ftp://oak.oakland.edu/pub/msdos/pktdrvr/pktd11.zip, 
ftp://oak.oakland.edu/pub/msdos/pktdrvr/pktd11c.zip. 


A-3. What is Windows Sockets?  Where can I get it?

  The idea for Windows Sockets was born at Fall Interop '91, during a
  Birds of a Feather session.  

  From the Windows Sockets specification:
  [courtesy of Mark Towfiq, mailto:towfiq@Microdyne.COM]:
 

    The Windows Sockets Specification is intended to provide a
    single API to which application developers can program and
    multiple network software vendors can conform. Furthermore, in
    the context of a particular version of Microsoft Windows, it
    defines a binary interface (ABI) such that an application
    written to the Windows Sockets API can work with a conformant
    protocol implementation from any network software vendor.

Windows Sockets will be supported by Windows, Windows for Workgroups,
Win32s, and Windows NT. It will also support protocols other than TCP/IP.
Under Windows NT, Microsoft will provides Windows Sockets support over
TCP/IP and IPX/SPX. DEC will be implementing DECNet. Windows NT will 
include mechanisms for multiple protocol support in Windows Sockets, 
both 32-bit and 16-bit.

Mark Towfiq writes:

    "Files and information related to the Windows Sockets API are
     available via ftp://sunsite.unc.edu/pub/micro/pc-stuff/ms-windows/winsock, 
     which is a mirror of ftp://microdyne.com/pub/winsock (SunSite has a much
     faster connection to the Internet, so you are advised to use
     that).

     If you do not have FTP access to the Internet, send a message
     with the word "help" in the body to either
     mailto:ftpmail@SunSite.UNC.Edu, or mailto:ftpmail@DECWRL.DEC.Com, to obtain
     information about the FTP to Mail service there."
 
  Alternative sources for the Windows Sockets specification include
  ftp://ftp.microsoft.com/ (an FTP server running NT), as well as the
  Microsoft forum on CompuServe (go msl).
  
  Currently NetManage (NEWT), Distinct, Spry, FTP and Frontier are shipping
  Winsock TCP/IP stacks, as is Microsoft (Windows NT and TCP/IP for
  WFW), Beame & Whiteside Software (v1.1 compliant), and Sun PC-NFS. 
  If you are looking for a Winsock.dll, you should first contact your TCP/IP
  stack vendor. Novell has one in beta for their Lan Workplace for DOS.

A-4 What is Trumpet Winsock? How can I get it to dial? 

Peter Tattam has released a shareware Windows Sockets compliant
  TCP/IP stack. You can obtain it via    
  ftp://ftp.utas.edu.au/pc/trumpet/winsock/winsock.zip, 
  ftp://ftp.utas.edu.au/pc/trumpet/winsock/winapps.zip
  ftp://biochemistry.bioc.cwru.edu/pub/trumpwsk/winsock.zip.   
  ftp://biochemistry.bioc.cwru.edu/pub/trumpwsk/winapps.zip.   
  
The first thing to do after you download WINSOCK.ZIP is to create
a directory for Trumpet Winsock, such as C:/TRUMPWSK, and put it
in your DOS PATH statement. 

Trumpet Winsock operates over packet drivers, or over a serial port
using its own built-in SLIP/CSLIP and PPP. If you are using a network
adapter, this means that you will have to locate a packet driver
for your adapter, and load it. Trumpet Winsock also comes with 
WINPKT, and this is loaded next, via the command
WINPKT.COM 0x60 [or whatever the software interrupt for your packet driver]

You will then enter Windows, and create a group in the Program Manager
for all the files that come with Trumpet Winsock. The stack itself is loaded
by executing TCPMAN. Applications that come with it include WinCHAT, 
a chatting program; PINGW, a ping utility; FTPW for FTP, WINARCH for Archie. 

When you first execute TCPMAN, you will be asked to fill out the setup 
information for the stack. Select whether you will be using a network
adapter or SLIP; you cannot use both. 

Since Trumpet Winsock now supports PPP, you do not need to load an Ethernet
simulation drivers such as EtherPPP. 

If for some reason you don't like Trumpet Winsock's scripting language, 
you can use any other comm program that doesn't drop carrier on exit, or 
the DIALER program, available via: 

ftp://ftp.cica.indiana.edu/pub/pc/win3/util/dialexe.zip. 

You can also use EtherPPP (ftp://merit.edu/pub/ppp/pc/etherppp.zip) 
instead of Trumpet Winsock's built-in PPP. This is an Ethernet simulation driver, 
so you will configure Trumpet Winsock as though it were running over an Ethernet 
Packet driver, i.e. by loading WINPKT 0x60, and setting the packet driver
vector in TCPMAN to 0x60. 

EtherPPP comes with its own dialer, so you will need to create a dialing script.
If your TCP/IP address will be changing, you will also need to write a little
batch script to capture the assigned IP address, and insert it into Trumpet's
initialization file.  EtherPPP takes up too much RAM (121K), but otherwise
works fine.  

As for Trumpet Winsock's built-in scripting language, the default dialout
script is LOGIN.CMD. A sample LOGIN.CMD file from Geoff Cox 
(mailto:geoff@satro.demon.co.uk):

#
# initialize modem
#
output atzm0\13
input 10 OK
#
# set modem to indicate DCD
#
output at&d2&c1\13
input 10 OK\n
#
# send phone number
#
output atdt0813434848\r
#
# my other number
#
#output atdt241644\13
#
# now we are connected.
#
#input 30 CONNECT
#
#  wait till it's safe to send because some modem's hang up
#  if you transmit during the connection phase
#
#wait 30 dcd
#
# now prod the terminal server
#
#output \13
#
#  wait for the username prompt
#
input 30 ogin:
username Enter your username
output \satro\r
#
# and the password
#
input 30 assword:
password Enter your password
output \my password\r
#
# we are now logged in
#
input 30 otocol:
#
# see who on for informational reasons.
#
output SLIP\r
input 30 HELLO


A-5. What publicly distributable TCP/IP applications are there for
     DOS?  Windows?

Right now there are a wealth of publicly distributable TCP/IP
applications running under DOS.  Windows also has a wealth of 
programs available, including implementations of Gopher, Mail
(POP3/SMTP), FSP, WWW, Telnet, FTP, IRC, and WAIS. 

See the Resource listings for information.


A-6. What software is available for doing SLIP?  Compressed SLIP?
     PPP?  For DOS?  For Windows?  For OS/2?

Trumpet Winsock now supports both PPP as well as SLIP/CSLIP. 

For SLIP or CSLIP use with DOS, I recommend using SLIPPER or CSLIPPER. 
These are packet drivers that can be used along with a dialer. For PPP, 
I recommend the EtherPPP packet driver described above.  

There is a special version of NCSA Telnet for PPP, available from
ftp://merit.edu/pub/ppp/pc. 

KA9Q supports SLIP/CSLIP as well as PPP, but unfortunately can not be used as a
TCP/IP protocol stack to run other apps.

I have heard good things about IBM's TCP/IP for OS/2, but haven't
used it msyelf. Please see the FAQ from news:comp.os.os2.networking for details.

IBM, FTP Software, Beame & Whiteside, Frontier, SPRY and Netmanage also 
offer SLIP support in their products. See the resource listings for details.  


A-7. What about the software included with various books?

The software included with various books (including mine) is usually
Chameleon Sampler from NetManage. Sampler supports SLIP/CSLIP/PPP, but
not connection over a network, and includes software for FTP, Telnet,
TN3270, and Mail. The stack included with Sampler (NEWT) is Winsock
compatible, so you can run any Windows Sockets-compatible application
over it. Installation is quite a bit simpler compared with going the
Trumpet Winsock route, so this is probably the best way to go assuming
that you are a dialup IP user.

However, be aware that Chameleon Sampler can cause problems if you
attempt to install it on a system that already has a version of TCP/IP,
such as one running Microsoft WFW TCP/IP-32. The installer does not
have an "applications only" option, which is unfortunate.  

Lately, some books are bundling Spyglass Mosaic. This is a good,
solid Mosaic implementation, but not as featureful or wizzy as
second generation browsers such as Netscape or BookLink. 

A-8. What diagnostic utilities are available to find problems with
my connection?  Where can I get them?

Frequently used diagnostic utilities include ifconfig (checks the
configuration of the network interfaces), ping (tests IP layer
connectivity), traceroute (traces the route that a packet takes
between two sites), netstat (checks the routing table), tcpdump
(protocol analyzer), arp (looks at the IP to Ethernet address
mappings). Microsoft TCP/IP-32 includes versions of all of these
except for tcpdump. 

KA9Q includes ifconfig, ping and traceroute functions. In KA9Q hop
check is the equivalent of traceroute. The Trumpet TCP/IP stack also
has a hopchk2 command that is a traceroute equivalent.

Etherload is very useful for network profiling, as well as packet 
analysis. Although it can't understand RARP or DHCP, it does
handle multiple protocols (AppleTalk, IP, IPX/SPX, NetBEUI),
lots of IP protocols (ARP, BOOTP, DNS, RIP, TFTP, TCP and UDP
statistics, Telnet, FTP). It even can handle NetBIOS traffic,
which UNIX tcpdump can't. One weakness is that it doesn't do
RARP or DHCP. 

The other major diagnostic utility I use is tcpdump, running
under UNIX. However, this is a TCP/IP only diagnostic tool,
can't be used with Netware, and doesn't know diddly about
NetBIOS. 

While Etherdump can be used for packet catching, I wish it would 
do more of the work for you, along the lines of TCPDUMP. Life's too
short to spend looking at hex packet traces, so I use EtherLoad
or tcpdump instead.  

Trumpet Winsock comes with Windows implementations of Ping and Traceroute. 

A-9. Is there a CD-ROM with the software included in this FAQ?

The Packet Driver, WinSock & TCP/IP CD-ROM is available from
CDPublishing for $29.95. This includes the packet drivers of course,
but also lots of other DOS and Windows TCP/IP stuff, including
Windows Sockets applications. It also includes the text of all
the RFCs. This is now somewhat out of date (it was cut in
December 1993), but is otherwise highly recommended. 

CDPublishing, (604)874-1430, (800)333-7565, fax: (604)874-1431,
mailto:info@CDPublishing.com, ftp://ftp.CDPublishing.com/,
Gopher site: gopher://gopher.CDPublishing.com/, WWW: http://www.CDPublishing.com/

A-10. Does Windows NT support SLIP? PPP? 

The Windows NT 3.5 supports PPP (client and server) and SLIP (client), 
both including support for Van Jacobson header compression. It also
supports DHCP, and Windows Internet Name Service (WINS). 

A-11. Where can I get Microsoft TCP/IP-32? 

Microsoft has now released a 32-bit TCP/IP stack for  
Windows for Workgroups v3.11. It's easy to set up,
fast, and has worked fine for me. It supports a host of very
nice new features, including DHCP automatic configuration, WINS
name resolution, and Windows Sockets v1.1. In addition it comes
with Telnet and FTP applications. However, please note that
it does not offer SLIP or PPP support. 
The final release is now available via:
ftp://ftp.microsoft.com/PerOpSys/WFW/tcpip/tcpip32.exe 


A-12. How do I get my BBS to run over TCP/IP? 

First off, let's clarify what we mean by "over TCP/IP." This can mean
everything from "accessible via Telnet" to being a full Internet
citizen, supporting Gopher, HTML, etc.

NovaLink Professional is today the only BBS software that includes
support for HTML in mail and news. For info, contact Res Nova
Software. Softarc's FirstClass package will soon be availble on Windows NT,
and has also promised HTML support. 

eSoft's IPAD is a full fledged SMTP, NNTP, DNS, FTP, Telnet, and SLIP/PPP
server, that can be hooked up to an existing TBBS system to provide
full Internet support. It comes with hardware and costs in the neighborhood
of $4K. 

The Major BBS now runs under UNIX, and thus offers Internet support;
the DOS version now has an Internet gateway that can handle telnet,
mail, and news, among other things. 

Support for a variety of BBSes is available from Murkworks. Their
BBSNet product provides a Telnet interface that looks like a FOSSIL driver.  
The first version runs partly as an NLM; some of the code resides on the server.
For info, contact 

BBSnet,MurkWorks, Inc., P.O. Box 631,Potsdam, NY 13676, 
+1 315 265 4717, mailto:info@MurkWorks.com

For further information on running BBSes on the Internet, see The
Online User's Encyclopedia, Addison-Wesley. 

A-13. Are there graphical servers out there?

Yes! For Windows there is a graphical SMTP daemon which is not very
functional (it can't do as much as KA9Q); several Web servers, including
a Windows version of NCSA's HTTP, and SerWeb. 

For Windows NT, The European Microsoft Windows Academic Consortium (EMWAC) has
released Windows NT servers for Gopher, WAIS, and WWW. These servers
are easy to install, and fast, and offer the full complement of capabilities,
including support for forms, access to WAIS indices from within HTTPS, 
installation as a Windows NT service, etc. Highly recommended. 

See the resource section for details.

A-14. What methods of address assignment are available?

Methods of address assignment include client/server protocols
(RARP, BOOTP, DHCP), as well as script-based methods 
(terminal server indicates, "your address is 192.187.147.2"). PPP
also supports assignment of addresses from the server. 

As part 2 of this FAQ discusses, there are RARP and BOOTP clients
and servers available for DOS. Typically the clients work by stashing
the IP address in a DOS environmental variable. It is then your responsibility 
to modify the appropriate config files to reflect this 
address. This can be done using a DOS batch script and a utility such as 
DOS awk. This same approach can be used to modify config files when using
EtherPPP; this does not place the IP address into a variable, but the output
of EtherPPP can be piped to a file and the IP address picked off and inserted 
in the appropriate locations. If this sounds complicated, it is; be warned.

Trumpet Winsock supports script-based assignment of addresses. Microsoft TCP/IP
supports a DHCP client and NT Server supports a DHCP server. 
There is also a forthcoming DHCP server for Sun. 

However, be aware that these products are not always RFC compliant. For example,
RFC covers interoperability between BOOTP and DHCP. This
RFC states states how a DHCP client can use a BOOTP server to determine its
parameters, and how a BOOTP client can interoperate with a DHCP server. 
However, I am not aware of a DHCP client or server that implements these
recommendations. 

A-15. How can I set up PPP server on a UNIX host? 

This is not the appropriate place to address that question, but lots
of info on this is available in the news:comp.protocols.ppp FAQ. 

A-16. What is WinSNMP? Why doesn't my TCP/IP stack support SNMP? 

WinSNMP is an API which provides a standard interface to to the
Simple Network Management Protocol (SNMP) for network
management applications running under Windows. Applications written to 
WinSNMP can run on any WinSNMP-compatible implementation. 

Vendors supporting WinSNMP include FTP Software, which supports it
in both OnNet 1.1, and PC/TCP 3.0. SNMP agents are also available
in Windows NT, Chameleon, and other packages. 

There are also freeware WinSNMP-compliant applications. See the 
Resource Section for details. 

However, if your chosen TCP/IP stack does not support an SNMP agent,
you are probably out of luck. This is because SNMP support cannot just
be tacked on; the stack must keep the statistics, and work closely
with the SNMP agent in order to allow these variables to be read
in response to SNMP queries. Without detailed knowledge of a particular
stack's operation, it is virtually impossible to write an SNMP agent
for it. 

A-17. What HTTP proxies are available for use with NCSA Mosaic? 

Mosaic and WinWeb now both support proxies via the CERN httpd, which
supports http, ftp, gopher, and wais proxies, as well as caching. 

A-18. Why doesn't my Web browser support direct WAIS queries?

If you've been trying out WinWeb, Netscape, Booklink, Windows Mosaic, 
or Cello, you've noticed that trying to resolve a WAIS URL results
in an error. You may have checked your URL syntax over and over,
trying to figure out what you did wrong. 

Guess what? The only Web browser that supports direct WAIS queries
is XMosaic v2.4 or later. On that browser, a WAIS query will generate
a request to port 210 on the destination WAIS site. (I know, because
I've run TCPDUMP to verify this).

On other browsers, you can reach WAIS sites that have set up a
Gopher or Web server to handle queries; however, you cannot
reach them directly. 

How did this come about? Windows Mosaic v1.0 contained 
support for a WAIS gateway operating at NCSA. This gateway took 
your incoming request, and forwarded it to the destination WAIS site, 
and when the response came back, forwarded the answer to you. However,
the NCSA WAIS gateway got bogged down, so support for WAIS gatewaying
was removed in v2.0. However, since they didn't put direct WAIS
support in, an error was generated.

In my opinion, this was (and is) handled lamely. Either put up a 
reasonable error message explaining that WAIS is not supported,
or put in direct support.

A-19. What is SOCKS? What TCP/IP stacks support it?

SOCKS is a type of proxy server that listens on port 1080. Instead
of sending HTTP requests to port 80, gopher to port 70, etc. a
SOCKS-compliant stack will instead route them to port 1080 on the
SOCKS server. The SOCKS server then examines the requests and decides
if they should be allowed or denied. 

To my knowledge, Trumpet Winsock v2.0 is the only SOCKS-compliant
TCP/IP stack, and it apparently has problems with rbind, which
can get gotten around by using FTP in PASV mode. 

It would be cool to have a list of which applications support
SOCKS, and which support other proxy servers such as CERN
HTTPd. Please send me mail if you have info on this. 

A-20. How can I handle authentication on my NNTP server?

A good way to handle this is to use the AUTHINFO extensions to NNTP
which are supported by the INN server, as well as clients supporting
AUTHINFO, such as WinVN, the Trumpet newsreader (DOS and Windows versions), 
and Internews on the Macintosh. With AUTHINFO, you can automatically 
allow hosts within a known subdomain to post without authentication, 
forcing users outside this domain to input their userID and password, 
which is the same as that needed to access a POP server running on the 
same machine. With AUTHINFO, the userID is automatically placed in the posting. 

A-21. What is SlipKnot? 

SlipKnot (TM) is a shareware Web browser for MS Windows that works with ordinary UNIX
shell accounts, without requiring a SLIP or PPP connection. SlipKnot provides
a UNIX shell terminal window so that you can still use your ordinary UNIX
commands, or you can switch into Web browser mode. With SlipKnot, up to
five documents can be visible at a time; previous requests are cached. Since
SlipKnot supports threading, you can look at an existing document while a 
new one is being retrieved. SlipKnot supports saving or printing of documents, 
including embedded images. 

SlipKnot requires a mouse, Windows 3.1 or WFW with 2 MB disk space, 4 MB of memory, 
with 8 MB recommended. On the UNIX side, you will need Xmodem or Ymodem support. 

See the resource section for details. 

A-22. What is TwinSock?

TwinSock is a freeware Winsock proxy client implementation. When using
TwinSock, you need not assign an official IP address. When a Windows
application makes a Windows Sockets call, the TwinSock client passes
the request to a version of TwinSock running on the firewall host. The
firewall host will then permit or deny the request, and will pass the response
to through to the requesting client. 

For more information on TwinSock, check out news:alt.dcomp.slip-emulators,
news:comp.os.ms-windows.networking.tcp-ip, news:comp.os.ms-windows.apps.comm


B. Questions about drivers

B-1. What do I need to know before setting up SLIP or PPP?

Before setting up your SLIP or PPP connection, you should
have available the following information:

* The domain name and TCP/IP address of your host.
* Whether your TCP/IP address will be assigned statically,
  dynamically, or from the server.
* If from the server, whether you will be using RARP, BOOTP or DHCP. 
* The domain name and TCP/IP address of your machine (if you are not
  configuring the address dynamically or via BOOTP)
* The domain name and TCP/IP address of the primary and secondary
  Domain Name Server.
* The subnet mask.
* The domain name and TCP/IP address of an NNTP server.
* Whether your host supports POP, and if so, what version.
* Whether the host supports compressed or uncompressed SLIP, or PPP.
* The size of the Maximum Receivable Unit (MRU). 


Do not attempt to connect to your host before you have this
information, since it will just waste your time and money, and may
cause problems for the network.  In particular, do not attempt to
initiate a connection using a made up TCP/IP address! It is possible
that your made-up address may conflict with an existing address.  

This is probably the quickest way to get people very angry at you.

Static addressing means that your TCP/IP address will always
be the same. This makes it easy to configure your setup files.
Dynamic addressing means that the host will send you a message
containing your TCP/IP address when you log on. This can be
problematic if your software doesn't support grabbing the address
and inserting it into the setup files. If not, then you may have
to edit your setup files every time you log on. Yuck!

Chameleon includes a version of SLIP which can handle dynamic
addressing. The most recent version of Novell's Lan Workplace for
DOS does as well. 

You can also retrieve your address using RARP, BOOTP or DHCP. RARP
is only available to those on the same LAN as the RARP server, since
it uses broadcasting. BOOTP clients can access BOOTP servers on other
LANs via BOOTP relay. DHCP is a BOOTP extension, which allows
complete configuration of a client from info stored on a DHCP server, and in
addition supports new concepts such as "address leases". Since
DHCP frames are very similar to BOOTP frames, devices supporting BOOTP relay 
will also support DHCP relay. 

Of course, for DHCP or BOOTP to work, you will need to set up a DHCP
or BOOTP server. DHCP servers are available for UNIX, and Windows NT;
BOOTP servers are available for UNIX (BOOTPD, from CMU).

PPP also supports server assignment of TCP/IP addresses.


B-2. How do I configure SLIPDIAL?


From Ashok Aiyar, mailto:ashok@biochemistry.bioc.cwru.edu:

PHONE Script Files:

PHONE comes with several scripts (for various modems) and for the
University of Minnesota Terminal server built into it.  The command
PHONE WRITE writes these scripts to an ASCII file called PHONE.CMD,
which can be edited to your needs (modem and slip server)

The documentation that accompanies PHONE, provides good instructions on
writing script files to get PHONE to dial SLIP servers other than
the University of Minnesota server.  For example here is a script
that I use to dial a CISCO server at the University that I attend.

Background:  To start a SLIP connection, I dial our terminal server,
and login with a username and password.  After doing so, I start a SLIP
session with the following command "slip username-slip.dialin.cwru.edu",
followed by my password -- again.

Here then is the relevant portion of the PHONE.CMD script file -
#
# CWRU-TS2 SLIP login script by Ashok Aiyar 3/26/93
# Last revised 3/28/93
Procedure    Host.CWRU.Login
TimeOut 60      'CWRU-TS2 terminal server is not responding'
Message         "CWRU-TS2 SLIP login script -- Version 1.1"
Message         'Waiting for SLIP server to respond'
Quiet ON
Expect 'Verification'
Message         'Request for User Verification Received from CWRU-TS2'
Message         'Sending your user name and password'
Quiet OFF
Expect   'Username:'
Send '%u<'
Expect   'Password:'
Private
Send '%p<'
Reject    'Access denied'   'Your user name or password was not accepted'
TimeOut 30    'SLIP server did not respond to your validation request'
Expect 'CWRU-TS2>'
Send 'SLIP<'
TimeOut 10    'SLIP server did not respond to SLIP command'
Expect 'IP hostname or address:'
Send '%u-slip.dialin.cwru.edu<'
TimeOut 10 'SLIP server did not respond to hostname'
Reject    'Bad IP address'   'Incorrect Hostname'
Expect 'Password:'
Send '%p<'
Reject    'Access denied'    'Password not accepted.'
TimeOut 10
Expect 'Header Compression will match your system'
Message 'Login to CWRU SLIP server successful'
Wait 1.0
EndProcedure   Host.CWRU.Login
#
#
Procedure      Host.CWRU.LogOut
# Nothing special needs to be done to logout
EndProcedure   Host.CWRU.LogOut
#
#   End of Script file
#
----------------------------------------------------------------------
How to use packet drivers other than UMSLIP with PHONE?

The quick answer -- there is no "clean" way.  Below is a batch file
hack that I wrote to use PHONE with other packet drivers.  In this
example, the packet driver is Peter Tattam's CSLIPPER.  To use a
batch file like this, you must know the parameters with which you
plan to use the packet driver -- i.e interrupt vector, baud rate,
port address, and IRQ.  This batch file requires UMSLIP.COM,
CSLIPPER.EXE, and TERMIN.COM to be in the same directory
or in your path ...

All that the BATCH file does is to let you dial the SLIP connection
using PHONE, load the appropriate packet driver, hangup the
connection, and unload the driver when you are done ...

-- being CWRUSLIP.BAT --
@echo off
rem   this batch file is an ugly hack of U. of Minn. "SLIP.BAT"
rem   awaiting a version of C/SLIPPER that can directly interact
rem   with PHONE
rem   CWRUSLIP.BAT file is used with PHONE.EXE to start a SLIP
rem   connection on CWRU-TS2
rem   last modified 3/28/93 -- Ashok Aiyar

@echo off
cls
goto start

:start
if %1. == ?.         goto help
if %1. == help.      goto help
if %1. == setup.     goto setup
if %1. == dial.      goto forceD
if %1. == hangup.    goto forceH
if %1. == quit.      goto forceH
if %1. == HELP.      goto help
if %1. == SETUP.     goto setup
if %1. == DIAL.      goto forceD
if %1. == QUIT.      goto forceH
goto bogus
goto unload

:forceH
termin 0x60
umslip >nul
phone force hangup
goto unload

:slipper
termin 0x60
REM  the following line must be changed to reflect the COM port,
REM  IRQ, baud rate, and software interrupt
lh c:\packet\cslipper com1 vec=60 baud=57600 ether
goto end

:forceD
termin 0x60
umslip >nul
phone force dial
goto slipper

:setup
termin 0x60
umslip >nul
phone setup
goto help

:unload
termin 0x60
goto end

:bogus
echo %1 is not a valid command.
echo Try "cwruslip help" for a list of valid commands
echo.

:help
echo --------------------------------------------------------------
echo           Case Western Reserve University SLIP Setup
echo                  using Univ. of Minnesota PHONE
echo --------------------------------------------------------------
echo cwruslip setup     modem settings, phone number, username etc.
echo.
echo cwruslip dial      DIAL and establish the SLIP connection
echo cwruslip quit      HANGUP the phone and unload the driver
echo cwruslip help      this screen
echo.

:end
-- end CWRUSLIP.BAT --
 


B-3. How do I install packet drivers for Windows applications?

The secret is to load the packet driver, then run Windows. If you
are running Trumpet Winsock, you will also have to load WINPKT
before running Windows, as follows:

winpkt 0x60

If you are running DOS applications within a virtual DOS session
under Windows, you should load PKTMUX after your packet driver, as
follows:

PKTMUX 4 [or however many sessions you want]
WIN [load windows]
 
Then within each DOS session, load PKTDRV, the virtual packet driver.

If you are running Trumpet Winsock along with other DOS apps in a 
virtual DOS session, then you will need to load PKTDRV prior to
loading Windows, and then load WINPKT on top of it, as follows:

PKTMUX 4
PKTDRV 0x62
WINPKT 0x62
PKTDRV 0x60
WIN

TCPMAN will then find the virtual packet driver at 0x62. 


B-4. When do I need to install  WINPKT? 

You only need to load WINPKT before Windows if you have a network
card in  your computer, or are running a packet driver that simulates 
such a card, such as EtherPPP, or CSLIPPER in Ethernet simulation mode. 
If you are using Trumpet Winsock via SLIP/CSLIP, there is no need to load
WINPKT, since you can use Trumpet Winsock's built-in CSLIP driver. 

PKTMUX and WINPKT both accomplish the same thing: allowing you to
connect to a DOS packet driver running in real mode from a virtual
DOS session under Windows. PKTMUX is useful when you are running
more than one TCP/IP stack, and since it takes up more RAM and is
slower than WINPKT, you should only use it when you want to run more
than one stack at a time. If you are running only one DOS app,
or are using Trumpet Winsock, stick with WINPKT. 

James Harvey (mailto:harvey@iupui.edu) notes:
Winpkt is only useful running DOS applications with built-in TCP/IP
stacks under Windows, and for some Windows-based stacks (like the
Trumpet winsock.dll).  When an application registers with a packet
driver TSR to receive packets of a specified protocol type, one of the
things it hasto pass as a parameter to the packet driver in the call
is the address of a routine in the application that the packet driver
is to call when it has a packet to pass back to the application.  In
the case of an application running in 386 enhanced mode in a DOS shell
under Windows that is using a packet driver loaded in real mode before
Windows was loaded, the packet driver must ensure that Windows has the
application in memory when it does the callback, otherwise the callback
jumps off into space and your system locks up.  Winpkt does a Windows
system call to force the app into memory before the callback is done.

Erick Engelke (mailto:erick@uwaterloo.ca) notes:
Windows in enhanced mode uses the protected mode of the
386 CPU to create multiple virtual machines.  Winpkt tells
Windows to switch to the correct virtual machine before
trying to pass up the packet.  This reduces the chances of
Windows crashing.


B-5. How to do I run both WinQVT and ODI?

My advice is to use the Windows Sockets version of WinQVT/Net, Trumpet
Winsock, and ODIPKT. ODIPKT will allow you to run packet driver software
over ODI. You will also need to load WINPKT for Trumpet Winsock. 

The loading sequence is:

LSL [Link support layer]
NE2000.COM [or other ODI driver]
IPXODI [IPX version supporting ODI]
NETX
ODIPKT 1 96
WINPKT 0x60
WIN [run windows]

Then run Trumpet Winsock, and load WinQVT/Net. 

B-6. Is it possible to use BOOTP over SLIP?

Yes, but it is easier to use dynamic address assignment to get 
your IP address. This is where the SLIP server outputs your IP address 
before switching to SLIP. 

If you need BOOTP, then you should run a BOOTP server on the SLIP
server so that it can tell which SLIP connection originated the
request. Of course, the BOOTP server will ignore the hardware address
of the request originator, but instead will keep track of the SLIP
interface the request came in on. See the question on adding BOOTP to
KA9Q for info on how to handle this on the PC. Under UNIX, you may
have to add BOOTP capability to your slip driver, and rebuild the
kernel. (Not recommended for the squimish). 


B-7. How do SLIP drivers work? 

Some TCP/IP applications are written to only support Class 1 (Ethernet)
packet drivers, but do not support Class 6 (SLIP). For these applications, you
need software to make the application think it is dealing with a class 1
interface. This is done by adding fake ethernet headers to incoming 
SLIP or PPP packets and stripping the headers off outgoing packets. 


B-8. When do I need to install PKTMUX?

PKTMUX is needed to allow you to use more than one TCP/IP stack at the same 
time. This is useful if you have applications that require different stacks. 
Note that you do not need PKTMUX to run different protocols, since packet 
drivers only look at packets in the protocol they're designed to handle, 
and therefore you can use more than one of these at a time without conflict. 
You also don't need PKTMUX if all your applications use the same TCP/IP stack. 

PKTMUX works by looking at outgoing datagrams, and caching information on 
source and destination ports and addresses. Using this information, PKTMUX
tries to sort incoming datagrams by TCP/IP stack. If it can't figure out
which stack to send a datagram to (as might be the case if you were running
a server application on a well-known port, and had not sent any outgoing
packets yet), PKTMUX will send the datagram to all stacks. If all stacks
do not complain about the datagram, PKTMUX will throw away the ensuing outgoing
ICMP error message, assuming that one of the stacks correctly received
the datagram. If all stacks complain, it will send a single ICMP message
and throw the rest away.

While PKTMUX does its job very well, there are some situations that it cannot
handle, such as port conflicts. If two applications open the same TCP port,
chaos is inevitable, and there is little that PKTMUX can do to help. 

B-9. Can NDIS be used underneath multiple protocol stacks of the same type?
No. There is no equivalent to PKTMUX for NDIS.  

B-10. Is there an NDIS over packet driver shim? 
Joe Doupnik writes:

"No. Packet Drivers work by having an application register
for a particular packet TYPE, such as 0800 for IP. NDIS works much
differently, by offering a peekahead of every packet to applications in turn,
a polling operation. The only way NDIS could gracefully sit on a PD would
be to run the Packet Driver in all-types mode and let NDIS see all pkts
not used by other clients. Needless to say, that's an undesirable situation.
The quick solution, costing about US$100 (at least at my place,
more at yours) is a second Ethernet board in the client together with a
second IP address (most important, please)."

B-11. How do I run NetBIOS over TCP/IP? 

NetBIOS over TCP/IP is discussed in RFCs 1001 and 1002, which defines
three types of NetBIOS nodes:  

* B nodes, which use UDP broadcast packets to distribute datagrams and
resolve names. 
* P nodes, which use point-to-point communications and which 
require NetBIOS Datagram Distribution (NBDD) and NetBIOS Name 
Servers (NBNS). P nodes do not listen to or use broadcast 
services, so they cannot be used alongside B nodes. Unfortunately NBNS, 
and NBDD servers were not widely implemented, and those
that do exist (such as an implementation from Network Telesystems) 
are not inexpensive.  
* M nodes, which use both point-to-point and broadcast. 

B node technology cannot be used on an IP internet without extensions, 
since UDP broadcast packets are not forwarded through routers. This 
is not a problem with use of NetBIOS over IPX/SPX, since in IPX/SPX 
broadcast packets can be forwarded. 

However, until very recently, M and P node technology was not supported 
by popular TCP/IP implementations. For example, PC/TCP supports
B node technology with extensions such as a broadcast file, host file, 
or DNS resolution of NetBIOS names. Windows NT and WFW TCP/IP uses an 
LMHOSTS file for resolving names. 

According to Chip Sparling of FTP Software:

"From what I remember from our discussions of a few years ago, P
nodes were only implemented by Ungermann Bass and 3COM (and they 
required you to use a NetBIOS name resolver which was non-rfc 1001, 1002 compliant), 
nobody did M nodes (as far as I remember) and PC-LAN, Lantastic and
LanManager used B node.  Also, if you did a P or M node it wouldn't be
compatible with a B node NetBIOS.  We decided that we could give the
compatibility and functionality (routability) with a B node plus
extensions implementation.  So, that's what we did." 

Without implementation of M and P node technology, the only way 
to run over an IP internet is to to implement B node technology 
with extensions, as FTP Software does in PC/TCP. According to Chip, 
"one way to handle large numbers of hosts on multiple networks is 
to use the broadcast file extenstion.  Instead of putting specific 
ip addresses in the broadcast file, use a subnet broadcast address 
like nnn.nnn.nnn.255. which will cover an entire subnet."

Assuming you don't need any of the extensions to RFC NetBIOS 
Microsoft created to make NetBIOS work smoothly in a routed environment 
(available only in their IP stack), you can choose from a wide variety of
commercial vendors. For example, FTP Software's PC/TCP includes RFC NetBIOS 
support; Performance Technologies has a NetBIOS that runs over packet drivers,
as does Accton (LANSoft). 

If any other vendors are reading this, I'd love to have information 
on how *you* implement NetBIOS over TCP/IP, and whether you'll be
supporting WINS, the new P-node technology name resolution service
from Microsoft. 

WINS support is included in the recent release of TCP-IP/32 which
is available for download on ftp.microsoft.com. Consult the 
release documentation for more information on this. 

Another recent development is the release of an NBNS and SMB
server for UNIX, known as Samba. Samba works great, I am using
it. See resource section for details. 


B-12. How do I run NFS along with another TCP/IP application?

The DOS NFS implementation by M. Durkin now supports a built-in
packet multiplexer that can handle NFS plus another stack loaded
simultaneously. If you need to load more stacks, then you will need
to run PKTMUX as well.

See the resource section for details. 

B-13. How do I run Trumpet Winsock along with KA9Q or NFS? 

The secret is to load WINPKT on top of the PKTDRV virtual
packet driver, if you are running PKTMUX. 

B-14. I am trying to run Netware and TCP/IP at the same time, using
      PDETHER. How do I do this?

Chris Badura (bad@flatlin.ka.sub.org ) writes:
"On one PC running odipkt over the ODI driver for the pocket ethernet
adaptor resulted in a 10x performance *decrease*.  So I switched to
running IPX/SPX over a paket driver for this adaptor wich performs
very well. The setup is like:

	pkdriver 0x60
	lsl
	pdether
	ipxodi
	netx
	winpkt 0x60

I had to get pde103.zip from netlab2.usu.edu to get IPX with Ethernet
II frameing to work.  The older pdether from simtel didn't work.
It seems also like winpkt has to be loaded last."

B-15. Sample Stick Diagrams

It has been proposed that we begin to collect some diagrams of working
combinations of hardware, drivers, shims, stacks, and applications. I'm
game, and have made a start below. If you've got some other exotic
configuration that works (or if you've tried one of the configurations below
and discovered it doesn't work, drop me a line).

   Running an individual DOS application under Windows

    NCSA telnet / DOS Trumpet / POPmail/ PC Gopher III
                 |
             DOS Session
                 |
             Windows 3.1
                 |
               WinPKT
                 |
            Packet driver or Shim
                 |
                DOS
		 |
           Network Adapter


DOS Trumpet, NCSA Telnet, and WinQVT/Net over Ethernet under Windows

                                                QVT/NET
                                                   |
     TRUMPET                    NCSA telbin        |
       |                             |             |            
     PKTDRV1                     PKTDRVn           |
       |                             |             |
     DOS Session                DOS Session    Windows Session
       +-----------+-----------------+             |
                   |                               |
                   +                               |             
             WINDOWS 3.1 .............        WINDOWS 3.1
                   |                               |
                   |                      PKTINT(QVT/NET own)
                   |                               |
                   |                           PKTDRVx
                   +-------------------------------+            
               PKTMUX n
                   |                   
          Packet Driver or SHIM
                   |              
                  DOS 
		   |
            Network Adapter

PC Gopher III, NCSA Telnet over CSLIP under Windows


                                                
                                                   
  PC Gopher III                 NCSA telbin        
       |                             |                         
     PKTDRV1                     PKTDRVn           
       |                             |             
     DOS Session                DOS Session   
       +-----------+-----------------+             
                   |                               
                   +                                            
             WINDOWS 3.1 
                   |                               
                   |                   
                   |                               
                   |                           
                   +           
                PKTMUX n
                   |                   
               CSLIPPER
                   |              
                  DOS
		   |
	      Serial Port  

PC Gopher II and NetWare on a LAN - Alternative I
[Didn't work for me, but it's supposed to be OK]

               NetWare
PC Gopher        |
  III            |
   |             |
DOS Session    NETX
   |             |
 Windows 3.1     |
   |           PDIPX
  WINPKT        /
     \         /
      \       /
       \     /
        \   /
     Packet Driver
          |
	 DOS
          |
     Network Adapter
   

PC Gopher III and NetWare on a LAN - Alternative II

                                  PC-Gopher III
				      |
				  DOS Session
				      |
			          Windows 3.1                 
                                      |
                                      |
                    NetWare            |
                        \            /
                      NETX         WINPKT
                          \        /
                        IPXODI   ODIPKT
                             \   /
                              \ /
			       |
			Link Support Layer
                               |
                            ODI driver
			       |
			      DOS
                               |
                          Network Adapter

WinQVT/Net and PC Gopher II and NetWare over a LAN - Alternative I

PC Gopher      
  III 
   |             Win QVT/Net      
 PKTDRV1            |      
   |                |   
DOS session      Windows 3.1
   |                |
Windows 3.1      PKTINT (QVT/NET own)
   |                |
   |             PKTDRVn
 WinPKT             |
   |                |          NetWare
   +----------------+            |
   |                             |
   |                             |
 PKTMUX n                      NETX
   |                             |
    \                          PDIPX
     \                           |
      \                          |
       \                         |
        \                        |
     Packet Driver --------------+
          |
	 DOS
          |
     Network Adapter

WinQVT/Net, PC Gopher III and NetWare over a LAN - Alternative II
	
	                                         QVT/Net
  PC Gopher III                 NCSA telbin        |
       |                             |             |            
     PKTDRV1        .....        PKTDRVn           |
       |           |                 |             |
     DOS Session                DOS Session    Windows Session
       +-----------+-----------------+             |
                   |                               |                                  
		   |                               |            
             WINDOWS 3.1 .......................WINDOWS 3.1
                   |                               |
                   |                      PKTINT(QVT/NET own)
                   |                               |
                   |                           PKTDRVx
              	   |	                           |      
		   |		                   |
		   |		                   |
		   |		                   |
		   +------------------+------------+
				      |
                    NetWare            |
                        \            /
                      NETX         PKTMUX n (use if >1 TCP/IP app)
                          \        /
                        IPXODI   ODIPKT
                             \   /
                              \ /
			       |
		        Link Support Layer
                               |
                           ODI driver
                               |
			  Network Adapter



PC Eudora and Windows Trumpet over CSLIP/PPP under Windows using Trumpet Winsock


 PC Eudora    Windows Trumpet
     \         /
      \       /
       \     /
        \   /
       TCPMAN
          |
     Windows 3.1
	  |
     WINPKT 0x60
          |
         DOS
          |
      Serial Port
      
PC Eudora and Windows Trumpet supporting Ethernet and CSLIP/PPP under Windows 
using NDIS supporting stack [Chameleon]

[Please note: this is not possible under Trumpet Winsock, since it can
only handle a single interface; it requires a stack that routes]

 PC Eudora    Windows Trumpet
     \         /
      \       /
       \     /
        \   /
      Chameleon NEWT
          |
      Windows v3.1
          |
	  +------------------+
	  |                  |
    Protocol Manager         |
          |                  |
      NDIS Mac Driver     Serial Port
          |
         DOS
          |
     Ethernet card



PC Eudora, Windows Trumpet, and KA9Q under Windows

       WinTrump   PC Eudora
            \     /
             \   /
 KA9Q         \ /
   |           |
 PKTDRV      TCPMAN
     \         |
      \       /
       \     /
        \   /
	 \ /
        Windows
          |
        PKTDRV 0x62
	  |
        PKTMUX 2
          |
     Packet Driver
          |
         DOS
          |
     Ethernet Card

HGopher, PC Eudora, and WinTrumpet Under Windows
(Whether the TCP/IP stack is loaded before or
after Windows depends on the stack)

       HGopher
         |       
         |
   PC    |