The 'Security Digest' Archives (TM)

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

ARCHIVE: TCP-IP Distribution List - Archives (1991)
DOCUMENT: TCP-IP Distribution List for November 1991 (207 messages, 100342 bytes)
SOURCE: http://securitydigest.org/exec/display?f=tcp-ip/archive/1991/11.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 Nov 91 00:56:06 GMT
From:      bill@uhunix.uhcc.Hawaii.Edu (William J. King)
To:        comp.protocols.tcp-ip
Subject:   TTL question





	Is there a default TTL?
	what is it set to be?
	where are these answers to be found?
	thanx
	wm.

--
    bill@uhunix.bitnet 
    bill@uhunix.uhcc.hawaii.edu 
    21.19 N.LAT by 157.52 W.LONG          
    Fish on the Byte

-----------[000001][next][prev][last][first]----------------------------------------------------
Date:      1 Nov 91 03:15:37 GMT
From:      klox@otago.ac.nz
To:        comp.protocols.tcp-ip
Subject:   slip protocols sought

We are seeking an implementation of the SLIP protocol for tcp-ip over serial
lines running under SCO Unix, SCO Xenix, and or DOS.

Also looking for RARP and BOOTP servers for same.

Any pointers, hints, names addresses etc gratefully received.

advTHANKSance

KLox

   0          
    00      Martin Cox                        
    0000    Computing Services Centre         E-mail   klox@otago.ac.nz
     00     University of Otago               Tel      +64 3 479 8531   
            Dunedin                           Fax      +64 3 479 8577 
    000     New Zealand
   000
  000       I Am A Bear Of Very Little Brain, And Long Words Bother Me
 000

-----------[000002][next][prev][last][first]----------------------------------------------------
Date:      1 Nov 91 03:59:53 GMT
From:      mdb@NSD.3Com.COM (Mark D. Baushke)
To:        comp.protocols.tcp-ip.domains,comp.protocols.tcp-ip
Subject:   Re: A good word for the new GSI/NIC

On 31 Oct 91 13:41:48 GMT, chris@visionware.co.uk (Chris Davies) said:

Chris> In article <1991Oct29.120907.1963@arizona.edu>
Chris> leonard@arizona.edu (Aaron Leonard) writes:

Aaron> But as far as I can tell, calling them up definitely *does*
Aaron> work.  Last week, I called them at 1-800-365-3642, because I

Chris> So what's an alternative number?  I can't call 1-800 numbers from the
Chris> UK, I need a "real" phone number if I want to get our record updated.

The following information was in the domain-template.txt file that I
picked up on 30 September 1991. (My copy of that form is available via
anonymous FTP from ftp.3com.com:/netinfo/domain-template.txt)

| For further information contact the DDN/INTERNET Network Information
| Center (NIC):
| 
|       Via electronic mail:  HOSTMASTER@NIC.DDN.MIL
|       Via telephone:        (800) 365-3642 or (703) 802-4535
|       Via postal mail:      DDN Network Information Center
|                             14200 Park Meadow Dr., Suite 200
|                             Chantilly, VA  22021

Enjoy!
-- 
Mark D. Baushke
mdb@NSD.3Com.COM

-----------[000003][next][prev][last][first]----------------------------------------------------
Date:      1 Nov 91 07:59:28 GMT
From:      shih@madrone.eecs.ucdavis.edu (Alan S. Shih)
To:        comp.protocols.tcp-ip
Subject:   Noivce need help in setting up SMTP


 Help, Netlanders...

 I am trying to set up SMTP on my SVR4.386 unix box. However, reading the
 manuals really confuse me (seriously) Hence I ask if someone
 can give me quick directions on setting it up.

 Thanx
-- 
|   ALAN SHIH  
|   UC.Davis EE dept
|   "When I realize my life is full of sh*t, it was good;
|      because I can start cleaning it UP."

-----------[000004][next][prev][last][first]----------------------------------------------------
Date:      1 Nov 91 11:07:51 GMT
From:      flog@bernina.ethz.ch (Florian Gutzwiller)
To:        comp.protocols.tcp-ip
Subject:   Token ring to Ethernet Bridges/Routers ?

Who can summarize experiences and recommendations ?

Thanks

-----------[000005][next][prev][last][first]----------------------------------------------------
Date:      1 Nov 91 12:45:49 GMT
From:      aw2t+@andrew.cmu.edu (Alex R.N. Wetmore)
To:        comp.protocols.tcp-ip.ibmpc,comp.protocols.tcp-ip
Subject:   Token ring driver for i286 Xenix

Does anyone know if there is a public domain token ring driver for
Xenix/286? I would like to be able to use Xenix with my IBM Token Ring
board for use with telnet, ftp, etc.  Ethernet isn't really an option
because the dorm that I am in doesn't support ethernet connections.

So is there some sort of equivelent to IBMTOKEN.COM for Xenix that I
could stick in /dev and use for networking?



--- Alex Wetmore (aw2t@andrew.cmu.edu)

-----------[000006][next][prev][last][first]----------------------------------------------------
Date:      1 Nov 91 16:35:20 GMT
From:      dc@micronics.com (Darren Croke)
To:        comp.protocols.tcp-ip
Subject:   Re: InterOp Debate

In article <10187@crackers.clearpoint.com> martillo@clearpoint.com (Martillo) writes:
>
>At InterOp, there was a debate over routing issues.  Radia Pearlman,
>Milo Medin, Dave Clarke and others took part.
>
>Pearlman was advocating the use of IS-IS routing in place of OSPF even
>though IS-IS does not guarantee the "best" possible route because she
>claimed it really did not make a difference if now and then a packet
>took a less than optimal route.
>
>If "best" is evaluated in terms of cost, such an assertion is really not
>valid, and many people would care if a significant fraction of packets
>were taking higher cost (i.e.  less than optimal routes).  Of course,
>the PTTs would not mind such non-optimal routing as it increases their
>profit. 
>

But Pearlman's argument wasn't simply based on that fact that occasional 
less than optimal routing "did not make a difference". Although I did not
attend the debate, I have read 2 articles written by her this month, and
she makes it quite clear in both that there is a tradeoff between optimal 
routing of traffic leaving level 1 areas, and the cost of maintaining 
level 2 routing information in level 1 routers. 

-- 
 Darren Croke         Micronics Computers
 dc@micronics.com     X Terminal Division
                          (415) 651-2300

-----------[000007][next][prev][last][first]----------------------------------------------------
Date:      1 Nov 91 16:57:36 GMT
From:      tcora@pica.army.mil (Tom Coradeschi <tcora>)
To:        comp.protocols.tcp-ip.domains,comp.protocols.tcp-ip
Subject:   Re: A good word for the new GSI/NIC

bob@MorningStar.Com (Bob Sutterfield) writes:
>In article <1991Oct29.204121.2590@decuac.dec.com> mjr@hussar.dco.dec.com
>(Marcus J. Ranum) writes:
>   ... CIX is in the business of providing these services ...
>
>Did the CIX, any CIX members, or ANS bid on the new NIC contract?  If
>not, then either they weren't offered the opportunity, or they decided
>they didn't want the business.  If the latter, then maybe they were
>simply wiser than GSI.  If the former, then one might rightfully ask
>why.
>
>If the CIX, a CIX member, or ANS submitted an unsuccessful bid for the
>new NIC contract, then since the contract is still being let by the US
>government, the bids and the records of the evaluation process might
>be open for public perusal.  However, I hardly know how to read the
>bureaucratese in a government services contract or a request for bids,
>so someone else who's handier with the Freedom of Information Act will
>have to do the honors.

If you get a copy of the Statement of Work, you've got the meat of the whole
contract. In our Army contracts, that's Section C (formally defined as
Description/Specs/Work Statement). All the rest is useless (at least from the
point of view of determining what the work to be performed is). If you can
get that, you should be able to read it (well, maybe with a little work).
-- 
 
                tom coradeschi    <+>    tcora@pica.army.mil

-----------[000008][next][prev][last][first]----------------------------------------------------
Date:      1 Nov 91 17:15:44 GMT
From:      rtt22@cbnewse.cb.att.com (R. Trinh Ta)
To:        comp.protocols.tcp-ip
Subject:   Dumb or Intelligent Ethernet card ?

I need to add a 386 based PC to an Ethernet network to do call processing
functions in my project. Does anyone know if TCP/IP processing takes a lot
of CPU power so that I would be better off buying an intelligent Ethernet
card instead of a dumb one ?

Intelligent cards on the market usually have a powerful processor (like the
68000) on board plus up to a megabyte of RAM, but they cost about 4 times
as much as a dumb card. Does the offloading of the TCP/IP processing onto
the Ethernet board substantially improve the system performance ?

R. Trinh Ta
AT&T Bell Labs,
Naperville, IL
(708) 979-7607
att!iexist!rtt

-----------[000009][next][prev][last][first]----------------------------------------------------
Date:      1 Nov 91 19:17:09 GMT
From:      purodha@purodha.gun.de
To:        comp.protocols.tcp-ip
Subject:   Re: Duplicate/old mailings

In article <9109191855.AA02165@NISC.SRI.COM>, 08071TCP@MSU.EDU (Doug 
Nelson) writes:
> I am receiving frequent multiple copies of letters, and letters 
> that are up to two weeks old.

This system tends to emit "very old" postings once a week, because we 
batch news only once a week by default - we feed noone, post little,
and are pretty "downstream" uucp. Off course rare postings from here 
of some actuality get batched & sent "same night" express.

-----------[000010][next][prev][last][first]----------------------------------------------------
Date:      1 Nov 91 19:17:49 GMT
From:      purodha@purodha.gun.de
To:        news.groups,comp.protocols.tcp-ip,ba.internet,dnet.inet,dnet.and.sub.general,eunet.news.group
Subject:   re: RFD:  comp.internet.{policy,routing,research}

In article <k8xn4=c@rpi.edu> Edward Vielmetti <emv@msen.com> proposes:

 comp.internet.research (moderated)   - Matters of internetworking
		interest relevant to future networks and which expand the
		knowlege and state of the art.
 comp.internet.com-priv	 - gatewayed from existing mailing list
		"commercialization and privatization of the internet" 
		(com-priv-request@uu.psi.com).

I want (and shall subscribe to) both, actually, I am on the mailing 
list already and looking out for the materials expected to be in 
comp.internet.research elsewhere, do not have the time to read 
comp.protocols.tcp-ip regularly, exactly as Ed suggested. 
I'll be an interested reader, hardly a poster to either group.

I find the names good and appropriate. 

Without knowing the charter of either ba.internet or dnet.inet, I 
recommend not much if any crossposting between them and comp.internet.* 
keeping posting of strictly local interest in the local groups and 
all others in comp.internet.*  

Also, in the course of the currently discussed sub.* and dnet.* 
reorganization, my suggestion is to eventually rename dnet.inet 
into ???.internet  (the name "???" being yet under discussion - 
"deu" is my favorite) and later substructure it like comp.internet.* 
if volume is high enough. 

As ???(deu).internet.* would be in german language, we should maybe 
have an eunet.internet group as well, for InterEunet *and* general 
postings concerning Europe, in english language?
 
Purodha

-----------[000011][next][prev][last][first]----------------------------------------------------
Date:      1 Nov 91 19:33:34 GMT
From:      rikitake@jrd.dec.com (Kenji Rikitake)
To:        comp.protocols.tcp-ip.domains,comp.protocols.tcp-ip
Subject:   Re: A good word for the new GSI/NIC

In article <RIKITAKE.91Oct30105247@jrdzzz.jrd.dec.com> rikitake@jrd.dec.com (Kenji Rikitake) writes:
>   HUH???? JUST A FIFTY-SIX KAY BPS for such an important site, and with
>   ENCAPSULATED link???? *SIGH*

I've got several email replies on this question. I didn't know MILNET
was using the state-of-the-art technology in the late 1970s :-)

>   What about ns.nic.ddn.mil, BTW?

Obviously the name server is on the same network.

-- Kenji // this is nonsense, though - putting the NIC on the NSFnet
         // is much better.
--
Kenji Rikitake // VMS/Japanese Development, DEC Japan R&D Center
Outside Japan: kenji@macrofield.org // From Japan: kenji@komaba.c.u-tokyo.ac.jp

-----------[000012][next][prev][last][first]----------------------------------------------------
Date:      1 Nov 91 20:27:28 GMT
From:      rew@hpindda.cup.hp.com (Rob Wolters)
To:        comp.protocols.tcp-ip
Subject:   Re: RFD: comp.protocols.fddi

>
> NAME:          comp.protocols.fddi
> MODERATION:    none
> DISTRIBUTION:  world


 Hi Randall

 Yes, I vote for the creation of a new group.  FDDI 
 knowledge is hard to find, it would be nice to 
 devote a list to it.  

 Cheers,

 =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
 Rob Wolters              |    rew@hpuerca.hp.com
 Network Support Engineer |    Atlanta Response Center
 Hewlett-Packard          |  * These are MY opinions *
 =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
 Home of the National League Champion Atlanta Braves!
 =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= 

-----------[000013][next][prev][last][first]----------------------------------------------------
Date:      1 Nov 91 20:56:46 GMT
From:      tcora@pica.army.mil (Tom Coradeschi <tcora>)
To:        comp.protocols.tcp-ip.domains,comp.protocols.tcp-ip
Subject:   Re: A good word for the new GSI/NIC

chris@visionware.co.uk (Chris Davies) writes:
>In article <1991Oct29.120907.1963@arizona.edu> leonard@arizona.edu (Aaron
>Leonard) writes:
>>I'd like to put in a *good* word for the new GSI/NIC.
 
>>But as far as I can tell, calling them up definitely *does*
>>work.  Last week, I called them at 1-800-365-3642, because I
>
>So what's an alternative number?  I can't call 1-800 numbers from the
>UK, I need a "real" phone number if I want to get our record updated.

OK. Here's an alternative. Use AT&T's USA Direct Plan. Dial 0800-89-0011.
This is a free call from the UK which will get you an AT&T operator in the
states. Give the operator the number you want (800-365-3642), and your call
should go through!
-- 
 
                tom coradeschi    <+>    tcora@pica.army.mil

-----------[000014][next][prev][last][first]----------------------------------------------------
Date:      1 Nov 91 21:20:39 GMT
From:      rsivaram@vela.acs.oakland.edu (SIVARAMAN R)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.domains,comp.protocols.tcp-ip.ibmpc
Subject:   networking pcs with apollo..help needed

i am new to these groups. I am working on networking a bunch of ibm pcs
with an apollo server. Right now I have a network of apollos only connected
in an ATR(Apollo token ring). My purpose is to hook up a group of ibmpcs
with an apollo server and run the application originally designed for 
apollos on these pcs using tcp-ip (Apollo gives tcp-ip support).
Could any one help me out offering me some good solution answering the
following typical questions?
1. After connecting these machines in the network using tcp-ip
do u have to run any software in all these nodes?
2. If so are there any such programs available in the market or do I have 
to write one?
3. Do i have to run x terminal emulators on all the client nodes in order
to support the motif based applications?
4.Are there any such emulators available in the market?
5.Do i have to change the environment in all the client nodes to support
 the server application(using setenv command or whatever)

thanks a lot

-siva

-----------[000015][next][prev][last][first]----------------------------------------------------
Date:      1 Nov 91 22:05:46 GMT
From:      medin@cincsac.arc.nasa.gov (Milo S. Medin)
To:        comp.protocols.tcp-ip.domains,comp.protocols.tcp-ip
Subject:   Re: A good word for the new GSI/NIC

In article <RIKITAKE.91Nov1133334@jrdzzz.jrd.dec.com>, rikitake@jrd.dec.com (Kenji Rikitake) writes:
|> In article <RIKITAKE.91Oct30105247@jrdzzz.jrd.dec.com> rikitake@jrd.dec.com (Kenji Rikitake) writes:
|> >   HUH???? JUST A FIFTY-SIX KAY BPS for such an important site, and with
|> >   ENCAPSULATED link???? *SIGH*
|> 
|> I've got several email replies on this question. I didn't know MILNET
|> was using the state-of-the-art technology in the late 1970s :-)
|> 
|> >   What about ns.nic.ddn.mil, BTW?
|> 
|> Obviously the name server is on the same network.
|> 
|> -- Kenji // this is nonsense, though - putting the NIC on the NSFnet
|>          // is much better.
|> --
|> Kenji Rikitake // VMS/Japanese Development, DEC Japan R&D Center
|> Outside Japan: kenji@macrofield.org // From Japan: kenji@komaba.c.u-tokyo.ac.jp

NASA, with the cooperation of GSI, DCA, NSF, and the SURAnet folks, should have
T1 connectivity into GSI from the FIX in 2 weeks or so.  The GSI folks have
ordered a T1 line from GSI to the FIX at SURA, and this line is due in on
11/5/91.  This will be mated to CSU/DSU's, and then connected to the NASA
Science Internet router at FIX-E, and an NSI router will be dispatched to
GSI to attach to the LAN at their end.  We should have everything up in
another week or so.  This connection is being implemented in record time
(at least for any Government organization :-)), something like a month.  

Please don't beat on me about why this wasn't done earlier.  Many of us warned
the proper people about the anticipated load.  What is important is that a 
meeting was held about 3 weeks ago and NSF agreed to install a T1 to the
GSI NIC.  For various reasons, NASA was asked to support the link, and we
gladly agreed.

I expect routing changes to be made soon and as soon as the pieces are in
place, T1 connectivity to the new NIC will be established.  NSI will operate
the link, and provide 24x7 network operations support and maint.  

Things should be better soon.  It's just as hard on us as it is on you...

						Thanks,
						   Milo

-----------[000016][next][prev][last][first]----------------------------------------------------
Date:      1 Nov 91 23:19:09 GMT
From:      reece@eco.twg.com (Reece R. Pollack)
To:        comp.os.vms,comp.protocols.tcp-ip
Subject:   Summary: VMS 5.4-3 - DEBNA, DEBNI - TCP/IP

As most of you have heard, there is a bug with the VMS V5.4-3 release
which results in problems with TCP/IP packages. Here is a summary of the
information available as of this evening (01-Nov-1991):

Q1. Is this problem specific to TCP/IP implementations?

A1. No -- this problem affects any protocol which uses Ethernet
broadcasts: TCP/IP, XNS, and AppleTalk are apparent casualties.


Q2. Does this problem affect all TCP/IP implementations?

A2. Yes -- the bug is in the VMS ETDRIVER, not in the TCP/IP
implementations. There is nothing that the TCP/IP vendors can do to
correct the problem; however, we have been working with DEC to help
resolve the problem.


Q3. Does this problem affect all Ethernet adaptors?

A3. No -- the problem appears to be specific to the DEBNA device.
Contrary to earlier reports, it appears the DEBNI device does not suffer
from this problem. Further, I have no indications that the problem
extends to any other Ethernet adaptors, such as the DELQA.


Q4. What is the bug, and how does it manifest itself?

A4. Under VMS V5.4-3, the DEBNA fails to receive packets transmitted to
the Ethernet broadcast address (FF-FF-FF-FF-FF-FF). Since this is the
mechanism the Address Resolution Protocol (ARP) uses to translate IP
address to Ethernet addresses, this prevents other nodes from reaching
the DEBNA-equipped VAX.


Q5. Is there a bug fix available?

A5. Net yet, although I understand a bug fix is being tested and should
be available shortly.

--
Reece R. Pollack
Senior Software Engineer
The Wollongong Group, Inc.

-----------[000017][next][prev][last][first]----------------------------------------------------
Date:      2 Nov 91 07:14:10 GMT
From:      nirad@newdelphi.ceu.uq.oz.au (Nirad Sharma)
To:        comp.protocols.tcp-ip,comp.unix.programmer
Subject:   getservent name server ?

To the best of my knowledge,  getservent consults the local /etc/services
file and returns the info.  What if a process on another host wants to glean
this info from a particular host ?  Is there such a function ?
--
Nirad Sharma  (nirad@ceu.uq.oz.au)			Phone : (+61 7) 365 7575
Continuing Education Unit				Fax :	(+61 7) 365 7099
The University of Queensland.  QLD  4072.  AUSTRALIA

-----------[000018][next][prev][last][first]----------------------------------------------------
Date:      2 Nov 91 07:25:02 GMT
From:      nirad@newdelphi.ceu.uq.oz.au (Nirad Sharma)
To:        comp.protocols.tcp-ip,comp.unix.programmer
Subject:   References / Source for sockets / daemons

I am going to be writing a network daemon soon and,  as I understand it,  will
be employing sockets to do the job.  Does anyone have any good articles / books
/ source code examples that I could look at ?  I am aware of the bsd networking
code being available on uunet.uu.net & munnari.oz.au (closer to me) which I
will be looking into.

Thanks for any help.
--
Nirad Sharma  (nirad@ceu.uq.oz.au)			Phone : (+61 7) 365 7575
Continuing Education Unit				Fax :	(+61 7) 365 7099
The University of Queensland.  QLD  4072.  AUSTRALIA

-----------[000019][next][prev][last][first]----------------------------------------------------
Date:      2 Nov 91 07:32:43 GMT
From:      nirad@newdelphi.ceu.uq.oz.au (Nirad Sharma)
To:        comp.unix.programmer,comp.protocols.tcp-ip
Subject:   Host / User authentication in Daemons -  Recommendations ?

Are there any libraries available that allow an application,  more specifically
a network daemon,  to authenticate any incoming connections to it ?  I was
looking for a scheme that had the following features :

	.	any user at specific host / IP address

	.	specific user at specific host / IP address

	.	any host in specific domain e.g. .ceu.uq.oz.au

	.	any host in subnet e.g. 130.102.130.xx


I'm a little new to daemon programming so this may be somewhat unreasonable but 
I can't see why it would be .  If nothing is available then I'm quite willing
to write it, though I didn't see the point in (possibly) reinventing the wheel.

If it matters,  this is for SunOS 4.1.1.
--
Nirad Sharma  (nirad@ceu.uq.oz.au)			Phone : (+61 7) 365 7575
Continuing Education Unit				Fax :	(+61 7) 365 7099
The University of Queensland.  QLD  4072.  AUSTRALIA

-----------[000020][next][prev][last][first]----------------------------------------------------
Date:      2 Nov 91 11:52:53 GMT
From:      pen@lysator.liu.se (Peter Eriksson)
To:        comp.unix.programmer,comp.protocols.tcp-ip
Subject:   Re: Host / User authentication in Daemons - Recommendations ?

nirad@newdelphi.ceu.uq.oz.au (Nirad Sharma) writes:

>Are there any libraries available that allow an application,  more specifically
>a network daemon,  to authenticate any incoming connections to it ?  I was
>looking for a scheme that had the following features :
 
>	.	any user at specific host / IP address
 
>	.	specific user at specific host / IP address
 
>	.	any host in specific domain e.g. .ceu.uq.oz.au
 
>	.	any host in subnet e.g. 130.102.130.xx

Nothing that is in _common_ use. You could always use the protocol specified
in RFC931 and either hope that the sysadmins on the other side has installed
on of the authd daemons available or install it yourself. This isn't a fool-
proof 'authentication' service though. (If the user on the other side is
using a PC he could fake that information to be anything he wants, however
most machines are more secure and require root access to be able to break
this protocol.) Another authentication service is to go to the Kerberos
universe. But that is a major investment in time, software and machinery...

RFC931-daemons are available from a number of sites. The one I wrote is available
from 'ftp.lysator.liu.se' in 'pub/net'. It works on Suns running SunOS 4.0.3
or 4.1.1 and Sequent Balances running Dynix 2.1.1 (haven't tested it on other
systems yet). Dan Bernstein has written a RFC931-daemon too and his package
includes a C-interface library for querying these daemons. 

>I'm a little new to daemon programming so this may be somewhat unreasonable but 
>I can't see why it would be .  If nothing is available then I'm quite willing
>to write it, though I didn't see the point in (possibly) reinventing the wheel.
 
>If it matters,  this is for SunOS 4.1.1.
>--
>Nirad Sharma  (nirad@ceu.uq.oz.au)			Phone : (+61 7) 365 7575
>Continuing Education Unit				Fax :	(+61 7) 365 7099
>The University of Queensland.  QLD  4072.  AUSTRALIA
--
Peter Eriksson                                              pen@lysator.liu.se
Lysator Computer Club                             ...!uunet!lysator.liu.se!pen
University of Linkoping, Sweden                                  "Try LysKOM!"

-----------[000021][next][prev][last][first]----------------------------------------------------
Date:      2 Nov 91 14:39:22 GMT
From:      edwin@cs.ruu.nl (Edwin Kremer)
To:        comp.protocols.tcp-ip,comp.unix.programmer
Subject:   Re: References / Source for sockets / daemons

In <nirad.689066702@newdelphi> nirad@newdelphi.ceu.uq.oz.au (Nirad Sharma)
writes:

   | I am going to be writing a network daemon soon and, will
   | be employing sockets to do the job. Does anyone have any good articles /
   | books / source code examples that I could look at ?

I've always enjoyed Jim Frost's easy to read Socket-Primer. Gives you  a
neat  introduction, shows some code fragments and is a pleasure to read.
There's both a troff and a LaTeX version of the primer;  here's  how  to
get a copy of it:

________________________________________________________________________
  We, Computer Science department, Utrecht University,  are  running  an
anonymous  FTP  server  on  one  of  our systems. In addition to the FTP
service we're also running a mail  server,  for  those  of  you  without
direct Internet access.


--> How to get 'socket-primer' via anonymous FTP:

	Site:		archive.cs.ruu.nl  [131.211.80.5]
	Login:		"anonymous" or "ftp"
	Password:	your own email address (you@your_domain)
	File:		DOC/socket-primer.shar.Z


--> How to get 'socket-primer' via e-mail from our mail-server:

    NOTE: In the following I have assumed that your mail address is
	  "fred_flintstone@stone.age.edu"; of course you must substitute
	  your own address for this.
	  ** PLEASE USE VALID DOMAIN ADDRESSES. DO NOT USE ADDRESSES **
	  ** WITH ! and @ MIXED !!!! BITNETTERS USE USER@HOST.BITNET **


    Send the following message to
		mail-server@cs.ruu.nl
    or the old-fashioned path alternative
		uunet!mcsun!hp4nl!ruuinf!mail-server


      begin
      path fred_flintstone@stone.age.edu (SUBSTITUTE *YOUR* ADDRESS)
      send DOC/socket-primer.shar.Z
      end


The path command can be deleted if we receive a valid from address in your
message. If this is the first time you use our mail server, we suggest you
first issue the request:

      send HELP


  A complete "ls-lR" listing of the archive is  kept  in  the  top-level
directory, it will be updated every night. To get it, issue the command:

     send ls-lR.Z



  That's all for now. If you encounter problems using  the  FTP  service
and/or the mail-server, feel free to drop me a line (by e-mail, please).


			regards,
						--[ Edwin ]--
-- 
Edwin Kremer, systems and network administrator.  [NIC-Whois handle: EHK3]
     Department of Computer Science,  Utrecht University,  The Netherlands
     Email: edwin@cs.ruu.nl  | UUCP to: ...!uunet!mcsun!hp4nl!ruuinf!edwin

-----------[000022][next][prev][last][first]----------------------------------------------------
Date:      2 Nov 91 15:14:14 GMT
From:      barrett@daisy.ee.und.ac.za (Alan P Barrett)
To:        comp.protocols.tcp-ip
Subject:   Re: TTL question

In article <15519@uhccux.uhcc.Hawaii.Edu>,
bill@uhunix.uhcc.Hawaii.Edu (William J. King) writes:

> Is there a default TTL?
> what is it set to be?
> where are these answers to be found?

RFC 1122 "Requirements for Internet Hosts -- Communication Layers"
section 3.2.1.7 "Time-to-Live: RFC-791 Section 3.2"
says:

    The current suggested value will be published in the "Assigned
    Numbers" RFC.

RFC 1060 "Assigned Numbers" says:

    The current recommended default TTL for the Internet Protocol (IP)
    RFC-791 [45,105] is 32.

--apb
Alan Barrett, Dept. of Electronic Eng., Univ. of Natal, Durban, South Africa
RFC822: barrett@ee.und.ac.za             Bang: m2xenix!quagga!undeed!barrett

-----------[000023][next][prev][last][first]----------------------------------------------------
Date:      2 Nov 91 17:21:32 GMT
From:      nsharma@softaid.UUCP (Nawien Sharma)
To:        comp.protocols.tcp-ip
Subject:   experiences with OpenConnect/APPC, /FTP, and /Telnet ?


I would appreciate any knowledge of experiences with those persons that have 
worked with OpenConnect/APPC, OpenConnect/FTP, and OpenConnect/Telnet by  
Mitek Systems Corp. when connecting to an AS/400.

Requirement: To run 5250 emulation from PC workstations on a Novell
network (version 2.15 and 3.11) and connect to an AS/400 via the token ring
or ethernet interface (the adapter has not been purchased yet).

One poss. solution: Run TCP/IP stack on workstations (ie. BW, FTP PC/TCP, 
etc) with TN5250 and access an RS/6000 via TCP/IP/ethernet which bridges 
to the AS/400 via token ring. The OpenConnect software runs a bridge in the 
RS/6000 and converts TCP/IP from ethernet to LU6.2 traffic over token ring.

Is performance acceptable ? Does it appear as a bottleneck ? Or is the AS/400
the bottleneck (please specify which model AS/400, number of users and 
token ring speed) ?
Is the OpenConnect software stable and robust enough for handling
errors ? Does it crash a lot ?
How satisfied are you with the support from OpenConnect or Mitek ?

Thank you.
-- 
Nawien Sharma, Soft-Aid Computer 
106 Norton Avenue, North York, Ontario, Canada M2N 4A5
(TEL): 416-512-7878 (UUCP): softaid.UUCP

-----------[000024][next][prev][last][first]----------------------------------------------------
Date:      2 Nov 91 21:59:05 GMT
From:      sljny@cc.usu.edu
To:        comp.protocols.tcp-ip,comp.unix.programmer
Subject:   Re: References / Source for sockets / daemons

In article <nirad.689066702@newdelphi>, nirad@newdelphi.ceu.uq.oz.au (Nirad Sharma) writes:
> I am going to be writing a network daemon soon and,  as I understand it,  will
> be employing sockets to do the job.  Does anyone have any good articles / books
> / source code examples that I could look at ?  I am aware of the bsd networking
> code being available on uunet.uu.net & munnari.oz.au (closer to me) which I
> will be looking into.
> 
> Thanks for any help.
> --
> Nirad Sharma  (nirad@ceu.uq.oz.au)			Phone : (+61 7) 365 7575
> Continuing Education Unit				Fax :	(+61 7) 365 7099
> The University of Queensland.  QLD  4072.  AUSTRALIA

-----------[000025][next][prev][last][first]----------------------------------------------------
Date:      2 Nov 91 23:18:37 GMT
From:      peterd@cs.mcgill.ca (Peter Deutsch)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.domains
Subject:   Re: Paying for the Revolution (was Re: Alternate NICs)

In article <1991Oct27.230743.3616@sulaco.Lonestar.ORG> merlin@sulaco.Lonestar.ORG (David Hayes) writes:
>In article <GRITTON.91Oct26152810@altar.ee.byu.edu> gritton@ee.byu.edu writes:
>>   Now comes the hard part: The U.S. government will not want to give up its
>>position of power in the internet. .  .  .
>
>Now for the hard part: Cost. .  .  .

Now comes the hard part - I was going to say implementing
charging mechanism on the Internet, but that's not
actually it. The hard part is to get the existing users to
accept that they really pay for services so they'd be
willing to pay for this one, too. Then we'd get rid of the
prejudice against commercial traffic on the Internet and
see the growth of services, and not just wires and routers.

I've sung this song before, so I'll keep it short, but
it's not enough to say people should institute chargeable
services such as a commercial NIC. How would you propose
it gets paid for? Have the regionals pay? That would
surely mean a surcharge on network connections and I've
seen no move by regionals to pay for such services
(and I should know, being the guy most responsible for
trying to find support to keep archie alive. Donations and
contributions to date: $0.00, although lots of good
volunteer work by many people).

How about a per use fee? Mention this to many people on
the net and you'll have your secret decoder ring stripped
from your typing finger and be drummed out of Usenet in no time.

I'd say what the net really needs now is a mechanism for
charging users for services on a per use basis, and the
willingness of the user community to pay those fees.
Technically it's not that tough, but culturally it's a
real problem. If you had that, commercial NICs would
spring up overnight. Half would of course go bankrupt in
no time, but hey, that's how it works... :-)


>One problem holding up commercial internet services is this: One of the best
>resources available on the net is the FTP archives. Commercial members who
>have not been "blessed" by the goverment can't get access to these sites.
>They can join Alternet, or PSInet, or CERFnet, but those commercial IP
>providers are not permitted to pass traffic from non-blessed sites to 
>the government-funded portions of the Internet.

There is a movement to connect the various commercial
providers to each other, to allow bypassing the NSFnet
backbone (I saw a picture at Interop that showed three out
of four talking to each other). There is nothing stopping
someone shadowing useful stuff onto a commercially
connected site now. Heck, there's a ton of stuff at UUnet
already, although the financial viability of UUnet
providing this service free to all the other commercial
service providers as the commercial net grows is
questionable to say the least.

>Perhaps if each of the NSF regionals would buy a T-1 span to one or more of
>the commercial providers, there would be a better market for commercial
>users interested in TCP/IP.

In the long term I would like to U.S. government to start
thinking of the backbone as something akin to the highway
system, and not have them worry about whether the traffic
is commercial or private. Then, maybe we wouldn't need all
this navel gazing about whether certain traffic is legal
or not. Sigh.

				- peterd

-----------[000026][next][prev][last][first]----------------------------------------------------
Date:      3 Nov 91 01:45:35 GMT
From:      mycroft@gnu.ai.mit.edu (Charles Hannum)
To:        comp.protocols.tcp-ip,comp.unix.aix
Subject:   Re: arp cache aging question


> 1) Which arp cache deletion method is correct (aging
> vs. no response) ?
> 
> 2) Is there a problem in the IBM arp implementation ?

I haven't extensively tested it, but arp entries are supposed to be
deleted periodically, by default 20 minutes.  Look up 'no' in the
Info Abhorer.

-----------[000027][next][prev][last][first]----------------------------------------------------
Date:      3 Nov 91 09:48:03 GMT
From:      nirad@newdelphi.ceu.uq.oz.au (Nirad Sharma)
To:        comp.unix.programmer,comp.protocols.tcp-ip
Subject:   Re: Host / User authentication in Daemons - Recommendations ?

nirad@newdelphi.ceu.uq.oz.au (Nirad Sharma) writes:

>Are there any libraries available that allow an application,  more specifically
>a network daemon,  to authenticate any incoming connections to it ?  I was
>looking for a scheme that had the following features :
 
>	.	any user at specific host / IP address
 
>	.	specific user at specific host / IP address
 
>	.	any host in specific domain e.g. .ceu.uq.oz.au
 
>	.	any host in subnet e.g. 130.102.130.xx

More specifically,  I mean authenticating incoming connections to THAT
particular daemon,  not for the system in general.
--
Nirad Sharma  (nirad@ceu.uq.oz.au)			Phone : (+61 7) 365 7575
Continuing Education Unit				Fax :	(+61 7) 365 7099
The University of Queensland.  QLD  4072.  AUSTRALIA

-----------[000028][next][prev][last][first]----------------------------------------------------
Date:      3 Nov 91 12:55:35 GMT
From:      frank@rniwi.rni.sub.org (Frank Mogaddedi)
To:        comp.protocols.tcp-ip
Subject:   Re: Token ring to Ethernet Bridges/Routers ?

In <1991Nov1.110751.11915@bernina.ethz.ch> flog@bernina.ethz.ch (Florian Gutzwiller) writes:

>Who can summarize experiences and recommendations ?
>Thanks

Well, we are using a 386 w/ 8 MB Ram, an NE2000 and an IBM 16/4 addapter.
The whole thing works with Novell 3.11

Runs nice

Bye
	Frank
-- 
+-----------------------------------------------------------------------------+
| e-mail: frank@rniwi.rni.sub.org | Disclaimer: "Read my lips" (George Bush)  |
|       If there are aliens watching us, why don't we hear them laughing?     |
|	  Beauty is only skin-deep, but ugliness goes clean to the bone	      |

-----------[000029][next][prev][last][first]----------------------------------------------------
Date:      3 Nov 91 23:31:32 GMT
From:      jik@athena.mit.edu (Jonathan I. Kamens)
To:        comp.protocols.tcp-ip,comp.unix.programmer
Subject:   Re: getservent name server ?

In article <nirad.689066050@newdelphi>, nirad@newdelphi.ceu.uq.oz.au (Nirad Sharma) writes:
|> To the best of my knowledge,  getservent consults the local /etc/services
|> file and returns the info.  What if a process on another host wants to glean
|> this info from a particular host ?  Is there such a function ?

No.  However, if you happen to be using a version of UNIX whose vendor has
included support for Hesiod, you can use hesiod functions to retrieve service
information, assuming that you are running a Hesiod nameserver at your site
and that you have a list of services loaded into it.

I assume that YP (oops, excuse me, NIS) has similar functionality, and in fact
NIS is probably more widespread than Hesiod.

(I just mentioned Hesiod first since it's an Athena thing. :-)

-- 
Jonathan Kamens						jik@MIT.Edu
MIT Information Systems/Athena		    Moderator, news.answers
	      (Send news.answers-related correspondence
		  to news-answers-request@MIT.Edu.)

-----------[000030][next][prev][last][first]----------------------------------------------------
Date:      4 Nov 91 06:17:13 GMT
From:      rpw3@rigden.wpd.sgi.com (Rob Warnock)
To:        comp.protocols.tcp-ip
Subject:   Re: RFD: comp.protocols.fddi

rew@hpindda.cup.hp.com (Rob Wolters) writes:
+---------------
| > NAME:          comp.protocols.fddi
|  Yes, I vote for the creation of a new group. FDDI knowledge is hard to find
+---------------

As I pointed out to the original proposer, such a group, if created,
should definitely NOT be called "comp.protocols.fddi", but rather
"comp.dcom.lans.fddi", since FDDI is not a "protocol" in the sense
of the "comp.protocols.*" groups (which are network layer and higher).

FDDI is a specific kind of LAN, and thus belongs under the "comp.dcom.lans"
tree, along with already existing groups such as "comp.dcom.lans.v2lni"
(Proteon token ring) and "comp.dcom.lans.hyperchannel" (NSC Hyperchannel).


-Rob

-----
Rob Warnock, MS-1L/515		rpw3@sgi.com
Silicon Graphics, Inc.		(415)335-1673
2011 N. Shoreline Blvd.
Mountain View, CA  94039-7311

-----------[000031][next][prev][last][first]----------------------------------------------------
Date:      4 Nov 91 09:05:27 GMT
From:      ojr@itk.unit.no (Ornulf Rodseth)
To:        comp.protocols.tcp-ip
Subject:   Redundant Ethernet and TCP/IP


Have anyone had experience with redundant Ethernets and TCP/IP. How
should this be handled?

1. On hardware layer: Switching interfaces at intervals, using same
   ethernet address for both interfaces etc...

2. On IP layer, I guess this is difficult ...

3. On TCP layer, Rewrite of tcp software, two different IP packets
   belong to same socket ...

4. On application layer, using two sockets and handling redundancy by
   sending and receiving on both.

I haven't though much about this, but I would appreciate any comments,
suggestions, real experiences etc anyone might have.

Please mail me directly and I'll summarize if I get anything ...

Regards

--
Ornulf Jan Rodseth M.Sc.	ojr@itk.unit.no
SINTEF Automatic Control	+(47 7) 594351 (direct) / 594375 (switchboard)
N-7034 TRONDHEIM, NORWAY	+(47 7) 594399 (fax)

-----------[000032][next][prev][last][first]----------------------------------------------------
Date:      4 Nov 91 13:04:58 GMT
From:      kes@FTP.COM (Keith Savage)
To:        comp.protocols.tcp-ip
Subject:   Re: INTEROP debate


    
    
    The October 1991 issue of ConneXions--The Interoperability Report
    carried two articles with "background reading" for the Great IGP
    Debate, one written by Radia Perlman (yes, that is how you spell her
    surname) and Ross Callon, the other (on OSPF) written by Milo Medin.
    
    I have plenty of left over copies of this particular issue and will
    attempt to accomodate any request for copies. Please send me your
    postal address in a form which I can use as a simple cut-and-paste
    label, e.g.,:
    

Keith Savage
FTP Software
26 Princess Street
Wakefield, MA 01880-3004


617 246 2920

-----------[000033][next][prev][last][first]----------------------------------------------------
Date:      4 Nov 91 14:43:35 GMT
From:      brent@uwovax.uwo.ca (Brent Sterner)
To:        comp.os.vms,comp.protocols.tcp-ip
Subject:   Re: Summary: VMS 5.4-3 - DEBNA, DEBNI - TCP/IP

In article <1991Nov1.231909.25398@eco.twg.com>, reece@eco.twg.com (Reece R. Pollack) writes:
> As most of you have heard, there is a bug with the VMS V5.4-3 release
> which results in problems with TCP/IP packages. Here is a summary of the
> information available as of this evening (01-Nov-1991):
> 
> Q3. Does this problem affect all Ethernet adaptors?
> 
> A3. No -- the problem appears to be specific to the DEBNA device.
> Contrary to earlier reports, it appears the DEBNI device does not suffer
> from this problem. Further, I have no indications that the problem
> extends to any other Ethernet adaptors, such as the DELQA.

   OK, I thought I had a DEBNA.  SYSGEN "show/bus" indictes a DEBNI.
   I plan to call FS.  Before I do, should I trust SYSGEN?  b.
--
Brent Sterner                       Manager, Academic Technical Support
Fast:  <BRENT@uwo.ca>               <BRENT@UWOVAX.BITNET>
       <129.100.2.13>               Telephone  (519)661-2151 x6036
Slow:  Computing & Communications Services, Natural Sciences Building
       The University of Western Ontario, London, Ontario, Canada  N6A 5B7

-----------[000034][next][prev][last][first]----------------------------------------------------
Date:      4 Nov 91 14:49:41 GMT
From:      pete@sinix.UUCP (Pete Delany)
To:        alt.sys.sun,info.sun-managers,info.sun-nets,comp.protocols.snmp,comp.info.snmp,comp.protocols.tcp-ip,comp.windows.news
Subject:   Re: Running Sun Net Manager on a non-Sun.

In article <1991Oct22.113453.27037@sol.acs.unt.edu> kev@sol.acs.unt.edu (Mullet Kevin Wright) writes:
>
>Are there any brave souls out there who are running Sun Net Manager
>on a non-Sun workstation?  Is there a reason why this shouldn't be done?
>Any and all experiences and opinions welcome.

I ported Sun's OSI stack to the Siemens/Nixdorf Targon-35 (a.k.a. Pyramid) 
about two years ago, including what existed of the Network Manager at that time.
The only major problem with portability was in the IEEE 802.3 code and the
Pyramid C compiler padding the ethernet header structures to short boundries.

I thought the NNTP code  was much better than the code we contracted the 
Frahnhoffer Institute to do; though I'm sure others would differ.  At that 
the time of the Enterprise Event only the server code was complete, so I 
focused more on the X25 and FTAM stuff.  Though I was able to get the server up,
the differences between the european flavor and that of the states precluded 
by testing it locally; and political problems prevented me from testing it 
with the members outside the CNMA project. 

I thought Mick Cox et. al. did a really nice job and highly recomend it.

I also would be interested in hearing about the experiences of others.

I also did a NeWS 1.0 nntp windows front end that displayed tcp/ip, NFS,
and nntp network statistics in subwindows that the user could select from
a list what were presented for the various protocols.  It was only about 
80k of postscript that I hacked from a Stock Market Demo called 'Tycoon'.
It was a smash at the Hanover Fair, however the nntp part left a lot to 
be desired.  If anyone is interested in it I can probably e-mail it or 
if substantial interest exist I could post a copy to the comp.windows.news 
news group; but I seriously doubt it.
-- 
E-Mail: pete@athen.xs2-3.sto.sni.de || 	pete@sinix.uucp || pete@athen.uucp
Phone: 	+49 (89) 636-49432	    || Fax: 636-48976   || Home: (49) 8671-2608 
Mail: 	Pete Delaney, Rm: 46-477; Siemens; Otto-Hahn-Ring 6; 8000 Munchen 83;

-----------[000035][next][prev][last][first]----------------------------------------------------
Date:      4 Nov 91 16:53:58 GMT
From:      neerma@cod.NOSC.MIL (Merle A. Neer)
To:        comp.protocols.tcp-ip
Subject:   Re: Anyone using Timeplex's TIME/LAN 100 Routers?

In article <1991Oct31.230614.20936@cs.ucla.edu> chinson@chin.cs.ucla.edu (Chinson Yi) writes:
>We are looking into getting routers for Ether local area networks.
>I have heard very little about Timeplex TIME/LAN 100 Routers.
>If you have any experience with TIME/LAN 100 Routers, I would like
>to hear from you on its performance, reliability, and other related 
>issues.
>
>Thank you in advance.
>-- 
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
>       Chinson Yi (chinson@cs.ucla.edu) - (213)206-3584
>    UCLA, Computer Science Dept.  Room 3285-B Boelter Hall
><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

Hi,

We received a shipment of TIMELAN routers. Our experience so far
is they dont work. Software bug in RIP I guess.  The patch might
be on the way. The lesson is ...when the routers are delivered,
test em before the warranty expiration. On the other hand, one
wonders why would a vendor ship a  $20000 product to you knowing
it doesnt work?...no, its not money, they have the fix on a floppy
in the drawer. Being that our routers dont work, our experience
is somewhat limited.

Merle

-----------[000036][next][prev][last][first]----------------------------------------------------
Date:      4 Nov 91 16:56:34 GMT
From:      ole@Csli.Stanford.EDU (Ole Jacobsen)
To:        comp.protocols.tcp-ip
Subject:   Re: INTEROP Debate

The October 1991 issue of ConneXions--The Interoperability Report
carried two articles with "background reading" for the Great IGP
Debate, one written by Radia Perlman (yes, that is how you spell her
surname) and Ross Callon, the other (on OSPF) written by Milo Medin.

I have plenty of left over copies of this particular issue and will
attempt to accomodate any request for copies. Please send me your
postal address in a form which I can use as a simple cut-and-paste
label, e.g.,:

Joe Bloe
Infomagicland, Inc.
12345 Barn Way
Reckwille, NJ 01763

Cheers,

Ole

Ole J Jacobsen, Editor & Publisher ConneXions--The Interoperability Report
Interop, Inc., 480 San Antonio Road, Suite 100, Mountain View, CA 94040,
Phone: (415) 962-2515  FAX: (415) 949-1779  Email: ole@csli.stanford.edu


-- 
Ole J Jacobsen, Editor & Publisher ConneXions--The Interoperability Report
Interop, Inc., 480 San Antonio Road, Suite 100, Mountain View, CA 94040,
Phone: (415) 962-2515  FAX: (415) 949-1779  Email: ole@csli.stanford.edu

-----------[000037][next][prev][last][first]----------------------------------------------------
Date:      4 Nov 91 16:57:33 GMT
From:      kjd@icxn.com (Kevin J. Dunlap)
To:        comp.protocols.tcp-ip,comp.unix.sysv386,comp.unix.xenix.sco
Subject:   ppp for System V r3.2


I am looking for the sources to PPP that will run under System V Release 3.2
Has anyone ported PPP to System V?  I have both Interactive and SCO flavors
if that matters.

Send replies directly to me, my news feed isn;t always getting news to me.

Thanks,

-Kevin

------------------------------------------------------------------------
Kevin J. Dunlap                                      kjd@icxn.icxn.com
InterConnections, Inc.                               uunet!icxn!kjd
14711 NE 29th Place, Bellevue, WA  98007             206/881-5773

-----------[000038][next][prev][last][first]----------------------------------------------------
Date:      4 Nov 91 20:48:10 GMT
From:      cawilco@afterlife.ncsc.mil (Chris A. Wilcox)
To:        comp.protocols.tcp-ip
Subject:   DNS troubles


First off, let me apologize if this is not the correct group
for such a question. I looked for a group devoted to DNS-related
problems, but didn't see anything, so this was my next best
guess.

My system is a Sun 3/280, running SunOS 4.1.1, connected to MILNET
via the following arrangement:

_______  ___________                          ___________        _____
| IMP |--| Router1 |--------------------------| Router2 |--------|Sun|
-------  -----------     56 Kb/sec Serial     ----------- E-net  -----

If I try to ping the MILNET interface of Router1 from the Sun, I get
the following behavior:

(afterlife:8)% ping -s 26.29.0.172
PING 26.29.0.172: 56 data bytes
64 bytes from charon-gw.ncsc.mil (26.29.0.172): icmp_seq=0. time=39. ms
64 bytes from charon-gw.ncsc.mil (26.29.0.172): icmp_seq=0. time=2040. ms
64 bytes from charon-gw.ncsc.mil (26.29.0.172): icmp_seq=0. time=4040. ms
64 bytes from charon-gw.ncsc.mil (26.29.0.172): icmp_seq=0. time=6060. ms
64 bytes from charon-gw.ncsc.mil (26.29.0.172): icmp_seq=0. time=8060. ms
(etc.)

Why would two seconds be added to the ping every time? The nearest I
can figure is that the DNS is having a tough time resolving the
IP address back into the name charon-gw.ncsc.mil. I have noticed
that there doesn't seem to be much of a structure for resolving
reverse names in the DNS on the Internet (I can find SOA records
for in-addr.arpa requests for some networks, not for others). So,
my questions are:

1) Has anybody seen this sort of behavior for ping before?
2) Does a reverse-addressing structure exist in DNS for the Internet?
3) Anybody know how to fix the above problem?????

Any help/insights are greatly appreciated.

							Chris Wilcox

-----------[000039][next][prev][last][first]----------------------------------------------------
Date:      4 Nov 91 21:10:33 GMT
From:      wcs@cbnewsh.cb.att.com (Bill Stewart 908-949-0705 erebus.att.com!wcs)
To:        comp.protocols.tcp-ip
Subject:   Re: TCP/IP and Ethernet in real time control?

In article <OJR.91Oct23084839@rudolf.itk.unit.no> ojr@itk.unit.no (Ornulf Rodseth) writes:
]I'm not sure if this is the correct place for a question like this,
]but anyway:
	As good as any, I suppose - there are newsgroups that deal
	with real-time, but over here we have Real Experts (tm) :-)

]Do anyone have a good reference on why TCP/IP on an Ethernet is good
]enough for real time control? The main argument against Ethernet is
]the lack of determinism. The argument against TCP/IP are many;
]complexity of the supporting software, no easy to use priority, high
]overhead etc.

It depends *radically* on what you mean by real-time.
For many people, RealTime is a religion, and TCP/IP and Ethernet
come from the Heretics who don't know Token Ring when they see it,
and an overloaded Ethernet certainly reinforces that belief.

Are events happening at microsecond time quanta, or millisecond, or
100 ms, or 10 second?  How late can you be?  How deterministic are
the processes generating the events?  Do you really care more about
determinism or is "going fast" good enough?  How much horsepower
do your computers have, and how much is used up by the applications
themselves?  How big are the messages you're sending?  (TCP has high
overhead if all your messages are 1 byte long, and low overhead if
they're all 16 KB long - but if you've only got 10 packets/second,
the fact that your traffic is 63/64ths overhead is irrelevant.

Can you use UNIX, which has good TCP/IP implementations, or are you
using a real-time kernel which may not have great TCP/IP?
Do you have a real-time Unix (System V Release 4 provides some
significant real-time capabilities, as do a number of vendors such
as whatever Masscomp/Concurrent is named these days.)

TCP/IP and UNIX are probably not good enough for 1 ms deterministic systems.
They're more than good enough for 1 second sloppy systems,
and probably good down in to the 10-100ms range on fast hardware.
They're often good enough for event speeds in the 1-10ms range
if you flexible about variances and can tolerate 10-100ms delays
as long as all the work gets done.

Priorities matter most on bottlenecked resources - if you have
burst periods where you need to haul 9 mb/s, you need priorities.
If your bursts have 90 kb/s, and the big packets are high priorities,
you don't.  The real non-determinism problems do come from
non-preemptability, rather than Ethernet itself - if you have a 
1 ms maximum response time, and somebody's sending a 16,000 bit packet
of low-priority data, you lose.  (If it's a big packet of HIGH-priority data,
then you lose also, but you need a faster wire to cure that.)
You can improve the problem somewhat by choosing appropriately-sized
MTUs to keep your packets small.
-- 
				Pray for peace;      Bill
#Bill Stewart +1-908-949-0705 erebus.att.com!wcs AT&T Bell Labs 4M312 Holmdel NJ
# The government borrowed another billion dollars today - like every day.
# That's >$4 they'll steal from you later.  How much did you spend on lunch?

-----------[000040][next][prev][last][first]----------------------------------------------------
Date:      4 Nov 91 21:15:58 GMT
From:      richb@kronos.com (Rich Braun)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.domains
Subject:   CIX (was Re: Paying for the Revolution)

peterd@cs.mcgill.ca (Peter Deutsch) writes:
>There is a movement to connect the various commercial
>providers to each other, to allow bypassing the NSFnet
>backbone (I saw a picture at Interop that showed three out
>of four talking to each other).

That movement is called CIX, the Commercial Internet Exchange.
Founded in April, this is a cartel consisting of CERFnet, PSINet, and
UUNET.  It appears to be driven mainly by UUNET, at least that's the
impression I get whenever a computer-related business lists its
address as Falls Church, Virginia.  (I know, there are Falls Church
businesses unrelated to UUNET: in high school I worked for the Red
Lobster there...but that's beside the point.)

I'd like to see Internet become more widespread and eclipse older
X.25-based technologies.  But I worry a bit about companies which
claim to be not-for-profit and which seem to be motivated by a desire
to be the sole provider of a given service.

Anyone else know the full details of this CIX startup?  (There's a
blurb about it in this month's _Unix Review_, page 13.)

-rich

-----------[000041][next][prev][last][first]----------------------------------------------------
Date:      4 Nov 91 22:08:16 GMT
From:      pramod@uts.amdahl.com (Pramod Sharma)
To:        comp.protocols.tcp-ip,comp.protocols.iso
Subject:   SNAP question


I have a few simple questions regarding IEEE 802 SNAP (Sub Network Access 
Protocol).

Does anyone use SNAP for sending IEEE 802.3 Ethernet packets ?

If yes, how do you indicate to the Data Link provider, whether or not to use 
SNAP in an outgoing (to the network) packet ?

Is SNAP required for FDDI ?

Thanks.

Pramod Sharma      pramod@uts.amdahl.com

-----------[000042][next][prev][last][first]----------------------------------------------------
Date:      4 Nov 91 23:08:36 GMT
From:      rick@uunet.uu.net (Rick Adams)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.domains
Subject:   Re: CIX (was Re: Paying for the Revolution)

In article <1991Nov04.211558.5167@kronos.com>, richb@kronos.com (Rich Braun) writes:
> That movement is called CIX, the Commercial Internet Exchange.
> Founded in April, this is a cartel consisting of CERFnet, PSINet, and
> UUNET.  It appears to be driven mainly by UUNET, at least that's the
> impression I get whenever a computer-related business lists its
> address as Falls Church, Virginia.  (I know, there are Falls Church
> businesses unrelated to UUNET: in high school I worked for the Red
> Lobster there...but that's beside the point.)

UUNET is donating office and secretarial space to the CIX Assoc.
UUNET is a member of the CIX. That is the only association. (E.g
The cix.org computer is in San Diego)

It's a basic way of keep costs down.

-----------[000043][next][prev][last][first]----------------------------------------------------
Date:      5 Nov 91 04:32:44 GMT
From:      vjs@rhyolite.wpd.sgi.com (Vernon Schryver)
To:        comp.protocols.tcp-ip
Subject:   Re: Redundant Ethernet and TCP/IP

In article <OJR.91Nov4100527@rudolf.itk.unit.no>, ojr@itk.unit.no (Ornulf Rodseth) writes:
> 
> Have anyone had experience with redundant Ethernets and TCP/IP. How
> should this be handled?
> 
> 1. On hardware layer: Switching interfaces at intervals, using same
>    ethernet address for both interfaces etc...
> 
> 2. On IP layer, I guess this is difficult ...
> 
> 3. On TCP layer, Rewrite of tcp software, two different IP packets
>    belong to same socket ...
> 
> 4. On application layer, using two sockets and handling redundancy by
>    sending and receiving on both.


Others have observed that the rough proposal for a draft for EARP,
to support dual-MAC FDDI stations, might be useful for this idea.
(I've called it "gatling net" when arguing for it with management.)


vjs

-----------[000044][next][prev][last][first]----------------------------------------------------
Date:      5 Nov 91 08:43:48 GMT
From:      pushp@nic.cerf.net (Pushpendra Mohta)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.domains
Subject:   Re: CIX (was Re: Paying for the Revolution)

In article <1991Nov04.211558.5167@kronos.com> richb@kronos.com (Rich Braun) writes:
>
>Anyone else know the full details of this CIX startup?  (There's a
>blurb about it in this month's _Unix Review_, page 13.)
>
>-rich

Some information is available via anonymous ftp on cix.org.
(132.249.22.41)
If you have specific questions, direct them to info@cix.org

--pushpendra
CERFnet 

-----------[000045][next][prev][last][first]----------------------------------------------------
Date:      5 Nov 91 09:49:19 GMT
From:      ojr@itk.unit.no (Ornulf Rodseth)
To:        comp.protocols.tcp-ip
Subject:   Re: TCP/IP and Ethernet in real time control?


In a previous post I asked the following question:

 Do anyone have a good reference on why TCP/IP on an Ethernet is good
 enough for real time control? The main argument against Ethernet is
 the lack of determinism. The argument against TCP/IP are many;
 complexity of the supporting software, no easy to use priority, high
 overhead etc.

I got a few replies and I also have a few references of my own. I
haven't had time to go through all the papers yet, but I guess I'll
try to post some kind of summary in the not to distant future.

Thanks to those that replied.

My reference list is currently as follows:

Adams W.M., ``LAN Performance for Distributed Manufacturing
Applications'', Proceedings of ISA 89, Philadelphia USA, pp. 693-702

Shaw W.T., ``Ethernet-Based Plant Automation Architectures'',
Proceedings of ISA 89, Philadelphia USA, pp. 669-678

Boggs D.R., Mogul J.C., Kent C.A., ``Measured Capacity of an Ethernet:
Myths and Reality'', ACM SIGCOMM Conference Proceedings, August 1988.

Sventek J., Greiman W., O'Dell M., Jansen A, ``Token Ring Local Area
Networks: A Comparison of Experimental and Theoretical Performance'',
Computer Networks 8, 1984, pp. 301-309

Zhang L., ``Why TCP Timers Don't Work Well'', ACM SIGCOMM Conference
Proceedings, August 1986.

Field B., Znati T., ``Experimental Evaluation of Transport Layer
Protocols for Real-Time Applications'', 16th Conference on Local
Computer Networks 1991, pg. 521-534.

Adams W.M., ``802.3/Ethernet LAN Performance in Process
Manufacturing'', Proceedings of ISA 90, New Orleans USA, pp.
1749-1763

Bux W., ``Local Area Subnetworks: A Performance Comparison'', IEEE
Transactions on Communications Vol COM-29 No 10, October 1981
--
Ornulf Jan Rodseth M.Sc.	ojr@itk.unit.no
SINTEF Automatic Control	+(47 7) 594351 (direct) / 594375 (switchboard)
N-7034 TRONDHEIM, NORWAY	+(47 7) 594399 (fax)

-----------[000046][next][prev][last][first]----------------------------------------------------
Date:      5 Nov 91 17:58:37 GMT
From:      wlb@admin1.memst.edu
To:        comp.protocols.tcp-ip
Subject:   Proxyarpd configuration help...Please

Hello,

    Memphis State (Class B IP network 141.225.0.0) is starting
a gradual transition from a non-subnetted network to a subnetted 
network.  We are trying to use Proxyarpd software on a SUN 4.1 BSD
system.

Main network backbone                SUN card le0         SUN card le1

 -----------------------------------------             -----------------
    141.225.1.0                       141.225.1.160       141.225.6.50
                                                       (141.225.6.0 subnet)
  (Internet router
   141.225.1.1)


     We seem to be having a problem getting the Proxyarpd software configured
so it will run.  The 4.0 SUN mods were made and the Proxyarpd table seems
to be configured correctly.  The Proxyarp program is started but disappears
with 30 - 60 seconds.  Anybody out there have any ideas where we might be
going wrong?

     BTW, I am a VMS person and the person in our Engineering department 
responsible for the SUN and that subnet is not extremely experienced 
either.

Thanks for any help or pushes in the right direction.

Bill Brown
Network Support/Postmaster
Memphis State University
wlb@memstvx1.memst.edu
wlb@memstvx1.bitnet

-----------[000047][next][prev][last][first]----------------------------------------------------
Date:      5 Nov 91 18:16:08 GMT
From:      richb@kronos.com (Rich Braun)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.domains
Subject:   Re: CIX (was Re: Paying for the Revolution)

I wrote:
>> CIX ... appears to be driven mainly by UUNET, at least that's the
>> impression I get whenever a computer-related business lists its
>> address as Falls Church

rick@uunet.uu.net (Rick Adams) writes:
>UUNET is donating office and secretarial space to the CIX Assoc.
>UUNET is a member of the CIX. That is the only association. (E.g
>The cix.org computer is in San Diego)
>
>It's a basic way of keep costs down.

Rick, I have to post this apology for the tone of my posting.  Looking
back at it, I'm ascribing all kinds of potential anti-trust behavior
to an organization about which I really know little (only that a price
quote I received some months ago was high, and that a number of local
companies have had minor problems with UUNET in the past).

Healthy skepticism might be warranted in this case (and in any business
which is related to information-handling), but I had no reason to broadside
UUNET in this way.

-rich

-----------[000048][next][prev][last][first]----------------------------------------------------
Date:      5 Nov 91 20:23:27 GMT
From:      merlyn@iWarp.intel.com (Randal L. Schwartz)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.domains
Subject:   Re: CIX (was Re: Paying for the Revolution)

In article <1991Nov05.181608.20110@kronos.com>, richb@kronos (Rich Braun) writes:
|						      (only that a price
| quote I received some months ago was high, and that a number of local
| companies have had minor problems with UUNET in the past).

In UUNET's defense, I must say that they've been more than cooperative
at giving us IP service for the last six months, and they are
noticeably better than any service we seemed to get (or not get) from
CSNET.  And the price was *very* reasonable.

Just another internetworker,
-- 
/=Randal L. Schwartz, Stonehenge Consulting Services (503)777-0095 ==========\
| on contract to Intel's iWarp project, Beaverton, Oregon, USA, Sol III      |
| merlyn@iwarp.intel.com ...!any-MX-mailer-like-uunet!iwarp.intel.com!merlyn |
\=Cute Quote: "Intel: putting the 'backward' in 'backward compatible'..."====/

-----------[000049][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 00:17:57 GMT
From:      Frank@mindlink.bc.ca (Frank I. Reiter)
To:        comp.protocols.tcp-ip
Subject:   named problem

I have an interesting time trying to get named running on our network.  There
are three machines on the network: rsoft, deep, and router.  Router is a ka9q
box and seems to be running jsut fine.  rsoft has been configured as a primary
nameserver for .rsoft.bc.ca and deep as a secondary.

Rsoft (the primary) can resolve names outside of our network but does not
recognize rsoft, router, or deep.

Deep (which gets it's infor from rsoft at boot time) DOES recognize local
machines.

What on earth could I have done to cause rsoft to be unable to lookup it's own
name even though it gives the correct info to deep?

Frank.
--
___________________________________________________________________
Frank I. Reiter                UUCP:  ...!van-bc!rsoft!frank
Reiter Software Inc.       INTERNET: frank@mindlink.bc.ca (prefered)
Surrey, British Columbia             frank@rsoft.bc.ca

-----------[000050][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 01:24:33 GMT
From:      martillo@clearpoint.com (Martillo)
To:        comp.protocols.tcp-ip
Subject:   Re: InterOp Debate

In article <1991Nov1.163520.10952@micronics.com>, dc@micronics.com (Darren Croke) writes:

> But Pearlman's argument wasn't simply based on that fact that occasional 
> less than optimal routing "did not make a difference". Although I did not
> attend the debate, I have read 2 articles written by her this month, and
> she makes it quite clear in both that there is a tradeoff between optimal 
> routing of traffic leaving level 1 areas, and the cost of maintaining 
> level 2 routing information in level 1 routers. 

Well, if this problem is the basis of Pearlman's argument, they were
debating apples and oranges.  No one should seriously worry about the
problems of IMPs (the equivalent of DECNET Level 1 routers) in designing
network routing protocols.  Technology and understanding of the related
issues has advanced just a wee bit since 1969. 

In a well designed networking architecture, the problems of
communications subnet host-to-host routing and the design of
internetwork area-to-area routing should be completely independent. 

If DECNET Level 1 routers have to maintain information about hordes of
Level 2 routers in order to route frames within an area, there may be a
wee problem with the design of DECNET.  Designing OSI or DOD IP network
routing protocols in order to compensate for the inadequacies of DECNET
level 1 routers is fairly close to the limits of folly.  A more correct
approach would be perhaps to rethink and redesign DECNET Level 1
routers.  Or better yet replace the DECNET Level 1 routers with MAC
bridges and WAN LAN transparent bridges. 

>  Darren Croke         Micronics Computers

Joachim Carlo Santos Martillo Ajami
-- 
The statements contained in this article solely represent the views of
the author and in no way do they reflect the official opinion or policy
of Clearpoint Research Corporation.

-----------[000051][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 02:32:39 GMT
From:      BDK@UNB.CA (Brian Kaye)
To:        comp.protocols.tcp-ip
Subject:   Mail systems for HP 3000

Does anyone out there know of any commercial, public domain or share-
ware SMTP mail systems for HP 3000 mini computers.

Brian Kaye
University of New Brunswick

-----------[000052][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 03:07:11 GMT
From:      dennis@gpu.utcs.utoronto.ca (Dennis Ferguson)
To:        comp.protocols.tcp-ip
Subject:   Re: InterOp Debate

martillo@clearpoint.com (Martillo) writes:
>dc@micronics.com (Darren Croke) writes:
>
>> But Pearlman's argument wasn't simply based on that fact that occasional 
>> less than optimal routing "did not make a difference". Although I did not
>> attend the debate, I have read 2 articles written by her this month, and
>> she makes it quite clear in both that there is a tradeoff between optimal 
>> routing of traffic leaving level 1 areas, and the cost of maintaining 
>> level 2 routing information in level 1 routers. 
>
>Well, if this problem is the basis of Pearlman's argument, they were
>debating apples and oranges.  No one should seriously worry about the
>problems of IMPs (the equivalent of DECNET Level 1 routers) in designing
>network routing protocols.  Technology and understanding of the related
>issues has advanced just a wee bit since 1969. 
>
>In a well designed networking architecture, the problems of
>communications subnet host-to-host routing and the design of
>internetwork area-to-area routing should be completely independent. 
>
>If DECNET Level 1 routers have to maintain information about hordes of
>Level 2 routers in order to route frames within an area, there may be a
>wee problem with the design of DECNET.  Designing OSI or DOD IP network
>routing protocols in order to compensate for the inadequacies of DECNET
>level 1 routers is fairly close to the limits of folly.  A more correct
>approach would be perhaps to rethink and redesign DECNET Level 1
>routers.  Or better yet replace the DECNET Level 1 routers with MAC
>bridges and WAN LAN transparent bridges. 

Is it just me, or is this a little bit incoherent?

(1) Who mentioned DECNET?

(2) It is OSPF which originally required that level two route summaries be
    maintained in the level one areas.  IS-IS never did.  On the next
    protocol revision the OSPF people made it possible to optionally
    reduce the level 2 routing to a default in level one areas, thereby
    mimicing the IS-IS behaviour.  I kind of prefer this decision be left
    up to the end user since there are tradeoffs between the two alternatives
    (gee, didn't Pearlman mention that?).

(3) As I understand it, in none of these cases did the level 1 areas ever
    have to keep track of level 2 routers or topology.  Routing information
    is summarized at area boundaries in both directions.

(4) Why would I want to be restricted to use bridges and spanning tree
    routing when I can have good SPF routing?  Why would I want the topology
    limitations forced on me?  Why would I want to be forced to flood
    multicast traffic throughout my level 1 area?  On the other hand, even
    if I did want to use bridges what is there about level 1 routing which
    prevents me from doing so?  This is a classic bridges-versus-routers
    tradeoff (an entirely different argument) and I'd much rather this
    decision was left to the end users (as both OSPF and IS-IS do) rather
    than pretending there is only One True Way.

Dennis Ferguson
University of Toronto

-----------[000053][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 04:27:34 GMT
From:      baw@terminator.cc.umich.edu (Brian Wolfe)
To:        comp.protocols.tcp-ip,comp.protocols.nfs,comp.dcom.lans
Subject:   DOS TCP/IP Mail package

I'm posting this for a friend, please respond to ddurbin@rpslmc.edu...




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

Help!

        Looking for a user based pc/e-mail package that runs on a pc
implementation of tcp/ip, ie. FTP, Wollongong, B&W, etc... We are currently
in an AT&T StarLAN environment and are using AT&T Starmail. Our users like
the interface and features of starmail, but we are migrating toward a
tcp/ip and NFS based network environment and Starmail doesn't work there.
I've looked at the mail veiwer that comes with FTP's PC/TCP and it is
not something I could get our users to migrate to. We want to keep our post
office Unix based, so some of the other commercial packages won't work
either.

Any help with this is greatly appreciated!

Dave Durbin

Rush Medical Center
Information Services
Advanced Technology Group

ddurbin@rpslmc.edu
(312)-942-2562



--
Brian Wolfe                    Internet: brian@rpslmc.edu
Rush Medical Center            Voice:    (312)-942-5781
Chicago, IL 60612 	       FAX:      (312)-942-2114 

-----------[000054][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 07:08:10 GMT
From:      rwedlock@isis.cs.du.edu (Robert S Wedlock)
To:        comp.protocols.tcp-ip
Subject:   arp cache & icmp frags

IP supports fragmentation... does this include ICMP? Of course its possible,
but I suspect ICMP may be something that just is never seen fragmented.

-rob

-----------[000055][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 09:23:55 GMT
From:      vjs@rhyolite.wpd.sgi.com (Vernon Schryver)
To:        comp.protocols.tcp-ip
Subject:   Re: arp cache & icmp frags

In article <1991Nov6.070810.26566@mnemosyne.cs.du.edu>, rwedlock@isis.cs.du.edu (Robert S Wedlock) writes:
> IP supports fragmentation... does this include ICMP? Of course its possible,
> but I suspect ICMP may be something that just is never seen fragmented.

Yes, at least 4.3+BSD does.
A variation of an old game--on a moderately fast workstation to a similar
one, connected with ethernet:

	ping -f -s4400 otherhost

If this doesn't collapse your ethernet, then something is wrong.

This is the 4.3 style ping with -f for flood, which goes at 100 packets/sec
or as fast as the other end responds, which ever is faster.  (I didn't come
up with this; a colleague intentionally stressing things did it.)

(By the way, the packets/sec # from `ping -f` can be increased by using
slightly sneaky buffering & buffer flushing on the output of periods (.)
and backspaces (^H).)


Vernon Schryver,   vjs@sgi.com

-----------[000056][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 12:40:05 GMT
From:      martinea@hawk.nstn.ns.ca (Michael Martineau)
To:        comp.protocols.tcp-ip,comp.protocols.nfs,comp.dcom.lans
Subject:   Re: DOS TCP/IP Mail package


Try POPmail from boombox.micro.umn.edu:/pub/POPmail.  It is a PC based
POP2/POP3 client.  You will need a UNIX based POP server but there
are lots of them floating around.

-----------[000057][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 13:53:13 GMT
From:      dmak@draper.com (David Mak)
To:        comp.protocols.tcp-ip
Subject:   Query for VT240/Tek4010/401

Mail*Link(r) SMTP               Query for VT240/Tek4010/4014 Emulator.
I apologize if this cross posting is a duplicate, but I was told my first try
had failed.

I'm trying to find a VT240 or Tektronix 4010/4014 graphics terminal emulator
(commercial) that works under the Excelan LANWorkplace software and EXOS ST215
ethernet interface on IBM PS/2 models.  If anyone knows of such a package, I
would greatly appreciate E-mail (dmak@draper.com) with the relevant info.

Thanks in advance,
David Mak
Charles Stark Draper Laboratory - Cambridge, MA
dmak@draper.com
617-258-1718

-----------[000058][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 14:39:31 GMT
From:      onno@titan.hts.hsa.nl (Onno Witvliet)
To:        comp.sys.novell,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc
Subject:   Problem installing Charon v2.0a (Novell LPR gateway)

I have some problems installing the Charon v2.0A (Novell/LPR printer Gateway).
I can't get it to work under Netware 386 v3.11

If I first login to Novell as a normal user and then start Charon, I get the
following messages:

C>charon -p p.dat -l lpd.dat
Packet driver (3C503)
Class 1 Type 12 Version 9 Extended 2
Info: Adding Queue (e000001) LUNA/unix --> triton:laser
10:12:35 Lpd-I: Adding printer testq --> queue LUNA/queue_0 (c00000e)
Info: Attaching to Queue LUNA/unix
Warning: Attach failed with error ea for Queue LUNA/unix, unlinking
Press <SPACEBAR> To Continue ...

The incoming TCP/IP print requests are handeled OK, but requests from Novell
to TCP/IP doesn't work.

If I don't login first, I get the following messages:

Packet driver (3C503)
Class 1 Type 12 Version 9 Extended 2
Info: Adding Queue (e000001) LUNA/unix --> triton:laser
10:14:16 Lpd-I: Adding printer testq --> queue LUNA/queue_0 (c00000e)
Info: Attaching to Queue LUNA/unix
Warning: Attach failed with error d3 for Queue LUNA/unix, unlinking
Press <SPACEBAR> To Continue ...

In this case neither TCP/IP nor Novell requests will work

I guess there is something wrong with the queue or print server definitions
under Novell, but I can't figure out what's wrong.

Can someone tell what's wrong?

Thanks Onno


--
Onno Witvliet, Hogeschool Alkmaar | Email: onno@titan.hts.hsa.nl
Bergerweg 200, 1817 MN Alkmaar    |

-----------[000059][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 15:18:29 GMT
From:      jbvb@FTP.COM (James B. Van Bokkelen)
To:        comp.protocols.tcp-ip
Subject:   Re: Dynamic Window Congestion Control

PC/TCP for DOS has processed Source Quench messages for a number of years
now.  When we get one, we act like the usable portion of the TCP window
is zero until another Ack arrives (see the SQID RFC).  We have neither been
praised nor flamed for this behaviour, and I've wondered if anyone noticed
it at all...

James B. VanBokkelen		26 Princess St., Wakefield, MA  01880
FTP Software Inc.		voice: (617) 246-0900  fax: (617) 246-0901

-----------[000060][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 15:18:30 GMT
From:      jbvb@FTP.COM (James B. Van Bokkelen)
To:        comp.protocols.tcp-ip
Subject:   Re: Help Request: More than One Packet Driver

Making the driver contain buffers has two problems; first, the minimum
memory usage of the driver is a lot larger, and given the nature of the
PDS, the stack has to have buffers too.  Second, shared buffer pools are
horrendously tricky to tune.  Novell put one into the LSL layer when they
designed ODI; their DOS TCP/IP uses it but Netware itself doesn't.

Saving the packet in a protocol's buffer and daisy-chaining it would
require the addition of a 3rd upcall (get_buffer, do_you_want_buffer,
return_buffer), and given the lack of operating system support for anything
on the order of a "fork process" (yes, I wish DOS were as smart as RT11 was
ten years ago), you could wind up with deadlocks and buffer starvation
when two or more stacks were active.

My own feeling about the PDS is that it is more or less what it is.  The
features that are being added in 1.10 (the draft Joe mentioned) are either
quite simple (look ahead) or break new ground (as_send_pkt).  If you want
ODI's or NDIS's feature set, use that spec; the politics of DOS device
drivers make me think that new boards of any popularity are likely to have
all three drivers before they've been out for a year....

James B. VanBokkelen		26 Princess St., Wakefield, MA  01880
FTP Software Inc.		voice: (617) 246-0900  fax: (617) 246-0901

-----------[000061][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 15:18:31 GMT
From:      jbvb@FTP.COM (James B. Van Bokkelen)
To:        comp.protocols.tcp-ip
Subject:   Re: NetBIOS on TCP/IP ?


    1)      A NetBIOS implementation that runs on top of TCP/IP?

Excelan (now Novell), Ungermann-Bass, FTP Software, Wollongong, 3Com/HP
and maybe others offer RFC 1001/1002 Netbioses for their DOS TCP/IP products.
    
            This is so we can connect PCs to UNIX workstations (SparcStations)
            for file sharing.

I think you'll need some sort of after-market software for your Sparc to
implement the server side (Netbios and SMB or whatever).  Maybe Syntax Inc.
    
    2)      A NetBIOS implementation that runs over a serial line?
 
Our Netbios runs on any of our kernels, including SLIP and PPP.  However, the
other end of the line has to be either the server itself, or something that
knows to forward the broadcast packets out onto the LAN behind it.  Note
that some Netbios-based software (e.g. LAN Manager) may need to have its
timeouts re-configured for use over slow media.  I don't know of any "RAW
Netbios over serial line" standards; there may be proprietary products.

    3)      An X11 program that acts as a PC's keyboard/display and
            runs over NetBIOS or TCP/IP?

There are many PC X servers, but no X client that I know of.

James B. VanBokkelen		26 Princess St., Wakefield, MA  01880
FTP Software Inc.		voice: (617) 246-0900  fax: (617) 246-0901

-----------[000062][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 15:36:30 GMT
From:      dale@interlan.Interlan.COM (Dale B)
To:        comp.protocols.tcp-ip
Subject:   Re: arp cache & icmp frags

In article <psupcjg@sgi.sgi.com> vjs@rhyolite.wpd.sgi.com (Vernon Schryver) writes:
>In article <1991Nov6.070810.26566@mnemosyne.cs.du.edu>, rwedlock@isis.cs.du.edu (Robert S Wedlock) writes:
>
>A variation of an old game--on a moderately fast workstation to a similar
>one, connected with ethernet:
>
>	ping -f -s4400 otherhost
>
>If this doesn't collapse your ethernet, then something is wrong.
>
>Vernon Schryver,   vjs@sgi.com

And what might that something wrong be?  It doesn't phase the systems I tried it
on, SCO 3.2.2 with TCP/IP 1.1.3f with the command being executed on a 486/25
with an AT as the target.  Doesn't bother our VAX running BSD either.  I didn't
check packet rates with a lanalyzer, but all my stuff has been tested with a
saturated net before.  So what in particular is the problem that one is
supposed to see?

Dale, dale@interlan.com
-- 
----------------------------------------------------------------------------
dale@interlan.interlan.com
An expert is someone who learns more and more about less and less until they 
know absolutely everything about nothing.  Author unknown.

-----------[000063][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 15:42:12 GMT
From:      ellozy@farber.harvard.edu (Mohamed Ellozy)
To:        comp.protocols.tcp-ip
Subject:   Weather servers (summary)

A few days ago I posted a request for information on weather servers
available over the Internet.  The one with coverage of the entire
country is the University of Michigan weather underground.  It can be
reached directly by telnetting to madlab.sprl.umich.edu at port 3000.
For those with telnet implementations that cannot select a port telnet
to hermes.merit.edu, and at the "Which Host?" prompt enter um-weather.

I did not save the names of the many kind persons who relied, though I
think I thanked them all individually.  If I did not thank anyone,
please accept my public thanks.

Mohamed

-----------[000064][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 15:44:48 GMT
From:      emv@msen.com (Ed Vielmetti)
To:        comp.protocols.tcp-ip
Subject:   Re: TTL question

In article <1991Nov02.151414.183@daisy.ee.und.ac.za> barrett@daisy.ee.und.ac.za (Alan P Barrett) writes:

   > Is there a default TTL?
   > what is it set to be?
   > where are these answers to be found?
 
  >RFC 1122 -> RFC 1060
  >RFC 1060 "Assigned Numbers" says:

       The current recommended default TTL for the Internet Protocol (IP)
       RFC-791 [45,105] is 32.

In the NSFnet, as it is currently configured, a default TTL of 32 will
not be enough under some circumstances.   It is possible for traffic
to go from Massachusetts to New Jersey by crossing the entire T3
backbone, reaching the interconnect on the west coast at San Diego,
and crossing the entire T1 backbone on its return trip.  Add in a
moderately complex regional network setup on each end and a largish
campus network and you will exceed 32.

I don't know the schedule for producing a new Assigned Numbers, but I
fully expect that this default TTL will be increased when the new
document is released.

-- 
Edward Vielmetti, vice president for research, MSEN Inc. emv@msen.com
       MSEN, Inc. 628 Brooks Ann Arbor MI 48103 +1 313 741 1120

-----------[000065][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 15:52:40 GMT
From:      rikitake@jrd.dec.com (Kenji Rikitake)
To:        comp.protocols.tcp-ip
Subject:   Re: CIX (was Re: Paying for the Revolution)

In article <1991Nov5.202327.23552@iWarp.intel.com> merlyn@iWarp.intel.com (Randal L. Schwartz) writes:
>   In UUNET's defense, I must say that they've been more than cooperative
>   at giving us IP service for the last six months, and they are
>   noticeably better than any service we seemed to get (or not get) from
>   CSNET.  And the price was *very* reasonable.

And I should mention that UUNET provides the same service as well to
non-US customers. I had no difficulties on obtaining a domain name for
my non-profit organization.

-- Kenji // just another UUNET subscriber
--
Kenji Rikitake // VMS/Japanese Development, DEC Japan R&D Center
Outside Japan: kenji@macrofield.org // From Japan: kenji@komaba.c.u-tokyo.ac.jp

-----------[000066][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 16:06:38 GMT
From:      atkinson@cmf.nrl.navy.mil (Randall Atkinson)
To:        news.announce.newgroups,news.groups,comp.protocols.tcp-ip,comp.dcom.lans
Subject:   RFD:  comp.dcom.lans reorganization


  Based on comments received thus far via email and news, I'm revising
the RFD for the FDDI newsgroup and expanding the area under discussion
slightly (at the request of others).  Each question will stand alone
on the ballot if this thing goes to a vote in another 2.5 weeks.  The
proposals are currently in the discussion period.  Under USENET rules,
discussion pro and con should be posted to news.groups.

  If there are existing mailing lists that cover the same charter,
they could be one-way or bi-directionally gatewayed if there were a
clear consensus in favor of that on BOTH USENET AND THE MAILING LIST.


FDDI Proposal
=============
Newgroup Name:  comp.dcom.lans.fddi
Moderated:	no

Charter:	Discussions of the FDDI, FDDI II, and FDDI Follow-on Lan
		(FFOL) protocols, proposals, implementations, and use.

Rationale:	Although FDDI is a MAN more than a LAN, this appears to
		be the consensus of the best name for the group.  Moderation
		doesn't appear needed.  FDDI and its related protocols
		are coming into wide use and existing discussions have
		been splintered across several groups.  This unites them
		in a single cohesive group whose size should be smallish
		but with sufficient daily traffic to justify the split.


Ethernet Proposal:
==================
Newgroup Name:	comp.dcom.lans.ethernet
Moderated:	no

Charter:	Discussions of Ethernet protocols, implementations,
		and use -- including both the traditional and the 
		IEEE 802.3 variants.

Rationale:	This group was proposed by various folks in posted news
		and in email sent to the originator of the FDDI proposal.
		It was asserted that there is sufficient volume in 
		comp.dcom.lans that is really ethernet specific to justify 
		creation of an Ethernet-specific newsgroup.  Also, it
		was asserted that the Ethernet postings tend to drown out
		the other postings to comp.dcom.lans


Renaming Proposal:
==================
Old Name:	comp.dcom.lans
New Name:	comp.dcom.lans.misc
Moderated:	no
Aliasing:	the old group would be aliased to the new name
		if the proposal passes to ensure that a smooth
		transition occurs and no traffic is lost.

New Charter:	Discussion of miscellaneous LAN and MAN protocols and
		implementations.

Rationale:	This group was proposed by various folks in posted
		news and in email reacting to the FDDI newgroup proposal.

		It has been the USENET experience that renamings of this
		variety have tended to significantly reduce inappropriate
		cross-posting among related groups.  This is because
		people have tended to crosspost most postings between 
		subsidiary groups (such as the two proposed above) and
		the top-level group but have tended not to cross-post
		between groups at the same level in the namespace.
-- 
----------------------------------------------------------------------
Randall Atkinson		Naval Research Laboratory
atkinson@cmf.nrl.navy.mil	Washington, DC 20375-5000
----------------------------------------------------------------------

-----------[000067][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 16:31:20 GMT
From:      roy@jupiter.sun.csd.unb.ca (Roy Nicholl)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc
Subject:   Telnet emulator for DOS (Colour and NL support)


I have a bit of a problem here: one of our customers is running PC-NFS on five 
DOS machines in a network to connect to UNIX file & database servers.  They are
also using the telnet (with vt100 emulator) that comes with PC-NFS to run an 
Ingres ABF application on the database server.  The "problem" is that 1) they 
have to be able to use both French and English when updating this database --
the emulator does not support (apparently) characters other than USA English;
2) They have (actually would like) to have colour [ie menu items in one colour,
input in another, query output in a third, etc].  I have contacted SUN with 
reference to their "enhanced telnet emulator" and was told that it does not
provide these capabilities either.  I know that NCSA telnet maps vt100 features
such as reverse text, flashing text, etc. to different colours, BUT it is 
another all-in-one package (the emulator is not separate from the lower level
network support) and hence will not run with PC-NFS.  Does anyone know of a
telnet/NFS package for DOS that provides both colour and National Language
support (I would think that a vt220/320 emulator should do the job).

All replies are welcomed.

In (Pseudo)reality, 

Roy


**************************************************************************
** Roy Nicholl					**  UNIX n. (1969)	**
** Atlantic Systems Group			**       Unbelievably	**
** Incutech Centre				**	 Nasty		**
** University of New Brunswick			**	 Information	**
** INTERNET: roy@jupiter.csd.unb.ca		**	 eXchange	**
**************************************************************************

-----------[000068][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 17:01:50 GMT
From:      vpmg9492@uxa.cso.uiuc.edu (Vance P. Morrison)
To:        comp.protocols.tcp-ip.ibmpc,comp.dcom.lans,comp.protocols.tcp-ip
Subject:   PCroute2.21 avalable

Hello,

PCroute2.21 is now available via anonymous FTP from accuvax.nwu.edu 
(129.105.49.1) in the pub/pcroute directory.  Note that this version
has no new features, but does include a bug fix or two.  If you have
experienced bizarre behavior, you may want to upgrade, otherwise it
is probably not worth it.

Vance

-----------[000069][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 17:05:35 GMT
From:      sss@nic.cerf.net (Marlene M. Eckert)
To:        comp.protocols.tcp-ip
Subject:   Remote control of equipment via terminal server


Hello all:

The environment:

1) SCO UNIX w/ ODT 1.1.0
2) Emulex 4000-T Terminal Server

The problem:

We will be remotely controlling various pieces of test equipment.  How
can we hang the equipment off the terminal server so that it is transparent
to our application whether the equipment is directly connected (/dev/ttyX)
or remotely connected (/dev/ttypX)?

We can remotely control the terminal server from the host to establish the
connection.  What's the best technique to get our application to talk
to the equipment via a pseudo tty?  Is a pseudo-device the best way to
go?

Thanks in advance!  Please email me directly, I'll summarize and post if
there is any interest.

-----------------------------------------------------------------------
Michael Reznick  (We use a shared account.)      Internet: sss@cerf.net
Structured Systems & Software (3S)
Laguna Hills, CA
-----------------------------------------------------------------------

-----------[000070][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 17:07:09 GMT
From:      vpmg9492@uxa.cso.uiuc.edu (Vance P. Morrison)
To:        comp.protocols.tcp-ip.ibmpc,comp.dcom.lans,comp.protocols.tcp-ip
Subject:   Bridge Routing with PCroute

Hello,

For any of you who have been using the verion of PCroute configured
to be a bridge router that is available from accuvax.nwu.edu (129.105.49.1)
in pub/pcroute/exp/broute, this notice is you.

Some major bugs have been found in that code which causes the bridge
to forward packets it should not, and to forward packets back on the
interface they came in on.  These problems have been fixed in the
code that is now on accuvax.   If you are using this code, I strongly
recommend getting a copy.

For those of you who have compiled bridge-routers from source, the 
PCroute version 2.21 source also incorperates these fixes.

Vance

-----------[000071][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 19:22:34 GMT
From:      dbeedle@rs6000.cmp.ilstu.edu (Dave Beedle)
To:        comp.protocols.tcp-ip
Subject:   What is SLIP?


     What is SLIP?  I've been asked and, while I've heard of it, I don't know
any details.  Thanks!

TTFN


-- 
  Dave Beedle                                    Office of Academic Computing
                                                    Illinois State University
  Internet:  dbeedle@rs6000.cmp.ilstu.edu                    136A Julian Hall  
    Bitnet:  dbeedle@ilstu.bitnet                          Normal, Il   61761

-----------[000072][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 21:58:27 GMT
From:      dixon@twisto.eng.hou.compaq.com (Dixon Dick)
To:        comp.protocols.tcp-ip
Subject:   Anon FTP location for Draft RFC's ?

Title sez it all. Anyone know ?


%  Dixon Chan Dick  (dixon@compaq.com)   	 Compaq Computer Corp.	 %
%									 %
% "All dumb opinions expressed here are mine and not Compaq's etc. etc." %

-----------[000073][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 22:18:09 GMT
From:      daveb@ingres.com (Dave Brower)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.domains
Subject:   Re: Paying for the Revolution (was Re: Alternate NICs)

In <1991Nov2.231837.12295@thunder.mcrcim.mcgill.edu>, peterd@cs.mcgill.ca (Peter Deutsch) writes:
|> (and I should know, being the guy most responsible for
|> trying to find support to keep archie alive. Donations and
|> contributions to date: $0.00, although lots of good
|> volunteer work by many people).

Just to get a sense of scale here, could we get some approximations--

    - how many "service requests" are made to archie per/interval
    - how much would it cost there be to support it as you'd like?
	
from which we could derive something like a cost-recovery model if we
wanted to somehow charge per transaction.  Then, think about what people
might be willing to pay for it.  One of my (and I suspect others)
problems with many existing charged information services is that the
things always seem to cost ten times more than I'd be willing to pay.  
For instance it might be worth $.05 or $.50 to make an archie request,
but if the charge would be $5.00 or $50.00, it would be hard to swallow.

Digression:

It's my impression the conventional wisdom for information owners is to
deny/withold access to data until they can figure out a way to make a
killing.   

One of the appeals of the Internet/Usenet community is that it has not
been subject to inflated information costs becuase it hasn't had any,
only transport cost and server support absorbed as justfied internally
by various organizations.

This discussion is best held elsewhere I believe, but I don't know
where.  Someone else please redirect followups to a likely spot.

thanks,
-dB

-----------[000074][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 22:49:47 GMT
From:      vjs@rhyolite.wpd.sgi.com (Vernon Schryver)
To:        comp.protocols.tcp-ip
Subject:   Re: arp cache & icmp frags

In article <1991Nov6.153630.22657@interlan.Interlan.COM>, dale@interlan.Interlan.COM (Dale B) writes:
> >
> >A variation of an old game--on a moderately fast workstation to a similar
> >one, connected with ethernet:
> >
> >	ping -f -s4400 otherhost
> >
> >If this doesn't collapse your ethernet, then something is wrong.
> 
> And what might that something wrong be? ...


Your machines are not "moderately fast".

You should see a meltdown, about 100% ethernet utilization as measured by
independent network monitors. 3rd parties try to transmit 16 times and
give up, and otherwise suffer heavy losses on their own pings.

I don't know precisely which workstations of other vendors are fast enough
for this command to melt the ether.  I assume there must be many, since all
currently shipping products of my employer should be at least that fast.
(Please excuse me for not doing exhaustive testing on the networks around
here.) I don't mean to brag.  I really think that current Sun and DEC
machines, to name only two obvious vendors, are probably fast enough,
although I don't have samples handy to test.  One might have to port the
4.3BSD ping command to get the -f flag.


Vernon Schryver,   vjs@sgi.com

P.S.  I mean "100%", not "80%" or the other low numbers claimed by
	anti-ethernet bigots.

-----------[000075][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 23:14:23 GMT
From:      eckert@immd4.informatik.uni-erlangen.de (Toerless Eckert)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.domains
Subject:   Re: Paying for the Revolution (was Re: Alternate NICs)

From article <1991Nov2.231837.12295@thunder.mcrcim.mcgill.edu>, by peterd@cs.mcgill.ca (Peter Deutsch):
> 
> Now comes the hard part - I was going to say implementing
> charging mechanism on the Internet, but that's not
> actually it. The hard part is to get the existing users to
> accept that they really pay for services so they'd be
> willing to pay for this one, too. Then we'd get rid of the
> prejudice against commercial traffic on the Internet and
> see the growth of services, and not just wires and routers.

And then we will enter the astonishing world of marketing,
product placement, commercials, merchandising and overpricing.
Commercialisation comes at a price and like in society there must be public
services to establish an open infratstucure. Openness is quite often
the opposite of capitalism, just take unix as an example and how
it mutated under the hands of all those companies telling you they are
open. They are not and they cannot really be open if they want to make
a profit. Likewise for networking. It must all be kept in balance to
keep up an Internet that's worth its name.

> In the long term I would like to U.S. government to start
> thinking of the backbone as something akin to the highway
> system, and not have them worry about whether the traffic
> is commercial or private. Then, maybe we wouldn't need all
> this navel gazing about whether certain traffic is legal
> or not. Sigh.

I think that's just the way it should happen.
---

             Toerless.Eckert@informatik.uni-erlangen.de
    /C=de/A=dbp/P=uni-erlangen/OU=informatik/S=Eckert/G=Toerless/
       SunSoft - to boldly go, where no bug has gone before

-----------[000076][next][prev][last][first]----------------------------------------------------
Date:      6 Nov 91 23:27:00 GMT
From:      barmar@think.com (Barry Margolin)
To:        comp.protocols.tcp-ip
Subject:   Re: DNS troubles

In article <1991Nov4.204810.9112@afterlife.ncsc.mil> cawilco@afterlife.ncsc.mil (Chris A. Wilcox) writes:
>First off, let me apologize if this is not the correct group
>for such a question. I looked for a group devoted to DNS-related
>problems, but didn't see anything, so this was my next best
>guess.

Well, there *is* a newsgroup devoted to the DNS:
comp.protocols.tcp-ip.domains (which I think is gatewayed to the
NAMEDROPPERS mailing list).  However, since your problem isn't DNS-related,
you lucked into the right group.

>(afterlife:8)% ping -s 26.29.0.172
>PING 26.29.0.172: 56 data bytes
>64 bytes from charon-gw.ncsc.mil (26.29.0.172): icmp_seq=0. time=39. ms
>64 bytes from charon-gw.ncsc.mil (26.29.0.172): icmp_seq=0. time=2040. ms
>64 bytes from charon-gw.ncsc.mil (26.29.0.172): icmp_seq=0. time=4040. ms
>64 bytes from charon-gw.ncsc.mil (26.29.0.172): icmp_seq=0. time=6060. ms
>64 bytes from charon-gw.ncsc.mil (26.29.0.172): icmp_seq=0. time=8060. ms
>(etc.)
>
>Why would two seconds be added to the ping every time? The nearest I
>can figure is that the DNS is having a tough time resolving the
>IP address back into the name charon-gw.ncsc.mil. I have noticed
>that there doesn't seem to be much of a structure for resolving
>reverse names in the DNS on the Internet (I can find SOA records
>for in-addr.arpa requests for some networks, not for others). So,
>my questions are:

The time is computed before translating the name, so your problem isn't
DNS-related.

I think the key to the problem is the "icmp_seq=0" on every line.  Ping
puts a sequence number in each transmitted packet, and matches it against
the sequence numbers in the reply to match them up and compute round trip
times.  I think charon-gw is failing to copy the data portion of the Echo
packet into the Echo Reply, so all the replies appear to be for packet #0.

>1) Has anybody seen this sort of behavior for ping before?

I've never seen it before.

>2) Does a reverse-addressing structure exist in DNS for the Internet?

Yes, the in-addr.arpa domains exist for precisely this purpose.  It's not
perfect (it only allows delegation at octet boundaries, which screws
organizations that need to have subnets that aren't multiples of 8 bits),
but it generally suffices.

>3) Anybody know how to fix the above problem?????

I think it's probably a bug in charon-gw's ICMP implementation.  Complain
to the router's manufacturer.
-- 
Barry Margolin, Thinking Machines Corp.

barmar@think.com
{uunet,harvard}!think!barmar

-----------[000077][next][prev][last][first]----------------------------------------------------
Date:      7 Nov 91 02:56:23 GMT
From:      lance@motcsd.csd.mot.com (lance.norskog)
To:        comp.protocols.tcp-ip
Subject:   Re: TCP/IP and Ethernet in real time control?

David Anderson's DASH group has done some very interesting
work in this area, particularly in real-time transfer of data
for audio and video from disk and across a local Ethernet.  

You can FTP their papers from: icsi.berkeley.edu 

-----------[000078][next][prev][last][first]----------------------------------------------------
Date:      7 Nov 91 06:39:07 GMT
From:      ngwl@nusdiscs.BITNET (Ng Wei Lin)
To:        comp.protocols.tcp-ip
Subject:   3+Open TCP and CUTCP


Hello,

	We are considering moving over to 3+Open TCP (which I heard 
is similar to LM/X) for our 3+Open network.  However, we hope to 
continue using the CUTCP (Packet driver) which we are so accustomed to.  
Is there any adapter available so that both can coexist?

	Thank you.


Wei-Lin Ng (email:ngwl@nusdiscs.bitnet)
National University of Singapore

ps: This message is posted to 2 lists, viz, tcp-ip and 3com-l.

-----------[000079][next][prev][last][first]----------------------------------------------------
Date:      7 Nov 91 06:44:55 GMT
From:      aimsuser@ncst.ernet.in (aimsuser citibank overseas)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.sys.hp,comp.sys.novell,comp.dcom.lans
Subject:   Novell and TCP/IP

We have a HP 832 business server. It is on a ethernet LAN
running TCP/IP. There are a few PC clients which run applications
that access ther server using SOCKETS.

The PC clients have Western Digital Ethernet Cards. 
Unfortunately our company does not have NFS, and the only
other things that the PC clients can use are telnet and ftp.
We cannot mount file systems.

We also have a Novell Server running Novell Netware 386 v 3.11.
This has TCP/IP support.

We were wondering if the clients currently accessing HP 832 could
logon to the Novell file server.

Some other questions that arise are

1. What additional software do we need at the workstation/servers.

2. Are the West. Dig. cards compatible with Novell Networks. If so
   are there different drivers to be used.

3. Is ftp available on Novell, so that the users on the HP 832
	could access it. Is ftp a adon ?

Can any one out there who has a similar setup can help us out ?

Thanks in advance.

P.S.
The PC clients acessing the HP currently have this software running

Resident ( loaded thru config.sys)
1. HP Resident TCP/IP Protocol Driver
2. West. Dig. Driver

Unloadable ( loaded thru AUTOEXEC)
1. HP DOS TCP/IP Protocol Driver
2. HP SOCKETS
3. HP Telnet
4. HP BAPI

-----------[000080][next][prev][last][first]----------------------------------------------------
Date:      7 Nov 91 11:20:57 GMT
From:      stephan@cs.kuleuven.ac.be (Stephan Biesbroeck)
To:        comp.protocols.ppp,comp.protocols.tcp-ip
Subject:   SLIP or PPP for Sequent/Dynix ??? Anyone


Hello,

I am looking for a version of SLIP or PPP that runs on a
Sequent with Dynix Release 3.0.17.9

Thanks for any info

Stephan Biesbroeck
stephan@cs.kuleuven.ac.be
stephan@Belgium.EU.net
stephan@ub4b.buug.be

-----------[000081][next][prev][last][first]----------------------------------------------------
Date:      7 Nov 91 13:41:15 GMT
From:      anthes@geocub.greco-prog.fr
To:        comp.protocols.tcp-ip
Subject:   FTP PORT command


I have a FTP server here on our local net, which rejects the following
PORT command:
PORT 1,2,8,2,1,157
while it's perfectly happy with  this one:
PORT 1,2,8,2,6,196

The ftp server software sends back a '530 Bad parameters for PORT command.',
in the first case and and a '200 PORT command okay.' in the second case.

Now I'd like to know, whether the server is full of it for not accepting
this port number, or if the client shouldn't be asking for a data connection
using this port?

-----------[000082][next][prev][last][first]----------------------------------------------------
Date:      7 Nov 91 13:45:04 GMT
From:      mlong@IASTATE.EDU (Michael C Long)
To:        comp.protocols.tcp-ip
Subject:   Re: NetBIOS on TCP/IP ?

In article <9111061518.AA22252@ftp.com>, jbvb@FTP.COM (James B. Van Bokke

     
>     2)      A NetBIOS implementation that runs over a serial line?
>  
> Our Netbios runs on any of our kernels, including SLIP and PPP.  However, the
> other end of the line has to be either the server itself, or something that
> knows to forward the broadcast packets out onto the LAN behind it.  Note
> that some Netbios-based software (e.g. LAN Manager) may need to have its
> timeouts re-configured for use over slow media.  I don't know of any "RAW
> Netbios over serial line" standards; there may be proprietary products.
> > 
I beleive IBM sells a product called Remote NETBIOS that alows one to run
NETBIOS over a serial line.

-----------[000083][next][prev][last][first]----------------------------------------------------
Date:      7 Nov 91 14:42:52 GMT
From:      dorn@fablap.uucp
To:        comp.protocols.tcp-ip
Subject:   CMU-TEK TCP/IP for VAX


Does anyone here know of a FTP or Mailserver site for downloading the
CMU-TEK TCP/IP implementation for Vax?

Dorn Hetzel

-----------[000084][next][prev][last][first]----------------------------------------------------
Date:      7 Nov 91 15:34:59 GMT
From:      oberman@ptavv.llnl.gov
To:        comp.protocols.tcp-ip,comp.protocols.nfs,comp.dcom.lans
Subject:   Re: DOS TCP/IP Mail package

In article <martinea.6@hawk.nstn.ns.ca>, martinea@hawk.nstn.ns.ca (Michael Martineau) writes:
> 
> Try POPmail from boombox.micro.umn.edu:/pub/POPmail.  It is a PC based
> POP2/POP3 client.  You will need a UNIX based POP server but there
> are lots of them floating around.

You will need a POP server. There is no reason for it to be Unix or anything
else. My POP server is definately NOT Unix. Like POP clients, POP servers can
(and probably do) exist for many platforms running IP protocols.

R. Kevin Oberman			Lawrence Livermore National Laboratory
Internet: oberman@icaen.llnl.gov	(510) 422-6955

Disclaimer: Don't take this too seriously. I just like to improve my typing
and probably don't really know anything useful about anything.

-----------[000085][next][prev][last][first]----------------------------------------------------
Date:      7 Nov 91 18:45:51 GMT
From:      bvk@scharat.uucp (Dr. B.van Kruechten)
To:        comp.unix.questions,comp.unix.aix,bit.listserv.aix,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,bit.listserv.ibmtcp-l
Subject:   Re: RS6000 + PCs in one Network

schommer@rhrk.uni-kl.de (Michael Schommer [Inf-FSE]) writes:



>  Hi Folks !!
 
>  I'm looking for somebody who has connected PCs to an IBM RS6000 via
>  Token Ring or Ethernet, using some kind of DOS TCP-software, and is
>  using the PCs as X-terminals (using some kind of X-software).
 
>  If you have successfully done this task - please feel free to share your
>  experiences with me. Simply send me some EMail.
 
>  Thank you in advance..
>                                                Saari
 
> Michael H. Schommer                                schommer@rhrk.uni-kl.de
> Am Glockenturm 10                                ixmo@informatik.uni-kl.de
> 6750 Kaiserslautern                                         IRCNICK: Saari
>                                                     Amateur Radio: DC 3 VR
> Vox: 0631 / 78514                   "Phantasie ist wichtiger als Wissen !"
> Btx: 063178455-0001                                      (Albert Einstein)


Pls. contect

   C.Schasiepen GmbH
   Harkortstrasse 21-23
   D 4030 Ratingen 1

  Tel: xxx/2102/49060



   bvk

-----------[000086][next][prev][last][first]----------------------------------------------------
Date:      7 Nov 91 18:46:24 GMT
From:      perrone@REMUS.RUTGERS.EDU (perrone)
To:        comp.protocols.tcp-ip
Subject:   University List



I am wondering if you could tell me from whom I may obtain a listing of
most of the Internet and Bitnet address names for the universities.
If you have this listing or can inform me where to obtain it, could you 
please reply to:     perrone@romulus.rutgers.edu

Thank you very much.

Paul J. Perrone

-----------[000087][next][prev][last][first]----------------------------------------------------
Date:      7 Nov 91 18:49:22 GMT
From:      pingn@hpindda.cup.hp.com (Teodora Ngo)
To:        comp.protocols.tcp-ip
Subject:   Re: re: InterOp Debate (response to trip report by Joachim Martillo)

/ hpindda:comp.protocols.tcp-ip / martillo@clearpoint.com (Martillo) /  1:28 pm  Oct 22, 1991 /

At InterOp, there was a debate over routing issues.  Radia Pearlman,
Milo Medin, Dave Clarke and others took part.

Pearlman was advocating the use of IS-IS routing in place of OSPF even
though IS-IS does not guarantee the "best" possible route because she
claimed it really did not make a difference if now and then a packet
took a less than optimal route.

If "best" is evaluated in terms of cost, such an assertion is really not
valid, and many people would care if a significant fraction of packets
were taking higher cost (i.e.  less than optimal routes).  Of course,
the PTTs would not mind such non-optimal routing as it increases their
profit. 

Dave Clarke and his opponent were debating the desirability of common
routing for IP and OSI packets.  Dave Clarkes observations on complexity
and true goals in general were eminently reasonable.  But in some cases
the use of different routing logic and routing suites for IP and OSI
packets seems absolutely necessary because OSI allows the construction
of protocol stacks where every protocol layer is connection oriented and
where higher protocol layers depend on the reliability of lower protocol
layers. 

OSI packets involved in such protocol stacks might require entirely
different routing procedures from IP packets, so that common routing of
IP packets and such OSI packets would seem inappropriate. 

Also I have to wonder whether OSI packets at corresponding protocol
layers always carry the same sorts of information useful to routing as
would be carried by comparable DOD Internet packets.  I could make a
good case that in some internets large IP packets which contain UDP
messages may optimally follow different routes than small IP packets
which contain TCP messages.  If in some cases routing different subsets
of IP packets differently makes sense, then forcing the common routing
of IP and OSI packets in general may be even quite unreasonable. 

In a sense, we can accept complexity and even try to make use of it
through the reasonable partitioning of the routing problem or we can
quixotically try to obliterate complexity and develop "uncomplex"
systems (like OSI) in which the routing problem has not be partitioned
and which will probably achieve rather uniformly mediocre yet uncomplex
performance. 

Joachim Carlo Santos Martillo Ajami


-- 
The statements contained in this article solely represent the views of
the author and in no way do they reflect the official opinion or policy
of Clearpoint Research Corporation.
----------

-----------[000088][next][prev][last][first]----------------------------------------------------
Date:      7 Nov 91 19:27:23 GMT
From:      cmaeda@EXXON-VALDEZ.FT.CS.CMU.EDU (Christopher Maeda)
To:        comp.protocols.tcp-ip
Subject:   multicast

I was wondering which commercial unixes (bsd or sysV) support ip
multicast.  I was also wondering which ip implementations for DOS
support it.  Please send email since I don't read this list.  I will
summarize.

Thanks,
Chris

-----------[000089][next][prev][last][first]----------------------------------------------------
Date:      7 Nov 91 19:38:56 GMT
From:      rcq@FTP.COM (Bob Quinn)
To:        comp.protocols.tcp-ip
Subject:   Re: References / Source for sockets / daemons

Hello Nirad!

>   I am going to be writing a network daemon soon and,  as I understand it,  will
>   be employing sockets to do the job.  Does anyone have any good articles / books
>   / source code examples that I could look at ?  I am aware of the bsd networking
>   code being available on uunet.uu.net & munnari.oz.au (closer to me) which I
>   will be looking into.

I support our PC/TCP Dev Kits for DOS and OS/2, both of which provide
Sockets API's so I'm asked this question quite often.  By far the
most comprehensive book for programming using Sockets (that I've found,
anyway) is:
             _Unix Network Programming_
             by W.Richard Stevens
             (c) 1990 Prentice-Hall, Inc.
             International Standard Book Number: 0-13-949876-1

As well as investigating inter-process communications paradigms and
ramifications, Stevens documents the Sockets Functions and provides a
nice variety of sample code (including examples of some common BSD
clients/servers).

 Bob Quinn                                        (617) 246-2920
 FTP Software, Inc.                                 rcq@ftp.com

-----------[000090][next][prev][last][first]----------------------------------------------------
Date:      8 Nov 91 02:36:21 GMT
From:      steve@mcnnet.UUCP (Steve Corso)
To:        comp.protocols.tcp-ip
Subject:   Looking for SLIP Driver for IBM Mainframe

Is anyone out there in netlan aware of a slip implementation of SLIP
to permit connectivity between IBM's TCP/IP Version 2 for MVS and
some other system (via an IBM 3745)?

I will summarize for the group if any positive responses are received.

Steve Corso
sharkey.cc.umich.edu!mcnnet!steve

-----------[000091][next][prev][last][first]----------------------------------------------------
Date:      8 Nov 91 03:09:41 GMT
From:      jm36+@andrew.cmu.edu (John Gardiner Myers)
To:        comp.protocols.tcp-ip
Subject:   A new twist on SMTP mis-implementation

Here's an unusual failure mode:

% telnet mvax.cc.ic.ac.uk smtp 
Trying 129.31.80.1...
Connected to mvax.cc.ic.ac.uk.
Escape character is '^]'.
421 mvax.cc.ic.ac.uk: Cannot resolve your address. '128.2.12.31'
Connection closed by foreign host.

This SMTP implementation gets my vote for most blatant violation of
the intent of a Host Requirement (namely 5.2.5) that "gets off" on a
technicality.  While it effectively refuses to receive mail due to
failure to back-verify an IP address, it does not wait for the HELO
command to decide to close the connection.

-----------[000092][next][prev][last][first]----------------------------------------------------
Date:      8 Nov 91 08:03:30 GMT
From:      aimsuser@ncst.ernet.in (aimsuser citibank overseas)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.dcom.lans,comp.sys.hp,comp.sys.novell
Subject:   Novell and TCP/IP

We have a HP 832 business server. It is on a ethernet LAN
running TCP/IP. There are a few PC clients which run applications
that access ther server using SOCKETS.

The PC clients have Western Digital Ethernet Cards. 
Unfortunately our company does not have NFS, and the only
other things that the PC clients can use are telnet and ftp.
We cannot mount file systems.

We also have a Novell Server running Novell Netware 386 v 3.11.
This has TCP/IP support.

We were wondering if the clients currently accessing HP 832 could
logon to the Novell file server.

Some other questions that arise are

1. What additional software do we need at the workstation/servers.

2. Are the West. Dig. cards compatible with Novell Networks. If so
   are there different drivers to be used.

3. Is ftp available on Novell, so that the users on the HP 832
	could access it. Is ftp a adon ?

Can any one out there who has a similar setup can help us out ?

Thanks in advance.

P.S.
The PC clients acessing the HP currently have this software running

Resident ( loaded thru config.sys)
1. HP Resident TCP/IP Protocol Driver
2. West. Dig. Driver

Unloadable ( loaded thru AUTOEXEC)
1. HP DOS TCP/IP Protocol Driver
2. HP SOCKETS
3. HP Telnet
4. HP BAPI



P.S. This is a repeat posting. Unfortunately I have many of the
replies to the previous posting due to a disk crash. I would be
obliged if the persons who had replied to the earlier posting
would mail the replies once again.
Sorry for the inconvinience caused.

-----------[000093][next][prev][last][first]----------------------------------------------------
Date:      8 Nov 91 14:14:29 GMT
From:      csmoko@relay.nswc.navy.mil (Chuck Smoko - E41)
To:        comp.protocols.tcp-ip,comp.sources.wanted
Subject:   LOOKING for FTP w/ automatic binary mode

When I was using a co-worker's SGI (Silicon Graphics) workstation to
ftp(1) to another Unix machine, I noticed that the ftp(1) program
automaticly goes into binary mode.  How nice!  I was wondering if
anyone knows where I can get the source to an ftp that does this.  All
of the ftp sources that I have 'played with' don't seem to have this
binary negotiation feature.  While I am talking about ftp(1), does
anyone know where I can get an ftpd(8) that logs to syslog who does a
login, who attempts a login, and what they do.  Most ftpd's log that
there was a connection from machine 'x' at time 't'.  After looking at
the ftpd(8) code, I can see how to do it.  But I would imagine that
someone has done this.  Ftpd's debug logging is way overboard; it even
logs passwords.

Please send e-mail suggestions and me-too's. I will do a follow-up if I
get enough requests for these programs.


                        csmoko@relay.nswc.navy.mil

   __                             __       ()
  /  ) /           /)            /  )      /\               /
 /    /_  __. __  // _  _       /--/      /  )  ____   ____/_  __
(__/ / /_(_(_/ (_(/_(<_/_)_    /  ( o    /__/__/) ) )_(_) / <_(_)

-----------[000094][next][prev][last][first]----------------------------------------------------
Date:      8 Nov 91 15:23:02 GMT
From:      kevin@msa3b.UUCP (Kevin P. Kleinfelter)
To:        comp.unix.aix,comp.protocols.tcp-ip
Subject:   Non-Passworded Account and FTP

I am trying to FTP to an AIX machine.  When I am prompted for a user-id,
I enter an ID for an account that has no password.  FTP on the AIX machine
responds with '331' message (account needs password). Correct behavior
or bug?

The account genuinely exists, and has no password, as verified by using
telnet to login on that account, without a password prompt.

(Actually, I'm writing an implementation of FTP on another machine, and
testing it by contacting an AIX machine.  When I send "USER username"
is when it responds with the 331.)
-- 
Kevin Kleinfelter @ DBS, Inc (404) 239-2347   ...gatech!nanovx!msa3b!kevin
Dun&Bradstreet Software, 3445 Peachtree Rd, NE, Atlanta GA 30326-1276
WARNING: I have been advised that email to kevin@msa3b.UUCP may bounce.
It looks like email will have to go via 'gatech' because that is well-known.

-----------[000095][next][prev][last][first]----------------------------------------------------
Date:      8 Nov 91 19:49:00 GMT
From:      gavron@alpha.sunquest.com (Ehud Gavron 602-885-7700x.2546)
To:        comp.protocols.tcp-ip
Subject:   Re: A new twist on SMTP mis-implementation

In article <8d6Tzpm00WBw8I3OwI@andrew.cmu.edu>, jm36+@andrew.cmu.edu 
(John Gardiner Myers) writes...
#Here's an unusual failure mode:
# 
#% telnet mvax.cc.ic.ac.uk smtp 
#Trying 129.31.80.1...
#Connected to mvax.cc.ic.ac.uk.
#Escape character is '^]'.
#421 mvax.cc.ic.ac.uk: Cannot resolve your address. '128.2.12.31'
#Connection closed by foreign host.
# 
#This SMTP implementation gets my vote for most blatant violation of
#the intent of a Host Requirement (namely 5.2.5) that "gets off" on a
#technicality.  While it effectively refuses to receive mail due to
#failure to back-verify an IP address, it does not wait for the HELO
#command to decide to close the connection.

	Same here, and my address *has* correct pointers.  I tried
	four times in a row, giving it enough time to percolate NS
	queries and cache them.  I'm not only disgusted at the
	implementation or at its failure to deal with a bona-fide
	host, but that I can't mail to its postmaster and whine :-)

	Ehud

--
Ehud Gavron        (EG76)     
gavron@vesta.sunquest.com
NEWARK has been REZONED!!  DES MOINES has been REZONED!!

-----------[000096][next][prev][last][first]----------------------------------------------------
Date:      8 Nov 91 19:57:51 GMT
From:      imp@solbourne.com (Warner Losh)
To:        comp.protocols.tcp-ip
Subject:   Re: FTP PORT command

In article <4148@geocub.UUCP> anthes@geocub.greco-prog.fr writes:
>I have a FTP server here on our local net, which rejects the following
>PORT command:
>PORT 1,2,8,2,1,157
>while it's perfectly happy with  this one:
>PORT 1,2,8,2,6,196
>The ftp server software sends back a '530 Bad parameters for PORT command.',
>in the first case and and a '200 PORT command okay.' in the second case.
>Now I'd like to know, whether the server is full of it for not accepting
>this port number, or if the client shouldn't be asking for a data connection
>using this port?

The first one is a request to connect to a port less than 1024.  Many
servers believe this is a security risk, so they don't honor these 
requests.  To obtain the widest range of interoperability, ftp clients
should request ports > 1024.

Warner
-- 
Warner Losh		imp@Solbourne.COM	  MMP to DoD #882
"Red hair is caused by sugar and lust," the woman, who was blond, confided.
"Highly evolved beings do not indulge in sugar and lust." -- Tom Robbins

-----------[000097][next][prev][last][first]----------------------------------------------------
Date:      8 Nov 91 20:16:12 GMT
From:      barrett@daisy.ee.und.ac.za (Alan P Barrett)
To:        comp.protocols.tcp-ip
Subject:   Re: A new twist on SMTP mis-implementation

In article <8d6Tzpm00WBw8I3OwI@andrew.cmu.edu>,
jm36+@andrew.cmu.edu (John Gardiner Myers) writes:
> 421 mvax.cc.ic.ac.uk: Cannot resolve your address. '128.2.12.31'
 
> While it effectively refuses to receive mail due to
> failure to back-verify an IP address, it does not wait for the HELO
> command to decide to close the connection.

Not only that, but the colon on the 421 line is illegal.

The first word of text on SMTP 220, 221 and 421 reply lines should be
the server's canonical domain name.  This requirement is implied by the
examples in RFC 821 and by the lists in section 4.2.1 and 4.2.2, and is
explicitly stated in section 4.3:

"        Note: all the greeting type replies have the official name of
"        the server host as the first word following the reply code.
"
"           For example,
"
"              220 <SP> USC-ISIF.ARPA <SP> Service ready <CRLF>

--apb
Alan Barrett, Dept. of Electronic Eng., Univ. of Natal, Durban, South Africa
RFC822: barrett@ee.und.ac.za             Bang: m2xenix!quagga!undeed!barrett

-----------[000098][next][prev][last][first]----------------------------------------------------
Date:      8 Nov 91 23:41:21 GMT
From:      root@gatech.edu (Operator)
To:        comp.protocols.tcp-ip
Subject:   Re: DNS troubles

In article <khgui4INNgtt@early-bird.think.com> barmar@think.com (Barry Margolin) writes:
>In article <1991Nov4.204810.9112@afterlife.ncsc.mil> cawilco@afterlife.ncsc.mil (Chris A. Wilcox) writes:
>>(afterlife:8)% ping -s 26.29.0.172
>>PING 26.29.0.172: 56 data bytes
>>64 bytes from charon-gw.ncsc.mil (26.29.0.172): icmp_seq=0. time=39. ms
>>64 bytes from charon-gw.ncsc.mil (26.29.0.172): icmp_seq=0. time=2040. ms
>>64 bytes from charon-gw.ncsc.mil (26.29.0.172): icmp_seq=0. time=4040. ms
>>
>>1) Has anybody seen this sort of behavior for ping before?
>

yes, we have seen this at georgia tech, with sun's version of ping.  
but, in our case, it was adding 1 second with every packet.

i can't remember the specifics, but it was nameserver related,
i also think that being on a subnetted class B network was
also a part of the problem, (fact that sun's ping didn't look
at netmasks).

sorry i can't be more precise.

carter bullard         carter@gatech.edu
network technologies
georgia institute of technology
atlanta, georgia 30332-730

-----------[000099][next][prev][last][first]----------------------------------------------------
Date:      9 Nov 91 04:33:42 GMT
From:      nelson@sun.soe.clarkson.edu (Russ Nelson)
To:        comp.protocols.tcp-ip
Subject:   Re: 3+Open TCP and CUTCP

In article <00951221.4A3B2AA0.15547@nusdiscs.bitnet> ngwl@nusdiscs.BITNET (Ng Wei Lin) writes:

   	We are considering moving over to 3+Open TCP (which I heard 
   is similar to LM/X) for our 3+Open network.  However, we hope to 
   continue using the CUTCP (Packet driver) which we are so accustomed to.  
   Is there any adapter available so that both can coexist?

Yes.  You can use dis_pkt, which is a packet driver that uses an NDIS driver
to do its work.  It's netlab.usu.edu:[.novell]dis_pkt9.zip.

--
--russ <nelson@clutx.clarkson.edu> I'm proud to be a humble Quaker.
Peace is not the absence of war.  Peace is the presence of a system for
resolving conflicts before war becomes necessary.  War never creates peace.

-----------[000100][next][prev][last][first]----------------------------------------------------
Date:      9 Nov 91 07:37:44 GMT
From:      kre@cs.mu.oz.au (Robert Elz)
To:        comp.protocols.tcp-ip,comp.sources.wanted
Subject:   Re: LOOKING for FTP w/ automatic binary mode

csmoko@relay.nswc.navy.mil (Chuck Smoko - E41) writes:

>I was wondering if anyone knows where I can get the source to an ftp
>that does this.

Try ftp.uu.net:bsd-sources/usr.bin/ftp/*
(or munnari.oz.au:pub/net/ftp.tar.Z which is probably not the same version)

>does anyone know where I can get an ftpd(8) that logs to syslog who does a
>login, who attempts a login, and what they do.

You can try munnari.oz.au:pub/net/ftpd.tar.Z for that one, though it
only logs "interesting" things they do (file transfers, cwd's, etc)
not noise (directory listings, status commands, ...)

There are others.

kre

-----------[000101][next][prev][last][first]----------------------------------------------------
Date:      9 Nov 91 09:07:23 GMT
From:      paul@actrix.gen.nz (Paul Gillingwater)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.sys.hp,comp.sys.novell,comp.dcom.lans
Subject:   Re: Novell and TCP/IP

In article <1991Nov7.064455.26381@ncst.ernet.in> aimsuser@ncst.ernet.in (aimsuser citibank overseas) writes:
> We have a HP 832 business server. It is on a ethernet LAN
> running TCP/IP. There are a few PC clients which run applications
> that access ther server using SOCKETS.
 
Is this HP Sockets?

> The PC clients have Western Digital Ethernet Cards. 
> Unfortunately our company does not have NFS, and the only
> other things that the PC clients can use are telnet and ftp.

I presumre you are using HP ARPA Service 2.1 for MS-DOS?

> We cannot mount file systems.

Four options exist for MS-DOS to mount files systems from HP-UX
that I know of:
(a) HP LAN Manager/X
(b) Portable Netware from Innovus
(c) PC-NFS on clients and NFS on HP-UX
(d) PC Interface(?)

> We also have a Novell Server running Novell Netware 386 v 3.11.
> This has TCP/IP support.
> 
> We were wondering if the clients currently accessing HP 832 could
> logon to the Novell file server.
 
Generally one doesn't ``login'' to a file server, except to access
files.  You can access data on the file server by using Novell
client software on your PC.  HP have a version of their ARPA
Services 2.1 for MS-DOS which will cooperate with Netware (even over
IPX).  You might get a trade-in if you already have ARPA 2.1.
-- 
Paul Gillingwater, paul@actrix.gen.nz

-----------[000102][next][prev][last][first]----------------------------------------------------
Date:      9 Nov 91 13:15:39 GMT
From:      mk@Materna.DE (Michael Kuschke)
To:        comp.protocols.tcp-ip,comp.sources.wanted
Subject:   Re: LOOKING for FTP w/ automatic binary mode


Hello,

if you've got a version of the ftp-client that comes with the newer
versions of ftpd you can define an init-macro in your .netrc which is
executed when the first connection is made.

My version of the init-macro is:

macdef init
bin
hash on
prompt off


	Ciao,
	  -Michael
	  
--
Michael Kuschke                    e-mail : mk@materna.uucp
Dr. Materna GmbH		   Voice  : +49 231 5599 200
Vosskuhle 37			   FAX    : +49 231 5599 100
D-4600 Dortmund 1,  West Germany

-----------[000103][next][prev][last][first]----------------------------------------------------
Date:      10 Nov 91 10:09:10 GMT
From:      nirad@newdelphi.ceu.uq.oz.au (Nirad Sharma)
To:        comp.protocols.tcp-ip,comp.unix.programmer
Subject:   Re: getservent name server ?

jik@athena.mit.edu (Jonathan I. Kamens) writes:

>In article <nirad.689066050@newdelphi>, nirad@newdelphi.ceu.uq.oz.au (Nirad Sharma) writes:
>|> To the best of my knowledge,  getservent consults the local /etc/services
>|> file and returns the info.  What if a process on another host wants to glean
>|> this info from a particular host ?  Is there such a function ?
 
>No.  However, if you happen to be using a version of UNIX whose vendor has
>included support for Hesiod, you can use hesiod functions to retrieve service
>information, assuming that you are running a Hesiod nameserver at your site
>and that you have a list of services loaded into it.

I take it,  then,  that Hesiod provides C library functions for the clients to
query the hesiod servers ?  Does anyone have instructions on installing the
Hesiod stuff on SunOS 4.1.1 ?  I'd recreated my libc to consult DNS but what
do I do now ?

Is Hesiod stable ?  From my understanding I replace my in.named with Hesiod's.

>Jonathan Kamens						jik@MIT.Edu
>MIT Information Systems/Athena		    Moderator, news.answers
--
Nirad Sharma  (nirad@ceu.uq.oz.au)			Phone : (+61 7) 365 7575
Continuing Education Unit				Fax :	(+61 7) 365 7099
The University of Queensland.  QLD  4072.  AUSTRALIA

-----------[000104][next][prev][last][first]----------------------------------------------------
Date:      10 Nov 91 10:21:58 GMT
From:      nirad@newdelphi.ceu.uq.oz.au (Nirad Sharma)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   How portable is RPC ?

What platforms are RPC offered on ?
--
Nirad Sharma  (nirad@ceu.uq.oz.au)			Phone : (+61 7) 365 7575
Continuing Education Unit				Fax :	(+61 7) 365 7099
The University of Queensland.  QLD  4072.  AUSTRALIA

-----------[000105][next][prev][last][first]----------------------------------------------------
Date:      10 Nov 91 14:59:17 GMT
From:      trepex@tmpmbx.UUCP (Ralf Moritz)
To:        comp.protocols.tcp-ip
Subject:   terminalserver with SLIP/PPP support

Hello,

I'm looking for a terminalserver to connect to my ethernet that will
support SLIP (and PPP). It should have >= 4 ports. All ports will be
running at 38400 bps with modems connected (USR V.32bis).
The ports _must_ support RTS/CTS handshake and modem control (DTR/CD).
There must be a simple login/authentification for the ports. Via this
authentification the "user/computer" dialing in must be able to get his
own IP number.

The final setup should provide dialup SLIP access to the internet.

Does anyone know where to get such a terminalserver ?
Is anyone _using_ such a setup right now (I know that the whole setup
could be done with a sun and some normal terminalservers but I would
prefer to have the whole setup without a "computer and operating system"
that could bring up some trouble).

Ralf
--
Ralf Moritz		Tel: +49-30-8555350	Fax: +49-30-8555395
trepex@netmbx.de	trepex@tmp.mbx.sub.org	trepex@tmpmbx

-----------[000106][next][prev][last][first]----------------------------------------------------
Date:      10 Nov 91 20:25:29 GMT
From:      jlu@doughbie.com (Jeff Urlwin (sysop))
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.sys.hp,comp.sys.novell,comp.dcom.lans
Subject:   Re: Novell and TCP/IP

In article <1991Nov7.064455.26381@ncst.ernet.in> aimsuser@ncst.ernet.in (aimsuser citibank overseas) writes:
>We were wondering if the clients currently accessing HP 832 could
>logon to the Novell file server.
You might want to check into Novell's LAN Workplace for DOS.  As of a few 
months ago, the WD8003 was not supported, but other WD's may have been.
You need to find out if Novell as an ODI driver for your card.

My info on this is not complete, but I hope this helps.

Jeff

--
UUCP uupsi!doughbie!jlu
	-- or --
     jlu@doughbie.uu.psi.com


-- 
UUCP uupsi!doughbie!jlu
	-- or --
     jlu@doughbie.uu.psi.com

-----------[000107][next][prev][last][first]----------------------------------------------------
Date:      11 Nov 91 05:56:10 GMT
From:      pwickman@carroll1.cc.edu (Paul J. Wickman)
To:        comp.protocols.tcp-ip
Subject:   comp.protocols.snmp


	Has anyone been able to get the SNMP libraries from MIT or CMU 
to compile and run sucessively on a SCO Unix 3.2 machine.  The reason
I ask is that I am developing an SNMP monitor and would like to
develop it at home on my SCO 3.2 machine.

	Please respond to me directly as I do not frequently read
news.

	Thanks a lot.




-- 
Paul J. Wickman - Systems/Network Administrator, Computer Science Dept.
Carroll College	 Waukesha, WI 53186
Internet: pwickman@carroll1.cc.edu	Phone: 414-524-7343

-----------[000108][next][prev][last][first]----------------------------------------------------
Date:      11 Nov 91 12:47:13 GMT
From:      jik@athena.mit.edu (Jonathan I. Kamens)
To:        comp.protocols.tcp-ip,comp.unix.programmer
Subject:   Re: getservent name server ?

In article <nirad.689767750@newdelphi>, nirad@newdelphi.ceu.uq.oz.au (Nirad Sharma) writes:
|> I take it,  then,  that Hesiod provides C library functions for the clients to
|> query the hesiod servers ?  Does anyone have instructions on installing the
|> Hesiod stuff on SunOS 4.1.1 ?  I'd recreated my libc to consult DNS but what
|> do I do now ?

The Athena version of Hesiod does include such library functions --
hes_getservbyname, etc.

You can get the Athena hesiod distribution via anonymous ftp from
athena-dist.mit.edu.  Probably other places (like uunet) as well.

I don't know what functionality that Hesiod stuff done by vendors, such as
DEC's in Ultrix, provides.

|> Is Hesiod stable ?  From my understanding I replace my in.named with Hesiod's.

Hesiod is an accepted Internet protocol augmenting the original the DNS
protocol.  The newest release of Bind from Berkeley supports Hesiod.  If
you've modified your C library to use DNS and have installed the newest Bind
named from Berkeley, then your named should already be capable of doing Hesiod.

-- 
Jonathan Kamens						jik@MIT.Edu
MIT Information Systems/Athena		    Moderator, news.answers
	      (Send news.answers-related correspondence
		  to news-answers-request@MIT.Edu.)

-----------[000109][next][prev][last][first]----------------------------------------------------
Date:      11 Nov 91 13:31:15 GMT
From:      jik@athena.mit.edu (Jonathan I. Kamens)
To:        news.newusers.questions,comp.protocols.tcp-ip
Subject:   Re: FINGER command.

In article <9107@s3.ireq.hydro.qc.ca>, srini@ireq-num.hydro.qc.ca (Srinivasan K..ca) writes:
|> The finger command if used as
|> finger john@xxx.xxx.xxx.xxx   seem to work with practically any site in
|> the world.

Presumably, by xxx.xxx.xxx.xxx, you mean an IP address, i.e. four numbers
between 0 and 255 separated by periods.

|> But when I try it with a valid e-mail address it fails. (except local ones)

There are several possible explanations for this.  The first is that the
"valid e-mail address" you are trying is not directly on the Internet, but
rather is fed its mail by a site on the Internet and therefore has an MX
record registered for itself.  If that's the case, then you can't finger it.

The second is that your site is running old software that tries to use a host
table file rather than Domain Name Server (DNS) lookup, and your host table is
empty.  You must update your software or ask your admin to get it updated so
that it uses DNS instead of a host table.  This is no small order, and may
require replacing many of the standard network utilities, and you may have to
do it by hand if your vendor has not released a newer version of that OS that
does DNS.

The third is that your clients are configured to use DNS properly, but your
named is not running or is not configured properly.  You have not told us
*how* it fails, so this possibility cannot be ruled out.

The fourth is that your network is not directly connected to the rest of the
network; this could be intentional or otherwise.  Again, we cannot rule out
this possibility since you have not described the precise failure mode.

Does "finger @pit-manager.mit.edu" work?  If so, then there is probably
nothing wrong with your configuration, and you've been trying to figure MX
hosts.  If not, and you get "Host unknown" immediately, you're probably using
a host table and need to switch to DNS.  If you get "Host unknown" after a
delay, then you may be using DNS without named running or with it running but
misconfigured, or you may be segmented from the network.  If you get
"Connection timed out" then you're probably segmented from the network.

You have not provided us with a lot of information that would have helped us
to answer your question.  I suggest you read:

Subject: Welcome to news.newusers.questions! (weekly posting)
Newsgroups: news.newusers.questions,news.answers

Available in the indicated USENET newsgroup(s), or via anonymous ftp from
pit-manager.mit.edu (18.72.1.58) in the file:

/pub/usenet/news.newusers.questions/W_t_n.n.q!_(w_p)

Also available from mail-server@pit-manager.mit.edu by sending a mail message
containing:

send usenet/news.newusers.questions/W_t_n.n.q!_(w_p)

Send a message containing "help" to get general information about the
mail server.

I have directed followups to comp.protocols.tcp-ip because your program, if
there is a problem, is with the tcp-ip Internet and because I don't know of
any newsgroup to put the discussion in (if someone in comp.protocols.tcp-ip
thinks I could have made a better choice, please E-mail and tell me what it
is).

-- 
Jonathan Kamens						jik@MIT.Edu
MIT Information Systems/Athena		    Moderator, news.answers
	      (Send news.answers-related correspondence
		  to news-answers-request@MIT.Edu.)

-----------[000110][next][prev][last][first]----------------------------------------------------
Date:      11 Nov 91 14:33:16 GMT
From:      larryp@sco.COM (Larry Philps)
To:        comp.protocols.tcp-ip
Subject:   Re: arp cache & icmp frags

In <pue0cl0@sgi.sgi.com> vjs@rhyolite.wpd.sgi.com (Vernon Schryver) writes:

> In article <1991Nov6.153630.22657@interlan.Interlan.COM>, dale@interlan.Interlan.COM (Dale B) writes:
> > >
> > >A variation of an old game--on a moderately fast workstation to a similar
> > >one, connected with ethernet:
> > >
> > >	ping -f -s4400 otherhost
> > >
> > >If this doesn't collapse your ethernet, then something is wrong.
> > 
> > And what might that something wrong be? ...
> 
> 
> Your machines are not "moderately fast".
> 
> You should see a meltdown, about 100% ethernet utilization as measured by
> independent network monitors. 3rd parties try to transmit 16 times and
> give up, and otherwise suffer heavy losses on their own pings.
> 
> I don't know precisely which workstations of other vendors are fast enough
> for this command to melt the ether.  I assume there must be many, since all
> currently shipping products of my employer should be at least that fast.
> (Please excuse me for not doing exhaustive testing on the networks around
> here.) I don't mean to brag.  I really think that current Sun and DEC
> machines, to name only two obvious vendors, are probably fast enough,
> although I don't have samples handy to test.  One might have to port the
> 4.3BSD ping command to get the -f flag.

Actually the 4.3BSD ping command does not have the -f flag (at least
not in my copy of the sources), it probably appeared in the tahoe or
reno sources.  Also, a quick glance at the ping source shows it is
coded in such a way that unknown flags are silently ignored.  (Try
"ping -q -z -b hostname" - works fine on my machine :-)

Anyway, just for the record, the SCO Unix TCP/IP 1.1.3f version of ping
does *not* support the -f option.  It just ignores the flag and
transmits at the normal slow rate.

---
Larry Philps,	 SCO Canada, Inc.
Postman:  130 Bloor St. West, 10th floor, Toronto, Ontario.  M5S 1N5
InterNet: larryp@sco.COM  or larryp%scocan@uunet.uu.net
UUCP:	  {uunet,utcsri,sco}!scocan!larryp
Phone:	  (416) 922-1937

-----------[000111][next][prev][last][first]----------------------------------------------------
Date:      11 Nov 91 16:53:39 GMT
From:      rzsan@sbusol.rz.uni-sb.de (Alfred Neisius)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc
Subject:   NCSA Telnet V2.3 "Retransmissions"

We are using NCSA-Telnet and we found it most satisfactory.
However, we found in the newest versions (2.3b12, 2.3.0 and
2.3.03) some strange things:
In the case of put from PC to a host the ftpbin receives
the TCP-Sequence-number and starts at this bytecount with
the transmissions again. If the host answeres too slowly
the whole transmission load may be up to 8 times greater.
In some cases this leads to a network overload.
Is this problem known ?
If yes, what could we do about it ?
If no, we'll gladly provide you with trace protocols and 
other printouts you might need.

Universitat des Saarlandes
Rechenzentrum
Alfred Neisius                 neisius@rz.uni-sb.de
Im Stadtwald                   tel:49-681-302-2566
W-6600 Saarbrucken             fax:49-681-302-4462  
------------------
Germany 

-----------[000112][next][prev][last][first]----------------------------------------------------
Date:      11 Nov 91 17:28:00 GMT
From:      art@opal.acc.com (Art Berggreen)
To:        comp.protocols.tcp-ip
Subject:   Re: SNAP question

In article <47n102QX0cxg00@amdahl.uts.amdahl.com> pramod@uts.amdahl.com (Pramod Sharma) writes:
>
>I have a few simple questions regarding IEEE 802 SNAP (Sub Network Access 
>Protocol).
>
>Does anyone use SNAP for sending IEEE 802.3 Ethernet packets ?

It is legal to send IP, but recommended against.
SNAP is an option for Novell Netware.
AppleTalk V2 uses SNAP.

Basically ANY protocol that doesn't have an ANSI/ISO LSAP assigned could
use SNAP.  Since most everyone uses SNAP OUI 0 (DIX Ethertypes), there is
a valid Ethernet encapsulation that is usually preferred on the Ethernet
LANs.  Anyone with an ANSI OUI can define their own private set of
protocol types.

>If yes, how do you indicate to the Data Link provider, whether or not to use 
>SNAP in an outgoing (to the network) packet ?

This is typically VERY implementation dependent.  Often the Data Link User
provides the MAC header (indicating SNAP) and the Data Link Provider
just gets the SNAP header as the first part of the User Data.

>Is SNAP required for FDDI ?

It is, to carry the protocols which don't have ANSI/ISO LSAPs.
(Novell, AppleTalk, XNS, DECNET-IV, ARP, etc.)

>Thanks.
>
>Pramod Sharma      pramod@uts.amdahl.com

Art

-----------[000113][next][prev][last][first]----------------------------------------------------
Date:      11 Nov 91 17:31:07 GMT
From:      milionis@shannon.ee.wits.ac.za (Taki Milionis)
To:        comp.sys.novell,comp.protocols.tcp-ip.ibmpc,comp.protocols.tcp-ip
Subject:   Campus Networking Strategy.

I am currently doing a literature survey for a paper that I am helping to
write. We hope to set some standards for a fledgling campus-wide lan and wish
to examine the solutions implemented elsewhere.

Here is an INTRO.
In most respects, having considered the current facilities of our University,
we can safely say that by far the most common platforms are PC's connected
to NOVELL servers, UNIX boxes, a supercomputer 
(running TCP-IP networking) and a main-frame (which used to be very central) but
is now connected to the network via an IBM3172 with TCP-IP gateway software.
So basically we want to mix NETWARE and TCP-IP protocols!!
Within these parameters, there are a number of competing specifications 
(NDIS, ODI, packet-drivers) and we wish to implement all of them (greedy hey!).
I expect that we will have a number of departments defecting to OS2  (if
v2.0 is any good (and it SOUNDS great)) and
LAN Server so the NDIS requirement is quite important.

What I would like from you.
Firstly, has a large survey been done to determine what systems people are
using accross the Internet? If so, could you let me know where I can get it?
I cannot access FAQ's (no full Internet conn yet).

Secondly, if you are a Network administrator, and you have a "How we did it?"
document, please post it to me.

Thirdly, if you have answers to the following questions, please mail me.

1) Approx no of UNIX machines	       --
   Other platforms (please spec)       --
   Approx no of Netware Servers	       --
   Approx no of OS2 Servers            --
   Use NDIS                            --
   Use ODI			       --
   Use Packet-drivers                  --
   
2) Campus backbone
	  fddi			-
          ethernet		-
	  Token/Ring		-
	  arcnet		-
	  other (specify)	-

3) Backbone protocols
	  mixed native protocols  -
	  only IP		  -
	  only IPX		  -
	  other (specify)         -
	  
4) Routers/bridges used
	  multi-protocls routers  -
	  only IP routers         -
	  only IPX brouters       -
	  PCRoute		  -
	  other freeware routers  -
	  only Mac-layer bridges  -
	  
5) TCP-IP packages used on PC side
	  LAN wp for DOS           -
	  3/COM TCP Package        -
	  OS2 TCP-IP               -
	  FTP Software		   -
	  CUTCP			   -
	  NCSA			   -
	  QVTNET		   -
	  other freeware (specify) -
	  locally developed TCP pack  -
	  

Thank you in advance for your time. If I do get a massive ;-) response, I
will try to complile a summary. If not, I will convert our news-server into
a toaster.

Cheers,
	  	  
	  
	  
		  	  		  	  
	
       	
    







 
-- 
                         *    H     * . 
Taki Milionis        . ( .   HHH  .   ___  milionis@odie.ee.wits.ac.za
Network Manager          ___  H ____  |.|  (..011) 716-5423
Wits Elec Eng        O __|.|_ H |. | _|.|  
Joburg               | |.|..| H | .| | .| "Sometimes my arms .."

-----------[000114][next][prev][last][first]----------------------------------------------------
Date:      11 Nov 91 17:40:04 GMT
From:      meht@unifybb.uucp (SONY)
To:        comp.protocols.tcp-ip
Subject:   Re: terminalserver with SLIP/PPP support

In article <3431@tmpmbx.UUCP> trepex@tmpmbx.UUCP (Ralf Moritz) writes:
>Hello,
>
>I'm looking for a terminalserver to connect to my ethernet that will
>support SLIP (and PPP). It should have >= 4 ports. All ports will be
>running at 38400 bps with modems connected (USR V.32bis).
>The ports _must_ support RTS/CTS handshake and modem control (DTR/CD).
>There must be a simple login/authentification for the ports. Via this
>authentification the "user/computer" dialing in must be able to get his
>own IP number.
>
>The final setup should provide dialup SLIP access to the internet.
>
>Does anyone know where to get such a terminalserver ?
>Is anyone _using_ such a setup right now (I know that the whole setup
>could be done with a sun and some normal terminalservers but I would
>prefer to have the whole setup without a "computer and operating system"
>that could bring up some trouble).
>
>Ralf
>--
>Ralf Moritz		Tel: +49-30-8555350	Fax: +49-30-8555395
>trepex@netmbx.de	trepex@tmp.mbx.sub.org	trepex@tmpmbx

While we are not using the SLIP capabilities we are using the ANNEX IIe
terminal server which will fully support SLIP operation.  In addition
it provides for a fairly high level of security of dial-up operation
and a second level of password protection above the UNIX password.

Tom
--
Tom Mehrens		Tel: +408-944-4155	Fax: +408-432-1874
San Jose, Ca. U.S.A.

-----------[000115][next][prev][last][first]----------------------------------------------------
Date:      11 Nov 91 18:23:39 GMT
From:      dwb@IMD.Sterling.COM (David Boyd)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: How portable is RPC ?


	Any vendor which supports NFS or NIS supports at least the
runtime requirements of Sun RPC.  Unfortunately not all vendors 
supply libraries for program development.  The good news is the
source for these libraries is freely available. (I can't remember
the FTP site off the top of my head).  I have successfully ported
these libaries to DECs with no problem.  They even interacted correctly
with the run-time portions supplied by DEC.  I seem to recall that
this code is dependent on BSD style sockets and select, but I do not
recall any other major portability problems.

-- 
David W. Boyd	             UUCP:     uunet!sparky!dwb
Sterling Software IMD        INTERNET: dwb@IMD.Sterling.COM
1404 Ft. Crook Rd. South     Phone:    (402) 291-8300 
Bellevue, NE. 68005-2969     FAX:      (402) 291-4362

-----------[000116][next][prev][last][first]----------------------------------------------------
Date:      11 Nov 91 19:04:08 GMT
From:      magnuson@skyler.mavd.honeywell.com
To:        comp.protocols.tcp-ip
Subject:   Need example source for RFC 868

I don't know if this is the appropriate newsgroup, but...

I'm looking for (client side) source code implementing RFC 868 (Time Protocol). 
Does anyone out there have non-copyrighted sources for this?

-- 
Randy R. Magnuson
magnuson@skyler.mavd.honeywell.com  (612-378-4557)

-----------[000117][next][prev][last][first]----------------------------------------------------
Date:      11 Nov 91 19:19:16 GMT
From:      johnf@persoft.com (John Freeborg)
To:        comp.protocols.tcp-ip
Subject:   TCP/IP on Token-Ring?


  How do most implementations "talk" TCP/IP on a 802.5 token-ring LAN?  I
always thought they used the 802.2 LLC SNAP header with the IP and ARP
types buried in the SNAP header using the 0xAA SAP for LLC.  However,
I recently saw some material which showed a listing of "assigned" LLC SAPs
and 0x06 was listed for IP.  The material was from IBM so does this mean
that IBM RS/6000 workstations on a token-ring use the 0x06 SAP for both
DSAP and SSAP with the control field being 0x03 for unnumbered info in IP
frames?  Or do they use the 0xAA SAP and SNAP method?  Or both?  If they are 
able to use the 0x06 SAP for IP which SAP do they use for ARP?

  Thanks!
   - John

-------------------------------------------------------------------------------
John Freeborg          Software Engineer         Persoft
johnf@persoft.com                                465 Science Dr.
608-273-6000                                     Madison, WI 53711

-----------[000118][next][prev][last][first]----------------------------------------------------
Date:      11 Nov 91 19:58:38 GMT
From:      srini@ireq-num.hydro.qc.ca (Srinivasan K..ca)
To:        comp.protocols.tcp-ip
Subject:   Re: FINGER command.


Thank you for the information.
I found out that the in.named process is not running in my unix system.
Next I found out there is no named.boot file. So I guess these things
can be configured only by a system Administrator, if and when he wants it.

That apart, In my IBM-PC, I configured the name server to one not far from
here. Wow! the finger command works from MSDOS (PC-TCP software) with
any site carrying, "edu", "ca" etc.

Fingering a site with "gov", results in "Disconnected by host" message.
So I guess they do not want it.

Fingering a "com" results in "Host un reachable" message.
But the host IS reachable, because I have the Internet number of one
such "com" host. It works fine as long as I use the internet ip number.

Probably I should use some other "name server" who knows "com"
addresses. There should be some way of finding one. Do you have any
suggestions?

Thanks.

Srini vasan K.    srini@ireq-num.hydro.qc.ca

-----------[000119][next][prev][last][first]----------------------------------------------------
Date:      11 Nov 91 20:16:34 GMT
From:      thurlow@convex.com (Robert Thurlow)
To:        comp.protocols.tcp-ip,comp.protocols.nfs
Subject:   Re: How portable is RPC ?

In <nirad.689768518@newdelphi> nirad@newdelphi.ceu.uq.oz.au (Nirad Sharma) writes:

>What platforms are RPC offered on ?

In Sun's "ONC/NFS - A Technology Guide to Distributed Computing", I
see 68 implementations of the RPC libraries, and that's _way_ too many
to list :-)  You should talk your local Sun rep into giving you a copy
of this guide, or get it from Dennis Freeman at Sun Microsystems, (415)
336-0955.  If you're after the freely-redistributable RPC sources,
version 4.0 can be had via anonymous ftp from archive-server@bcm.tmc.edu.

Rob T
--
Rob Thurlow, thurlow@convex.com
Recent poll results show that more Canadians believe that Elvis Presley
is alive than would vote for the current Prime Minister, Brian Mulroney.

-----------[000120][next][prev][last][first]----------------------------------------------------
Date:      11 Nov 91 21:08:13 GMT
From:      a2934@mindlink.bc.ca (T. William Mitchell)
To:        comp.protocols.tcp-ip
Subject:   tcp/ip

I have a Mac Plus computer with 4 megs and an IBM XT compatible.  I am new to
modeming, and wish to find out how I can do offline mail/usenet reading.  I do
not have MacTCP.  Can anyone tell me where I can obtain it, and what I need to
get started?  Is it possible for me to use ftp on the computers I have?
Please e-mail your response.
T. William Mitchell

-----------[000121][next][prev][last][first]----------------------------------------------------
Date:      11 Nov 91 22:05:17 GMT
From:      vatsan@risky.Convergent.COM (Srivatsan)
To:        comp.protocols.tcp-ip
Subject:   Documents on fddi.

I understand NIC distribution centre at SRI has only RFC's .
Is it possible to ftp or ask for documentation for 
Token ring/FDDI from IEEE  "database". Is it public domain?.

I  would  appreciate your reponse. 

Mail to vatsan@Convergent.COM.

-----------[000122][next][prev][last][first]----------------------------------------------------
Date:      12 Nov 91 01:10:30 GMT
From:      lamy@sobeco.com (Jean-Francois Lamy)
To:        comp.protocols.tcp-ip
Subject:   Re: terminalserver with SLIP/PPP support

In <291EC076.7687@unifybb.uucp> meht@unifybb.uucp (SONY) writes:

>While we are not using the SLIP capabilities we are using the ANNEX IIe
>terminal server which will fully support SLIP operation.  In addition
>it provides for a fairly high level of security of dial-up operation
>and a second level of password protection above the UNIX password.

They indeed do all that (including Van Jacobson Compressed Header SLIP).  Note
however that Annexes apparently require manual routing tables for them to
be aware of what's beyond the point-to-point link.  This becomes a concern
if you have several such that need to be aware of each other.

Jean-Francois Lamy         lamy@sobeco.com, lamy@sobeco.ca, uunet!sobeco!lamy
Research and Development, Sobeco Technical Services, Montreal, Canada H2Z 1Y7

-----------[000123][next][prev][last][first]----------------------------------------------------
Date:      12 Nov 91 12:57:29 GMT
From:      ercm20@castle.ed.ac.uk (Sam Wilson)
To:        comp.protocols.tcp-ip
Subject:   Re: A new twist on SMTP mis-implementation

I will forward the discussion of this feature on to an appropriate
person.  However the UK Academic Community has a very strange situation
at the moment whereby wwe are allowed to use SMTP on campus but not
nationally nor internationally, even though we are slowly becoming
connected to the Internet for protocols other than mail.  There has been
good (well...) mail connectivity for some time via an application relay
called NSFNET-RELAY.AC.UK - this converts beween SMTP/RFC822 and the
UK's own Grey Book Book mail system.  This has all sorts of interesting
implications for how we do MX records and such and there is considerable
pressure for this to be changed - please be patient. 

I suspect that what is happening here (by intention, rather than by how
well it's actually done) is that the machine has code to refuse SMTP
connections from the Internet but accept calls from on site. 

The real question that needs an answer is why anyone is trying to make
an SMTP connection to that particular machine from outside the UK
anyway.  In the UK's own naming scheme it is not advertised as a mail
service.  If its name is appearing in outbound mail that suggests it is
misconfigured and needs fixing.  I suggest you contact
postmaster@cc.ic.ac.uk (MXed via nsfnet-relay.ac.uk) for more details. 


Sam Wilson
Network Serices, Computing Service
University of Edinburgh, Scotland, UK

[ Member of Shoestring project to introduce IP services across the UK's
    JANET Network ]

-----------[000124][next][prev][last][first]----------------------------------------------------
Date:      12 Nov 91 14:28:14 GMT
From:      gdal@gdalsrv.sas.upenn.edu (GDAL Manager)
To:        comp.protocols.tcp-ip
Subject:   pop server for SCO ODT


  Has anyone ported a pop mail server to SCO ODT? Either pop2 or pop3
is ok. Pointers to C source appreciated. Thanks in advance,

George
-- 
gdal@gdalsrv.sas.upenn.edu		Social Science Computing
theall@rm105serve.sas.upenn.edu		University of Pennsylvania

-----------[000125][next][prev][last][first]----------------------------------------------------
Date:      12 Nov 91 14:47:18 GMT
From:      dyer@spdcc.com (Steve Dyer)
To:        comp.protocols.tcp-ip,comp.unix.programmer
Subject:   Re: getservent name server ?

In article <nirad.689767750@newdelphi> nirad@newdelphi.ceu.uq.oz.au (Nirad Sharma) writes:
>I take it,  then,  that Hesiod provides C library functions for the clients to
>query the hesiod servers ?

Yes.  There are versions of getservby_().

>Does anyone have instructions on installing the Hesiod stuff on SunOS 4.1.1 ?
>I'd recreated my libc to consult DNS but what do I do now ? Is Hesiod stable ?
>From my understanding I replace my in.named with Hesiod's.

Actually, if you don't want to run named locally, you can just point to a
Hesiod-ized named using /etc/resolv.conf.  The Hesiod named is just a version
of the BIND named which also understands the HS class.

-- 
Steve Dyer
dyer@ursa-major.spdcc.com aka {ima,harvard,rayssd,linus,m2c}!spdcc!dyer

-----------[000126][next][prev][last][first]----------------------------------------------------
Date:      12 Nov 91 14:53:18 GMT
From:      foxworth@bnlux1.bnl.gov (Bob Foxworth)
To:        comp.protocols.tcp-ip
Subject:   Help with using RG213 on ethernet

I posted an article on 6 Nov (1991Nov6.210418.7150@bnlux1.bnl.gov)
asking for help with use of ordinary RG213 as a section of backbone
on a ethernet LAN, and use of different tap devices. (This was article
18266 at our site). I received zero replies, so perhaps there is another
group that would be more appropriate to post in. ANy guidance from any
readers would be appreciated. Thanks!

-----------[000127][next][prev][last][first]----------------------------------------------------
Date:      12 Nov 91 16:28:07 GMT
From:      sfancher@spdcc.com (Steven Fancher)
To:        comp.unix.aix,comp.protocols.tcp-ip
Subject:   DLC managers in AIX (help needed)

Does anyone out there have any experience with the AIX Data Link Control
package? I want to use the services of 'dlcether' and 'dlctoken' but cannot
seem to get the DLC_ENABLE_SAP ioctl() to work.

Any help would be very welcome!! (email to sfancher@clam.com)

Thanks,

Steven Fancher   CLaM Associates, Cambridge, MA. (617) 621-2542

-----------[000128][next][prev][last][first]----------------------------------------------------
Date:      12 Nov 91 22:17:58 GMT
From:      ggm@brolga.cc.uq.oz.au (George Michaelson)
To:        comp.protocols.tcp-ip,comp.unix.programmer
Subject:   Re: getservent name server ?

dyer@spdcc.com (Steve Dyer) writes:
>Actually, if you don't want to run named locally, you can just point to a
>Hesiod-ized named using /etc/resolv.conf.  The Hesiod named is just a version
>of the BIND named which also understands the HS class.

I don't want to be a damp squib in all this, but as part of the site network
admin for nirads campus, I'd prefer to see a PLAN for doing hesiod/kerberos
accross all departments before large-scale adoption of this here.

If its plausible and sensible for isolated sub-nets to go ahead and try this
stuff out, so much the better and power to Nirads elbow.

How easy is it going to be to retro-fit this into an existing DNS? will the
HS type freak out the slave servers for his sub-net for instance? the primary
for the campus sub-net is bind-4.8.3 because of fears about the Ultrix 4.2
named, which alas DOES have HS in it, unlike the 4.8.3 replacement. (or am
I out of date again...) and I dont want everybody else's nameservice to go
bottom-up.

-George
--
                         George Michaelson
G.Michaelson@cc.uq.oz.au The Prentice Centre      | There's no  market for
                         University of Queensland | hippos in Philadelphia
Phone: +61 7 365 4079    QLD Australia 4072       |          -Bertold Brecht

-----------[000129][next][prev][last][first]----------------------------------------------------
Date:      12 Nov 91 22:57:01 GMT
From:      vjs@rhyolite.wpd.sgi.com (Vernon Schryver)
To:        comp.protocols.tcp-ip
Subject:   Re: arp cache & icmp frags

In article <1991Nov11.143316.720@sco.COM>, larryp@sco.COM (Larry Philps) writes:
> 
> Actually the 4.3BSD ping command does not have the -f flag (at least
> not in my copy of the sources), it probably appeared in the tahoe or
> reno sources.....

Yes, checking old 4.3BSD source shows that -f did not appear until reno.
I mis-remembered that BSD picked it up not long after 1988 when we got -f
from the original source of ping, BRL.  It has been about 2 years since
Cliff Frost & others put -f, -R, and other stuff into the BSD sources at
Berkeley.

In any case, `ping -f` is in 4.3BSD-reno on your nearest BSD archive.  So
is `ping -l`, an even more effective way to figuratively melt coax.  In the
4.3BSD-reno version, while -f (flood) requires UID=0, -l (preload) does
not.


Vernon Schryver,   vjs@sgi.com

-----------[000130][next][prev][last][first]----------------------------------------------------
Date:      13 Nov 91 03:27:58 GMT
From:      jik@athena.mit.edu (Jonathan I. Kamens)
To:        comp.protocols.tcp-ip,comp.unix.programmer
Subject:   Re: getservent name server ?

In article <ggm.689984278@brolga>, ggm@brolga.cc.uq.oz.au (George Michaelson) writes:
|> the primary
|> for the campus sub-net is bind-4.8.3 because of fears about the Ultrix 4.2
|> named, which alas DOES have HS in it, unlike the 4.8.3 replacement. (or am
|> I out of date again...)

You are out-of-date.  The 4.8.3 bind distribution supports hesiod.

-- 
Jonathan Kamens						jik@MIT.Edu
MIT Information Systems/Athena		    Moderator, news.answers
	      (Send news.answers-related correspondence
		  to news-answers-request@MIT.Edu.)

-----------[000131][next][prev][last][first]----------------------------------------------------
Date:      13 Nov 91 13:57:59 GMT
From:      jos@bull.nl (Jos Vos)
To:        comp.sources.wanted,comp.unix.sysv386,comp.protocols.tcp-ip
Subject:   SLIP wanted for Interactive 2.0.2

Who can supply me with information about available SLIP implementations
for Interactive Unix V.3.2 ?  Please reply by MAIL.

Thanks!

-- 
--    Jos Vos <jos@bull.nl>    (UUCP: ...!{uunet,mcsun,hp4nl}!nlbull!jos)
--    Bull Nederland NV, Product Support Unix, Amsterdam, The Netherlands

-----------[000132][next][prev][last][first]----------------------------------------------------
Date:      13 Nov 91 14:43:52 GMT
From:      swb@zoyd.ae.utexas.edu (Steve W Bova)
To:        comp.protocols.tcp-ip
Subject:   Sun OS 4.1.1 and Tahoe release


We recently installed Van Jacobson's C-SLIP on a Sun SPARC 1 under Sun
OS 4.1.1.  The performance is not so hot; there are a lot of problems
with timeouts.  In the documentation, there is a comment that one
would prefer to run the tahoe release to avoid these problems.
However, that comment is in an INSTALLATION file that specifically
states that it is *not* for Sun OS 4.x.  Furthermore, the header file
that accompanies the tahoe release does not mention Sun OS 4.x.  (This
is understandable, since the release date in the Makefile is pre-4.x)

My question is the following:  Is the tahoe release compatible with
Sun OS 4.x in general and 4.1.1 in particular?  
--
_______________________________________________________________________________

Steven W. Bova                             "Dammit Smithers! This isn't
Computational Fluid Dynamics Lab              rocket science, it's brain 
Bldg. WRW 111                                 surgery!  Now hand me that
The University of Texas                       ice cream scoop!"
Austin, TX, 78712                  
     (512) 471-4069                            -Mr. Burns    
                                      
swb@cfdlab.ae.utexas.edu              

-----------[000133][next][prev][last][first]----------------------------------------------------
Date:      13 Nov 91 16:18:32 GMT
From:      gary@sci34hub.sci.com (Gary Heston)
To:        comp.protocols.tcp-ip
Subject:   Re: Help with using RG213 on ethernet

In article <1991Nov12.145318.14164@bnlux1.bnl.gov> foxworth@bnlux1.bnl.gov (Bob Foxworth) writes:
>I posted an article on 6 Nov (1991Nov6.210418.7150@bnlux1.bnl.gov)
>asking for help with use of ordinary RG213 as a section of backbone
>on a ethernet LAN, and use of different tap devices. (This was article
>18266 at our site). I received zero replies, so perhaps there is another
>group that would be more appropriate to post in. ANy guidance from any
>readers would be appreciated. Thanks!

At the moment, there's not a newsgroup dedicated to LAN hardware issues,
to the best of my knowledge.

RG-213/U is equivalent to Belden 8267; the Belden thick Ethernet cable 
is specified as 9880. The 8267 is slightly lighter weight, has a stranded
center conductor instead of solid, a bare copper braid shield instead of
foil and tinned copper braid, a slower velocity of propagation (66%c
instead of 78%c), more capacitance per foot (30.8pf instead of 26pf), 
and higher attenuation (.83db/100' instead of .52db/100').

It is the same diameter and impedance (.405"/50 ohm) as the specified
Ethernet cable.

In summary: Yes, you can use it, but it's not as good. You'll have shorter
segements and higher signal losses, and probably somewhat lower reliability
from the RG-213/U than you would from the 9880. One good point about the
RG-213/U is that (according to Belden) it passes the VW-1 Vertical Wire
Flame Test. It's not marked as a plenum cable, though--you'd need the
Belden 89880 for that. If you're doing a 500m segement, you probably
won't notice the difference between the RG-213/U and the 9880.

This information is from the Belden Master Catalog (for RG-213/U) and
the Belden Comprehensive LAN Catalog.

I don't recall your questions about taps, and have little experience with
those anyway. Sorry.

-- 
Gary Heston   System Mismanager and technoflunky   uunet!sci34hub!gary or
My opinions, not theirs.    SCI Systems, Inc.       gary@sci34hub.sci.com
Become a pheresis donor. Loan your blood to the Red Cross for a couple
of hours. They, and cancer patients, will appreciate it.

-----------[000134][next][prev][last][first]----------------------------------------------------
Date:      13 Nov 91 17:38:55 GMT
From:      henry@zoo.toronto.edu (Henry Spencer)
To:        comp.protocols.tcp-ip
Subject:   Re: Help with using RG213 on ethernet

In article <1991Nov13.161832.23428@sci34hub.sci.com> gary@sci34hub.sci.com (Gary Heston) writes:
>At the moment, there's not a newsgroup dedicated to LAN hardware issues,
>to the best of my knowledge.

Try comp.dcom.lans.
-- 
SVR4:  the first system so open that    | Henry Spencer @ U of Toronto Zoology
everyone dumps their garbage there.     |  henry@zoo.toronto.edu  utzoo!henry

-----------[000135][next][prev][last][first]----------------------------------------------------
Date:      13 Nov 91 18:04:51 GMT
From:      jalsop@seachg.uucp (John Alsop)
To:        comp.protocols.snmp,comp.protocols.tcp-ip
Subject:   Recommendations wanted for network management

We have a client with a moderately complex network environment:

	- 6 TCP/IP hosts
	- ~10 terminal servers (8 port)
	- an Appletalk network connected to the main Ethernet using a
	  GatorBox
	- a Cabletron MMAC-3 

The network will probably double in size/complexity over the next 18
months.

The hosts and terminal servers are SNMP-aware.  Not sure about the GatorBox
and MMAC.

I'm looking for recommendations for a network management solution suitable
for this size network.  Basic requirements are to monitor traffic, identify
problems, and help problem resolution.

An ideal solution would probably run on a PC, but I'm interested in
anything that will do the job.  It should be useable by mere mortals.

If anyone can help with suggestions, I would be most grateful. Please reply
by e-mail - I'll summarize in a couple of weeks.  Responses from vendors are
welcome.

-- 
John Alsop

Sea Change Corporation
6695 Millcreek Drive, Unit 8
Mississauga, Ontario, Canada L5N 5R8
Tel: 416-542-9484 Fax: 416-542-9479
UUCP: ...!uunet!mnetor!aimed!seachg!jalsop

-----------[000136][next][prev][last][first]----------------------------------------------------
Date:      13 Nov 91 18:57:26 GMT
From:      brunner@practic.UUCP (Thomas Eric Brunner)
To:        comp.protocols.tcp-ip
Subject:   Re: arp cache & icmp frags

There appears to be a minor but persistent confusion about the flood flag
for ping. This version is available via anonymous ftp from brl.mil and will
source icmp echo request packets as fast as the host can service i/o for
any user level process.

This version and several other versions allow for the pre-load of packets
(you write the packet body bytes) as well as to set the packet size. Setting
the size to be greater than the link MTU (or MRU for correct ppp jargon) will
allow you to see fragmentation (rememeber the original question?) with any
of several ethernet monitors: tcpdump, LAN Watch, or a purpose-built network
analyzer, or a router which makes its ip fragmentation counters available,
either via telnet interaction or an snmp query from some manager, or simply
allows tracing on an interface.

If you are looking for upper-limits on your i386/i486 host's network
implementation, which is what I recollect was tangential to the thread,
both one or more icmp floods or ttcp network cats are useful, an alternative
is to set up a routing loop between two hosts-as-routers and inject packets
into the loop. Another approach is to use a purpose-built network analyzer
which can source packets addressed to your host and flood your host at
various rates and packet sizes to determine at what point packets are
dropped.


-- 
#include <std/disclaimer.h>
Eric Brunner 4bsd/RT Project
uucp: uunet!practic!brunner or brunner@practical.com
trying to understand multiprocessing is like having bees live inside your head.

-----------[000137][next][prev][last][first]----------------------------------------------------
Date:      13 Nov 91 20:38:40 GMT
From:      leonard@arizona.edu (Aaron Leonard)
To:        comp.dcom.lans,comp.protocols.tcp-ip
Subject:   Dialup SLIP access for remote LANs

I'm looking for a method to allow people with TCP/IP
LANs (subnets) at home to dial up to the Internet.

At present, we allow users with home end systems to use
dialup SLIP to connect up.  In this technique, the user
manually dials from his workstation to our modem bank. 
The modems connect up to a cisco terminal server.  The 
user enters the command "SLIP" on the terminal server, 
which causes the terminal server to ask him for a host 
name and a password.  The terminal server then queries an
authentication server out on the network.  If the password
validates, then the cisco dynamically assigns an IP address
to the port based upon the contents of a local host table. 
The terminal connection thereafter is placed into SLIP mode,
and the cisco handles IP routing  by proxy ARPing for the
workstation out on its Ethernet side.

This is the "right" way to do dialup SLIP, because it ensures
that the SLIP client always gets the same IP address when it
connects up to the net, regardless of which physical terminal
server port it happens to grab.  Thus "r" commands, SMTP, etc.
all work as they should.

Now we need a way to allow users with LANs at home a way to
accomplish the same thing.  That is, we want to support a 
topology like this:

[es]   [es]     [es]   [es]     [es]   [es]    <= end systems
 |      |        |      |        |      |
===========     ===========     ===========    <= home LANs
     |               |               |
    [is]            [is]            [is]       <= SLIP-capable "gateway"
     |               |               |            machines at home
  [modem]         [modem]         [modem]
                                
                 
  [ modem pool ] ===== [ magic box ] ==== campus backbone


Each of the remote IS's should be able to dial up any modem
in the pool and select the SLIP service.  Magic Box should 
authenticate the remote user and then begin routing
IP traffic for the remote subnet thru that serial link.

Magic Box should NOT send routing information thru the 
serial link for obvious reasons, nor should it depend on 
hearing routing information from the remote LAN.  Rather, 
it should somehow "know" the identity of the remote subnet
based upon the authenticated host ID.  

Magic Box should RIP-advertise a route to the remote subnet 
onto the campus backbone.  When Magic Box detects hangup on 
the serial port, it should cease issuing RIP for the remote 
subnet, and should delete the route from its internal routing 
tables.

It is NOT a requirement of the proposed system that either
side automagically dial the other upon receipt of traffic;
we assume that the person on the far side will manually
cause his IS to dial up our modem pool.  It IS a requirement
that no modems in the pool, nor ports on Magic Box, be 
dedicated to this application, but that they be available
for ordinary terminal connections when unused by SLIP.

I'm well aware of the NetBlazer product, and it looks quite
nice, but it's just too expensive for this application.

Is anyone supporting dialup LAN access of the sort I
describe?  Does anyone have any (wacky) notions about how to
accomplish it?  If so, I'd appreciate hearing of your ideas;
send me mail, and I'll summarize if appropriate.

Aaron

Aaron Leonard (AL104)

University of Arizona Network Operations, Tucson AZ 85721
Leonard@Arizona.EDU     LEONARD@ARIZONA.BITNET     uunet!arizona!leonard

-----------[000138][next][prev][last][first]----------------------------------------------------
Date:      13 Nov 91 23:23:33 GMT
From:      osmana@well.sf.ca.us (Osman Arslaner)
To:        comp.protocols.tcp-ip,comp.dcom.lans
Subject:   AT&T 3B2's and Multiple Ethernet Cards.


Can we add a second Ethernet card to our 3B2 system (600 series) ?.
Presently, we have one ethernet card and are using Wollongong's TCP/IP
WIN/3B and have about 25 users accessing through the network.

 Is it possible to add a second card and give it an  IP address on the
same network without doing subnet routing as described in the TCP/IP
WIN/3B Installation and Admin Guide ?

Thanks.

Osman Arslaner
CN Rail,
Montreal, Quebec.

-----------[000139][next][prev][last][first]----------------------------------------------------
Date:      13 Nov 91 23:59:24 GMT
From:      albers@ka3ovk (Jon Albers)
To:        comp.protocols.ppp,comp.protocols.tcp-ip
Subject:   Point to Point and Terminal Server questions.


Hello.  I have 2 questions for the group:

1.	I have seen a number of postings about people having slip or ppp links
set up over standard phone lines, using V.32 modems or Telebits, from say,
their office to their home.  I was wondering how to set this up without
having something like a Netblazer at each end.  Our office has a number of
UNIX minis (Unisys, Sequent, etc), and a number of UNIX micros running
SCO.  They are linked to each other via thick-ethernet TCP/IP.  We also have
several Netblazers installed, linking a couple of nets across town (or 
across the street).  I have a micro at home running SCO as well.  How could I
dial up and connect my home machine to the network via modem?

2.	We also have a number of people in the office with DOS machines.  I
would like to get them up on the TCP/IP net.  Right now, I have done it easily
by hardwiring them to one of the minis, but we are moving from our current
rooms in the building, and I may not be able to run RS-232 connections around
for everyone.  Assuming I can run coax/twisted pair/whatever from the TCP/IP
net into the office, what software/hardware would be good to use for say,
8-10 PCs?  I would like them to be able to have telnet/ftp and maybe even
SMTP to their PCs if possible, but it wouldn't be required, so maybe a 
terminal server would be all I'd need?  What alternatives are out there?

						Thanks,
							Jon

-- 
| Jon Albers, IRS, Information Systems Management, Support and Installation.  |
| Office Symbols: ISM:S:S:SI   voice: (202/FTS)535-3729                       |
| UUCP:(media|teemc|tcsc3b2|ki4pv)!ka3ovk!albers       Packet: KA3OVK@WA3ZNW  |

-----------[000140][next][prev][last][first]----------------------------------------------------
Date:      14 Nov 91 00:34:03 GMT
From:      tab@tfs.com (Theresa Breckon)
To:        comp.protocols.tcp-ip
Subject:   info needed regarding data line to UK



We are investigating installing a private data line
from CA to the UK. We have gotten quotes from AT&T for
their international ACCUNET 56kbs full duplex private
line service carried on fiber optic cable. On the
UK end we have a choice of either British Telecom or
Mercury. Installation time about 3 months. This is
a concern.

I have also thought about the possiblility of using
a service like Uknet. Advantages would be faster installation,
lower cost. But, we are concerned about having no guarantee of
available end-to-end bandwidth.

We won't be doing development, but we will be wanting to log on
remotely, check status of running processes, transfer large amounts of
data, executables every couple of weeks.

Some questions I have for those with experience with
AT&T, UKnet, British Telecom, or Mercury:

1.  Has anyone had any experience with UKnet other than
   reading mail and news.

2. Has anyone had experience with British Telecom or Mercury
   in the UK?

-----------[000141][next][prev][last][first]----------------------------------------------------
Date:      14 Nov 91 01:22:04 GMT
From:      sss@nic.cerf.net (Marlene M. Eckert)
To:        comp.protocols.tcp-ip
Subject:   Use of TLI t_sync function

Hello Netters:

I hope this is the right group, I have a question regarding the TLI t_sync
function.

According to the SCO UNIX and SunOS t_sync man pages one should be able to
use an open system call followed by a t_sync call to open an endpoint. For
instance:

    sock = open("/dev/inet/tcp", O_RDWR);
    if (sock < 0) {
      perror("opening new endpoint");
      exit(1);
    }

    rval = t_sync(sock);
    if (rval < 0) {
      t_error("syncing endpoint");
      exit(1);
    }

In his book, "UNIX Network Programming", Richard Stevens also leads me
to believe that this should work.  Well, the open is successful, however
the t_sync fails (on both SCO UNIX and SunOS) with "Illegal file descriptor".

Any clues?  Is this a known bug in the original AT&T code?  Please
email me directly and I'll post if there is any interest.

Thanks in advance!

-----------------------------------------------------------------------
Michael Reznick  (We use a shared account.)      Internet: sss@cerf.net
Structured Systems & Software (3S)
Laguna Hills, CA
-----------------------------------------------------------------------

-----------[000142][next][prev][last][first]----------------------------------------------------
Date:      14 Nov 91 05:36:02 GMT
From:      ferriby@perot.com
To:        comp.protocols.tcp-ip
Subject:   News read for Macintosh

Is anyone aware of any newsreaders for the macintosh?  I'm
familiar with Intercon's version that is bundled with their
TCPconnect product and I am looking for others. (Public Domain,
Shared, Proprietary products, etc...)

--                             
John Ferriby

Perot Systems Corporation    Telephone: +1-313-641-3660
4555 Corporate Drive         Internet: ferriby@perot.com
Troy, MI 48098-6353          UUCP: uunet!perot!ferriby

-----------[000143][next][prev][last][first]----------------------------------------------------
Date:      14 Nov 91 06:51:13 GMT
From:      martell@ucs.ubc.ca (Jonn Martell)
To:        comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.sys.hp,comp.sys.novell,comp.dcom.lans
Subject:   Re: Novell and TCP/IP

In article <1991Nov10.202529.2625@doughbie.com> jlu@doughbie.com (Jeff Urlwin (sysop)) writes:
>You might want to check into Novell's LAN Workplace for DOS.  As of a few 
>months ago, the WD8003 was not supported, but other WD's may have been.
>You need to find out if Novell as an ODI driver for your card.

WD8003 ODI drivers come with NW Lite, haven't tried it yet. I have NW Lite
but only briefly looked at it.  LANtastic is technically better but I suspect
NW Lite (IPX) has better integration with TCP/IP.   What PD is available for
NetBIOS (protocol for LANtastic)
?

-----------[000144][next][prev][last][first]----------------------------------------------------
Date:      14 Nov 91 11:56:07 GMT
From:      rstevens@noao.edu (W. Richard Stevens)
To:        comp.protocols.tcp-ip
Subject:   Re: Use of TLI t_sync function

Between the open() and t_sync() you must ioctl(fd, I_PUSH, "timod")
to push the "timod" streams module onto the stream.  If the
descriptor is coming from a fork() or dup() you don't need to
do this, but with open() it's required.

	Rich Stevens  (rstevens@noao.edu)

-----------[000145][next][prev][last][first]----------------------------------------------------
Date:      14 Nov 91 18:55:41 GMT
From:      rgoldstone@OAVAX.CSUCHICO.EDU (Robin Goldstone)
To:        comp.protocols.tcp-ip
Subject:   getting to .mil

I have read about the problems with nic.ddn.mil.  Does this apply to MILNET
in general? A user of mine has been trying to send an e-mail reply to a 
usenet posting he read.  The address he is trying to reach is:
<someone>@uservx.afwl.af.mil

If I do an nslookup on this address, I get:
*** hq.csuchico.edu can't find "uservx.afwl.af.mil": Non-existent domain

Is this truly a "non-existent" domain, or is this really just a problem 
getting to MILNET?  BTW, today I was able to successfully ftp to nic.ddn.mil
for the first time in a long time.

Thanks for any help you can give me.
-Robin
***********************************************************************
Robin Goldstone, Systems Software Specialist
California State University, Chico  Computing Services
rgoldstone@oavax.csuchico.edu

-----------[000146][next][prev][last][first]----------------------------------------------------
Date:      14 Nov 91 19:47:02 GMT
From:      morgan@ms.uky.edu (Wes Morgan)
To:        comp.sys.att,comp.protocols.tcp-ip
Subject:   RFS woes (3B2/1000's with Wollongong TCP/IP)


I've configured our RFS to run over tcp.  Connections are made
with no problems, and access is fine.  However, the connections
seem to just drop dead with no notice.  tcp/ip connectivity is
still there, but RFS just decides to drop the connection.  This
is an intermittent problem; the connections may last for hours
or minutes.  There doesn't seem to be any rhyme or reason to
this one.  I've gone through the configuration for RFS, with
no luck.

Does anyone happen to have an answer?

Wes

-- 
 morgan@ms.uky.edu    |Wes Morgan, not speaking for|     ....!ukma!ukecc!morgan
 morgan@engr.uky.edu  |the University of Kentucky's|   morgan%engr.uky.edu@UKCC
 morgan@ie.pa.uky.edu |Engineering Computing Center| morgan@wuarchive.wustl.edu

-----------[000147][next][prev][last][first]----------------------------------------------------
Date:      14 Nov 91 20:58:39 GMT
From:      richb@kronos.com (Rich Braun)
To:        comp.protocols.ppp,comp.protocols.tcp-ip
Subject:   Re: Point to Point and Terminal Server questions.

albers@ka3ovk.UUCP (Jon Albers) writes:
>1.	I have seen a number of postings about people having slip or ppp links
>set up over standard phone lines, using V.32 modems or Telebits, from say,
>their office to their home.  I was wondering how to set this up without
>having something like a Netblazer at each end.

A Netblazer is nothing but an expensive modem with an Ethernet transceiver
and some SLIP/PPP software (plus a lot of other fancy features).  You can
accomplish the same functionality using either a Unix host with regular
modems hanging off it, connected using RS-232 ports, or using any of a
number of terminal servers which offer SLIP/PPP over their RS-232 ports
(which can be connected to modems).

I'm not arguing for or against the Netblazer, which is a fine product; I'm
just getting you to ask yourself exactly what you'd need to add to your
existing setup in order to support the dialup IP service you want.

>2.	We also have a number of people in the office with DOS machines.  I
>would like to get them up on the TCP/IP net.  Right now, I have done it easily
>by hardwiring them to one of the minis...
>what software/hardware would be good to use for say,
>8-10 PCs?  I would like them to be able to have telnet/ftp and maybe even
>SMTP to their PCs if possible, but it wouldn't be required, so maybe a 
>terminal server would be all I'd need?

The terminal server approach would work fine for your local systems but it
would be more expensive and offer a much lower level of service than a simple
Ethernet setup using public-domain software (if you don't already have a
terminal server and aren't planning to use one to solve problem 1).

What we use here at my company are standard Ethernet LAN cards with
thin-wire connectors.  They cost less than US$200 each (probably
closer to $100 if you shop around), and you just run coax between the
systems using BNC connectors.  You then download the Clarkson
University TCP package for DOS from omnigate.clarkson.edu, or any of a
few other freebie DOS TCP packages, and you're in business.  If you
want a commercially-backed TCP software package, you can buy one from
any of a number of vendors, but you don't necessarily have to do this
to "get your feet wet" with a DOS-based Ethernet.

If you haven't the money for Ethernet, you could use the SLIP/PPP
drivers included with the Clarkson packages.

-rich

-----------[000148][next][prev][last][first]----------------------------------------------------
Date:      15 Nov 91 01:04:51 GMT
From:      James P H Coleman
To:        comp.protocols.tcp-ip
Subject:   Re: News read for Macintosh

In article <1991Nov14.003602.6680@perot.com> ferriby@perot.com writes:
>Is anyone aware of any newsreaders for the macintosh?  I'm
>familiar with Intercon's version that is bundled with their
>TCPconnect product and I am looking for others. (Public Domain,
>Shared, Proprietary products, etc...)
>
>--                             
>John Ferriby
>
>Perot Systems Corporation    Telephone: +1-313-641-3660
>4555 Corporate Drive         Internet: ferriby@perot.com
>Troy, MI 48098-6353          UUCP: uunet!perot!ferriby
>
I am using TheNews V2.1 by Bill Cramer.  It is very good.
I do not know where I got it from but it is available on various
archive sites around the world.

James P H Coleman
Northern Territory University
Australia

-----------[000149][next][prev][last][first]----------------------------------------------------
Date:      15 Nov 91 08:22:41 GMT
From:      taybh@hpsgm2.sgp.hp.com (Tay Beng Hang)
To:        comp.protocols.tcp-ip
Subject:   Is the number of max retry in TCP configurable

Hi guys,
	I am running HP-UX 7.0 on HP9000 (all models) and would like to 
configure the MAX RETRY and TIMEOUT TIMER parameters for TCP. I found this file 
/etc/conf/netinet/timer.h and I believed it contains the definition for the
above parameters as follows: 
	MAX RETRY = TCP_MAXRXTSHIFT
	TIMEOUT TIMER = TCP_REXMTMAX

However, I don't have the necessary TCP source code. Can I still configure 
the above parameters? If not, is there any other ways that I can 
configure these 2 parameters?
	Thanks.

p/s: I understand that TCP retransmit timer is adaptive and is not
configurable. However, I believe that there is an upper bound for this value and
TIMEOUT TIMER mentioned above refers to this upper bound value.

 ____________________________________________________________________________
| Beng-Hang Tay                       | Telnet:    520 8732                  |
| Singapore Networks Operation        | Phone:     (65) 279 8732             |
| Hewlett-Packard Singapore Pte. Ltd. | Fax:       (65) 272 2780             |
| 1150 Depot Road                     | Internet:  taybh@hpsgm2.sgp.hp.com   |
| Singapore 0410                      |            taybh@hpsgnlc.sgp.hp.com  |
| Republic of Singapore		      | HPDesk:    HP3200/67                 |
 ----------------------------------------------------------------------------
    

-----------[000150][next][prev][last][first]----------------------------------------------------
Date:      15 Nov 91 13:00:26 GMT
From:      JFL4@psuvm.psu.edu
To:        comp.protocols.ppp,comp.protocols.tcp-ip
Subject:   Re: Point to Point and Terminal Server questions.

Not to change the subject, but....

I have need to get IP from building-to-building at a small campus.  I
started to work on the idea of using PCBridge and line drivers over
the existing four-wire that exists between these buildings;  however,
administrative types are *real* uncomfortable with this approach
(its not high-tech enough and it doesn't cost enough money) :)

I've seen reference to the NetBlazer (sp?) but I don't really need
the modem technology (i.e., straight serial connection w/ line drivers).
What's the best commercial approach for connecting two LANS, either
Ethernet -> Ethernet or Ethernet -> AppleTalk (encapsulated IP).
I'd rather use PPP than SLIP if it is available.  Is there anything
like a box that has an ethernet interface, processor, PPP software
and a serial connection?

All responses are appreciated, even if it's just to say my first
approach was the best (or the most cost-effective).

Thanks in advance for any help.

Jeff Luck
CES Systems Office
Penn State University
JFL4@PSUVM.PSU.EDU

-----------[000151][next][prev][last][first]----------------------------------------------------
Date:      15 Nov 91 14:30:11 GMT
From:      jrm@crest.uucp (John Morris)
To:        comp.sys.att,comp.protocols.tcp-ip
Subject:   Re: RFS woes (3B2/1000's with Wollongong TCP/IP)

morgan@ms.uky.edu (Wes Morgan) writes:


>I've configured our RFS to run over tcp.  Connections are made
>with no problems, and access is fine.  However, the connections
>seem to just drop dead with no notice.  tcp/ip connectivity is
>still there, but RFS just decides to drop the connection.  This
>is an intermittent problem; the connections may last for hours
>or minutes.  There doesn't seem to be any rhyme or reason to
>this one.  I've gone through the configuration for RFS, with
>no luck.
 
>Does anyone happen to have an answer?
 
>Wes
 
>-- 
> morgan@ms.uky.edu    |Wes Morgan, not speaking for|     ....!ukma!ukecc!morgan
> morgan@engr.uky.edu  |the University of Kentucky's|   morgan%engr.uky.edu@UKCC
> morgan@ie.pa.uky.edu |Engineering Computing Center| morgan@wuarchive.wustl.edu

	Wes,
	
	What version of TCP/IP and UNIX are you on.  I've heard that there

    were many problems with TCP/IP version < 3.2 .

				John Morris
				jrm@crest

-----------[000152][next][prev][last][first]----------------------------------------------------
Date:      15 Nov 91 15:27:48 GMT
From:      hari@metrix.UUCP (T hariharan)
To:        comp.protocols.misc,comp.protocols.tcp-ip,comp.protocols.tcp-ip.ibmpc,comp.mail.misc
Subject:   NOVELL<->UNIX EMAIL INFO (WANTED)

We have a LAN (Ethernet - 1 segment) with both Novell (PC clones
running MS DOS) and IP (sun3, sparcstation, sun386i) traffic.

We are looking for an email application for the PC's.
The requirements are:

1. Allow you to read your mail from the suns (that's sitting in your
   mailbox, say /usr/spool/mail/USERNAME or /home/USERNAME/mail/inbox)
   AND from the other Novell users.

2. Be able to send mail from your PC to other Novell users AND to the
   people on the sun's.

We'd like to know your experience with email packages (public domain and
commercial) that satisfy the above requirements. Information from
commercial vendors is also welcome.

Please email me a copy of your reply because I don't subscribe to these
newsgroups. I'll summarize the interesting responses.

Thanks in advance,
Hari
Metrix
1 Tara Blvd
Nashua NH 03062
(603)888-7790
uunet!metrix!hari (or) metrix!hari@uunet.uu.net

-----------[000153][next][prev][last][first]----------------------------------------------------
Date:      15 Nov 91 16:16:21 GMT
From:      jra@lawday.DaytonOH.NCR.COM (John R. Ackermann x2966)
To:        comp.protocols.tcp-ip
Subject:   Datagram Assembly Timer?

We've found that our TCP/IP implementation has a very short timer for
packet assembly; an application that does screen output slowly will
generate <lots> of tiny packets in a telnet session, when it could do a
few larger ones.  On an ethernet, this isn't really a problem, but we're
running on a half-duplex, slow channel and the clutter caused by all
these two or three character packets causes some problems.

At the moment, I'm looking for some general information, rather than
implementation-specific: what's the algorithm for determing when to stop
waiting and ship the current data out?  Is this part of the telnet
daemon, the tcp code, or something else?  Would SVR3 streams parameters
have anything to do with it?

And, finally, is there usually a tunable parameter that would allow one
to force the system to wait longer (to collect more data) before sending
out small packets?

Thanks...

-- 
John R. Ackermann, Jr.        Law Department, NCR Corporation, Dayton, Ohio
(513) 445-2966		      John.Ackermann@daytonoh.ncr.com
Packet Radio: ag9v@n8acv      tcp/ip: ag9v@ag9v.ampr [44.70.12.34]

-----------[000154][next][prev][last][first]----------------------------------------------------
Date:      15 Nov 91 18:17:29 GMT
From:      U47526@uicvm.uic.edu (David E. Thomas)
To:        comp.sys.att,comp.protocols.tcp-ip
Subject:   Re: RFS woes (3B2/1000's with Wollongong TCP/IP)

In article <1991Nov14.194702.28108@ms.uky.edu>, morgan@ms.uky.edu (Wes Morgan)
says:
>
>I've configured our RFS to run over tcp.  Connections are made
>with no problems, and access is fine.  However, the connections
>seem to just drop dead with no notice.  tcp/ip connectivity is
>still there, but RFS just decides to drop the connection.  This
>is an intermittent problem; the connections may last for hours
>or minutes.  There doesn't seem to be any rhyme or reason to
>this one.  I've gone through the configuration for RFS, with
>no luck.
>
We had a similar problem on our 3b2 400's a while back.  We were
able to characterize the problem a bit further.  What was happening
was: the RFS link was dropping,  RFS realized the link was dropped
and set it back up again but the original link was not deleted from
the table, thus eventually the table was full and the link could
not be remade.  We did not have a problem with applications 'bombing'
until the table was full, so we figure the link must have been reset
pretty quickly.  (Although the RFS mounted file systems were only
used sporadically and then as starlan virtual drives through dos).

We had AT&T support at the time so we called the hotline.  Three
different tiers of at&t techs logged on to the machine and investigated
the problem.  All of them thought it was a very interesting problem
but none could solve it.  The easiest solution for us was to migrate
the RFS functions to 386's.  An interim 'quick but dirty fix' was a
cron function that kicked off at 2:00 A.M. which brought RFS down,
and back up with an empty table.  The table never filled up within
one day, so this worked (after a fashion - we still had to deal with
the links going down periodically, but at least they came right back
up on their own).

We had 5 3B2's running RFS, all of them configured the same, but only
had the problem on one of the machines.  We were never able to
identify why it happened on that machine and not the others.


David E. Thomas
University of Illinois at Chicago

-----------[000155][next][prev][last][first]----------------------------------------------------
Date:      15 Nov 91 20:27:00 GMT
From:      gcreager@life.jsc.nasa.gov (CREAGER, GERRY)
To:        comp.protocols.tcp-ip
Subject:   TGV Multinet traceroute question

Using Multinet 2.2 on a Vax 3000, I've been getting a |permission denied" 
nasty-gram whenever I try to use the traceroute facility.  Both the system 
manager and I are stymied, and have been unable to decipher the meaning of 
the message, as to whether it refers to a particular socket, if the 
permission-denied may be a normal termination of the traceroute, etc.  The 
MultiNet manuals are not particularly helpful on this facet.

Comments, or help?  Please E-Mail and I'll summarize.

Thanks, Gerry
gcreager@life.jsc.nasa.gov

-----------[000156][next][prev][last][first]----------------------------------------------------
Date:      15 Nov 91 21:17:33 GMT
From:      bvj@ESD.3Com.COM (B.V. Jagadeesh)
To:        comp.protocols.tcp-ip
Subject:   SVNC-92, CAll for Papers Date Extended to Dec 4th


                    Silicon Valley Networking Conference - 1992
		    -------------------------------------------
			   Call For Papers

Papers are solicited for the Silicon Valley Networking conference (SVNC-92)
to be held April 27th to 29th 1992 at Santa Clara convention center,
Santa Clara CA 95052, USA. 

Papers are solicited in the following areas.

Distributed Systems
Internetworking
Network Management
X-windows
Advanced File servers
High Speed Networking
Standards activities (IEEE, CCITT, IETF etc ) 
Network Monitors
WireLess Networking

SVNC typically attracts over 400 engineers every year and is a nice forum
to discuss system design architecture and solutions to complex networking 
problems.

If you are interested in presenting a paper, please send me an abstract
of the paper before DEC-4th 1991. If accepted for submission, a rough draft 
of the paper should be submitted before JAN 10th 1991 and camera ready copy
should be submitted before Feb 1 1992. 
Please include your Address, Tel number and Fax Number in the abstract and
mail it to 

B.V. Jagadeesh
Silicon Valley Networking Conference
1248, Olive Branch Lane
San Jose, CA - 95120
USA.

Fax Number: (408)- 997-8265

Thanks
/Jagadeesh
bvj@3Com.com
(408)-764-5169

-----------[000157][next][prev][last][first]----------------------------------------------------
Date:      15 Nov 91 22:51:19 GMT
From:      FORDE@POLYVM.BITNET
To:        comp.protocols.tcp-ip
Subject:   lpr discription

I am look for the discription of the lpr protocol.  Any help would
be greatly apreciated.

-----------[000158][next][prev][last][first]----------------------------------------------------
Date:      15 Nov 91 23:22:56 GMT
From:      brnstnd@kramden.acf.nyu.edu (Dan Bernstein)
To:        comp.protocols.tcp-ip,alt.security
Subject:   New network security mailing list: rfc931-users

RFC 931, the Authentication Server, stops the most common type of mail
and news forgery, increases the security of other TCP connections, and
helps security organizations trace Internet attackers. It is supported
by at least three (completely interoperable) UNIX server implementations,
used by several clients including the newly released wuarchive ftpd, and
installed at sites throughout the United States and Europe, ranging from
the Air Force to companies as large as 3M. It is currently the only
freely available wide-area TCP security protocol, yet it can run in
tandem with other security protocols such as Kerberos.

I've created a new mailing list, rfc931-users, for people who want to
use the Authentication Server to solve problems. The mailing list is
rfc931-users@kramden.acf.nyu.edu; to join, send mail to brnstnd@nyu.edu.
Announcements like ``new RFC 931 servers in Kalamazoo'' are welcome, as
are technical discussions about network security, new client support,
and current server implementations. You're also welcome to join if you
don't use the Authentication Server but want to know what it's all
about.

RFC 931 is not yet an IETF-approved standard. If we want vendors to
raise the minimum standard for Internet security, we have to get
together and show them that we really care! As soon as the flow of
subscriptions ebbs a bit, I'll send my proposed RFC 931 revision to
the list to see what people think. This is your chance to see and
affect Internet politics in action---don't miss it...

---Dan

-----------[000159][next][prev][last][first]----------------------------------------------------
Date:      17 Nov 91 03:02:29 GMT
From:      skl@CS.SFU.CA (Samuel Lam)
To:        comp.protocols.tcp-ip
Subject:   Re: lpr discription

In article <91319.175119FORDE@POLYVM.BITNET>, you wrote:
>I am look for the discription of the lpr protocol.

It's in RFC 1179.

...Sam
-- 
Samuel Lam <skl@cs.sfu.ca>
Network Support Group, Centre for Systems Science, Simon Fraser University

-----------[000160][next][prev][last][first]----------------------------------------------------
Date:      17 Nov 91 21:46:21 GMT
From:      erick@sunee.waterloo.edu (Erick Engelke)
To:        comp.protocols.tcp-ip
Subject:   Re: lpr discription

Samuel Lam <skl@cs.sfu.ca> writes:
>In article <91319.175119FORDE@POLYVM.BITNET>, you wrote:
>>I am look for the discription of the lpr protocol.
>
>It's in RFC 1179.
>

LPR is not based on RFC 1179, rather the reverse is true.  The definative
source for LPR is the unix source from which it was derived.

RFC 1179 was an optimistic view of what LPR could be like.  Do not assume
that it is the standard, BSD and BSD derivatives (for example) do not accept
the zero length field as meaning a undetermined length.  A lot of micro
computers running NFS would operate differently if it were true.

Your best bet is to read 1179, it's well done and clean, sit down with the
unix source and work out the protocol between the two of them.

Erick
-- 
----------------------------------------------------------------------------
Erick Engelke					   Watstar Computer Networks
Network Developer				      University of Waterloo
Erick@Development.Watstar.UWaterloo.ca		    (519) 885-1211 Ext. 2965

-----------[000161][next][prev][last][first]----------------------------------------------------
Date:      18 Nov 91 15:05:51 GMT
From:      D.Nash@utexas.edu (Donald L. Nash)
To:        comp.protocols.tcp-ip
Subject:   Re: TGV Multinet traceroute question

In article <1991Nov15.183041.22415@aio.jsc.nasa.gov>, 
gcreager@life.jsc.nasa.gov (CREAGER, GERRY) writes:
>Using Multinet 2.2 on a Vax 3000, I've been getting a |permission denied" 
>nasty-gram whenever I try to use the traceroute facility.

This should have been sent to service@tgv.com, but we'll overlook that
for now.  The problem is that the MULTINET_TRACROUTE.EXE image needs to
be installed with SYSPRV privilege.  START_MULTINET.COM does not install
this image with privilege, so you can only run it from an account with
SYSPRV enabled.  You can safely install this image with SYSPRV, although
you'll need to do it at system startup if you want it to last across
reboots.  Try this:

	$ install create multinet:multinet_traceroute.exe /open /priv=(sysprv)

				++Don Nash

				The University of Texas System
				Office of Telecommunication Services

Internet:  D.Nash@utexas.edu
THEnet:    THENIC::DON
BITNET:    DON@THENIC

-----------[000162][next][prev][last][first]----------------------------------------------------
Date:      18 Nov 91 15:32:23 GMT
From:      poorman@convex.com (Peter W. Poorman)
To:        comp.protocols.tcp-ip
Subject:   Re: lpr discription

Samuel Lam <skl@cs.sfu.ca> writes:
>In article <91319.175119FORDE@POLYVM.BITNET>, you wrote:
>I am look for the discription of the lpr protocol.

Another source of information (in addition to the RFC and the source code)
is the book "UNIX Network Programming" by W. Richard Stevens.  This book
contains a description of the protocol, and includes the source to a simple
client.

--Pete Poorman
  poorman@convex.com

-----------[000163][next][prev][last][first]----------------------------------------------------
Date:      18 Nov 91 19:35:09 GMT
From:      reece@eco.twg.com (Reece R. Pollack)
To:        comp.protocols.tcp-ip
Subject:   Re: TGV Multinet traceroute question

In article <62131@ut-emx.uucp>, D.Nash@utexas.edu (Donald L. Nash) writes:
|>In article <1991Nov15.183041.22415@aio.jsc.nasa.gov>, 
|>gcreager@life.jsc.nasa.gov (CREAGER, GERRY) writes:
|>>Using Multinet 2.2 on a Vax 3000, I've been getting a |permission denied" 
|>>nasty-gram whenever I try to use the traceroute facility.
|>
|>This should have been sent to service@tgv.com, but we'll overlook that
|>for now.  The problem is that the MULTINET_TRACROUTE.EXE image needs to
|>be installed with SYSPRV privilege.  START_MULTINET.COM does not install
|>this image with privilege, so you can only run it from an account with
|>SYSPRV enabled.  You can safely install this image with SYSPRV, although
|>you'll need to do it at system startup if you want it to last across
|>reboots.  Try this:
|>
|>	$ install create multinet:multinet_traceroute.exe /open /priv=(sysprv)
|>
|>				++Don Nash
|>
|>				The University of Texas System
|>				Office of Telecommunication Services

It's not necessary to install traceroute with SYSPRV if it is being used
only by users who have access to the SYSPRV privilege. You may wish to
leave it uninstalled (or installed without SYSPRV) to prevent abuse by
non-privileged users. Traceroute, like Ping, can generate a fair amount
of network traffic, and most users have no need for it.

--
Reece R. Pollack
Senior Software Engineer
The Wollongong Group, Inc.

-----------[000164][next][prev][last][first]----------------------------------------------------
Date:      18 Nov 91 20:31:40 GMT
From:      rees@paris.citi.umich.edu (Jim Rees)
To:        comp.protocols.ppp,comp.protocols.tcp-ip
Subject:   Re: Point to Point and Terminal Server questions.

In article <91319.080027JFL4@psuvm.psu.edu>, JFL4@psuvm.psu.edu writes:

  I have need to get IP from building-to-building at a small campus...
  I've seen reference to the NetBlazer (sp?) but I don't really need
  the modem technology (i.e., straight serial connection w/ line drivers)...
  Is there anything
  like a box that has an ethernet interface, processor, PPP software
  and a serial connection?

That's exactly what the Netblazer is.  It has no modems.  It has an ethernet
connection and a bunch of serial rs-232 ports.  It can talk slip or ppp and
does vj tcp header compression.  It also does some things a standard Unix
host normally can't do.  For example, it does net routes right (with a
netmask) and doesn't require a subnet number for each slip line.  I've got
one here and I'm pretty happy with it.

A pair of Netblazers with line drivers will do what you want quite well, but
it may not be the cheapest solution.

-----------[000165][next][prev][last][first]----------------------------------------------------
Date:      18 Nov 91 22:11:54 GMT
From:      sob@hsdndev.UUCP (Scott Bradner)
To:        comp.protocols.tcp-ip
Subject:   the slimeballs are at it again

Well, here it is a few weeks after a new set of testing results were
presented at Interop (and put on hsdndev.harvard.edu in pub/rtests/10.91)
and someone out there seems to want to trash the messenger rather than
defend their product (or whatever).  Again, the word is being spread
that Cisco is somehow paying me.  Just so it is clear, the only
thing I've gotten from Cisco is the same thing I got from most of the
vendors that came in to get their boxes tested - lunch (and a cheap one at
that) (if you can't get lunch out of your vendor salesdroid they you
are not playing the game)

I do not yet have specific information on who is telling this tale but if
anyone out there was told this please let me know so I can get to the
source directly. Two vendors have given me the same source for the
"informaton" but I do not yet have any direct user experience and would
like to get any that there is.

I suppose it could all be a misunderstanding (sure it could!).   The lab
that I've setup at Harvard will, in a few weeks, be rentable by anyone,
vendor or user, to do the same sort of tests.  Cisco (and a number of
other vendors) have told me that they hope to make use of the facility as
part of their development cycles.  They can run the tests themselves
and keep the results private or have someone from Harvard run them and
then have some stamp of validity on the results. (please don't take this
as some sort of promo for the test lab, I'm just trying to put a "clean"
understanding on what they are saying - even if I don't for a moment think
that there is any chance that this is what is going on. - when promo time
comes, I'll put one in the right newsgroup -newprod)

To be clear;
	NO VENDOR HAS PAID SCOTT BRADNER OR HARVARD ANYTHING FOR TESTING
ROUTERS OR BRIDGES OR TO ADVISE ON THE DESIGN OF THE SAME OR FOR ANY OTHER
REASON RELATED TO THIS.
	I've been paid by Interop, IBM and Chipcom for giving tutorials on
building and managing data networks.  (That is FULL disclosure.)

	I'm insulted by this and hope to get enough information from you to show
just who is doing this.  I trust that if you do hear this from a vendor
you will take this sort of dirty pool into account when choosing a product
for your own networks. (and also let me know)


	Thanks for reading this flame

			Scott

-----------[000166][next][prev][last][first]----------------------------------------------------
Date:      18 Nov 91 22:55:38 GMT
From:      ce1zzes@prism.gatech.EDU (Eric Sheppard)
To:        comp.protocols.tcp-ip
Subject:   IP maps?

I wish to determine the name of a machine from its ip address.  Where
can I find host tables that I can search?  Please note that I am NOT talking
about nslookup, which pops out a number given an internet name.  I need to
do the opposite.

Eric
-- 
Eric Sheppard      Georgia Tech    |   "Of course the US Constitution isn't
Atlanta, GA                        | perfect, but it's a lot better than what
ARPA: ce1zzes@prism.gatech.edu     |             we have now." -Unknown
uucp: ...!{allegra,amd,hplabs,seismo,ut-ngp}!gatech!prism!ce1zzes

-----------[000167][next][prev][last][first]----------------------------------------------------
Date:      18 Nov 91 22:58:40 GMT
From:      FORDE@POLYVM.BITNET
To:        comp.protocols.tcp-ip
Subject:   RCP

I like to thank all the users whou sent me informantion on lpr.
I am now look for which RFC contains a discription of RCP.

-----------[000168][next][prev][last][first]----------------------------------------------------
Date:      19 Nov 91 00:44:41 GMT
From:      subbu@hpindda.cup.hp.com (MCV Subramaniam)
To:        comp.protocols.tcp-ip
Subject:   A TPI/XTI question.


I have a question regarding TPI/XTI connection establishment functionality.

TPI specs mention that after a TP-user receives a connection indication, it
can send down a T_CONN_RESP on the "listen" queue, while also specifying
a pointer to another queue to "accept" the connection on. The provider
processes the T_CONN_RESP, and sends up a T_OK_ACK message.

The question is, should the message go up the "listen" queue, or the new
queue (assuming they are different queues).

In XTI terms,  I have a program that does the following:

	while(1) {
		t_listen(lfd);
		fork();
		if (child) {
			call t_accept on to accept the connection on a
				new file descriptor.
			close lfd.
			Service connection on new file descriptor.
			exit(0);
		}
	}

When the child process does a t_accept, the provider sends up a T_OK_ACK
message that gets delivered to the parent process, when the parent process
is expecting a T_CONNIND! As a result, I get a TLOOK error that returns
0 on a t_look call!!

Is there anything in XTI or TPI specs that say that I should not fork 
before accepting a connection? (all the programming examples that I saw
forked *after* accepting the connection, and therefore the problem went
away).

The problem would go away if the T_OK_ACK is sent up on the new file
descriptor instead of the listen file descriptor, but I dont see anything in
the TPI specs that say one way or the other.

Any pointers will be appreciated.

Thanks.

-Subbu
--
email: subbu@hpindlm.cup.hp.com               Voice: (408)447-2693.

-----------[000169][next][prev][last][first]----------------------------------------------------
Date:      19 Nov 91 05:00:49 GMT
From:      glen@drifter.nodak.edu (Glen Overby)
To:        comp.protocols.tcp-ip
Subject:   Re: IP maps?

In article <41748@hydra.gatech.EDU> ce1zzes@prism.gatech.EDU (Eric Sheppard) writes:
>I wish to determine the name of a machine from its ip address.  Where
>can I find host tables that I can search?  Please note that I am NOT talking
>about nslookup, which pops out a number given an internet name.  I need to

Host tables?  OH!  you mean the DNS :-) 

You can find what you're looking for in a properly-configured domain using
nslookup and looking for "IP-in-reverse.in-addr.arpa".

For example, the name to 134.129.125.252 is found using:

$ nslookup
> set query=ptr
> 252.125.129.134.in-addr.arpa

i.e.:

#!/usr/local/bin/perl

while($#ARGV > -1) {
	$_ = shift;
	($a, $b, $c, $d) = split(/\./);

	$addr .= "$d.$c.$b.$a.in-addr.arpa.\n";
}

open(NS, "|nslookup") || die "you have to install nslookup first\n";
print NS "set q=ptr\n";
print NS $addr;
close(NS);

(ok, so I don't post-process the output to look for JUST the names).

Glen Overby
overby@plains.nodak.edu

-----------[000170][next][prev][last][first]----------------------------------------------------
Date:      19 Nov 91 05:38:07 GMT
From:      rpw3@rigden.wpd.sgi.com (Rob Warnock)
To:        comp.protocols.tcp-ip
Subject:   Re: Datagram Assembly Timer?

jra@lawday.DaytonOH.NCR.COM (John R. Ackermann x2966) writes:
+---------------
| We've found that our TCP/IP implementation has a very short timer for
| packet assembly; an application that does screen output slowly will
| generate <lots> of tiny packets in a telnet session...
+---------------

See RFC 1122, "Requirements for Internet Hosts -- Communication Layers",
and look at the references to "Nagle's algorithm" (esp. section 4.2.3.4),
which is specified as a "SHOULD" feature. It does what you want, and is
self-adjusting to slow lines.

Caution: Some graphical UI applications (e.g., X Windows) will disable
Nagle's algorithm (in BSD networking, via setsockopt TCP_NODELAY) in order
to achieve "smooth" image motion or mouse input. The ability for the
application to disable Nagle's algorithm is a "MUST" requirement in RFC 1122.


-Rob

-----
Rob Warnock, MS-1L/515		rpw3@sgi.com
Silicon Graphics, Inc.		(415)335-1673
2011 N. Shoreline Blvd.
Mountain View, CA  94039-7311

-----------[000171][next][prev][last][first]----------------------------------------------------
Date:      19 Nov 91 12:21:43 GMT
From:      fabianf@informatik.uni-stuttgart.de (Franz Fabian)
To:        comp.protocols.tcp-ip
Subject:   Implementaion of Internet Stream Protocol (RFC 1190)


1) Does anyone out there know of an implementation of the experimental
internet stream protocol as described in RFC 1190?

2) Has anybody used this protocol for audia/video transmisson over
computer networks?

3) Are there other protocols for voice/video transmission (on the
subnetwork-, network- or transport-level)?

4) has anybody experience in using UDP (or TCP) as a protocol for
voice/video transmission (RFC 1257: isochronous protocols are not
necessary for voice/video transmission, but has anybody done it?).


--
------------
Franz Fabian			Tel.: 0711/7816-385
Informatik, Univ. Stuttgart
Breitwiesenstr. 20/22		SMTP: ffabian@informatik.uni-stuttgart.de
D-7000 Stuttgart 80		X400: ffabian@informatik.uni-stuttgart.dbp.de

-----------[000172][next][prev][last][first]----------------------------------------------------
Date:      19 Nov 91 12:53:44 GMT
From:      vaillan@ireq.hydro.qc.ca (Clement Vaillancourt)
To:        comp.protocols.tcp-ip
Subject:   Re: IP maps?

In article <41748@hydra.gatech.EDU> ce1zzes@prism.gatech.EDU (Eric Sheppard) writes:
>I wish to determine the name of a machine from its ip address.  Where
>can I find host tables that I can search?  Please note that I am NOT talking
>about nslookup, which pops out a number given an internet name.  I need to
>do the opposite.
>Eric Sheppard      Georgia Tech    |   "Of course the US Constitution isn't

You can try to get a copy of dig at one of those sites (only 2 of the
many I got by asking archie.cs.mcgill.ca:

Host uxc.cso.uiuc.edu

    Location: /net
      DIRECTORY drwxr-xr-x       1024  Oct 16 1990  dig

Host valhalla.ee.rochester.edu

    Location: /pub/src
      DIRECTORY drwxrwxr-x        512  Nov 24 1990  dig

Dig with a name will give you the ip number and dig -x ip number will
give you the ip name.

Here is two examples:

>dig uxc.cso.uiuc.edu
	;; ANSWERS:
	uxc.cso.uiuc.edu.       62653   A       128.174.5.50


>dig -x 128.174.5.50
	;; ANSWERS:
	50.5.174.128.in-addr.arpa.      50400   PTR     uxc.cso.uiuc.edu.

I also think that nslookup with the proper syntax will give you an ip number
from a name.

Clem.
-- 
   Clement Vaillancourt, Analyste,   |   Institut de Recherche d'Hydro-Quebec
   Responsable du Reseau Ethernet    |        1800 Montee Ste-Julie, Varennes
   (Analyst, Network Manager)        |             P. Quebec, Canada, J3X 1S1
   vaillan@ireq.hydro.qc.ca   VE2HQJ |Tel:+1 514 652 8238 Fax:+1 514 652 8309

-----------[000173][next][prev][last][first]----------------------------------------------------
Date:      19 Nov 91 15:52:58 GMT
From:      kovar@eclectic.com (David C. Kovar)
To:        misc.entrepreneurs,comp.protocols.tcp-ip
Subject:   Network services - what will be commercially provided in the future?


  I've been doing network related work on and off for about six years.
Initially, vendors providing network services were providing consulting
services, hardware, and software. These days, network services
covers a lot more ground. Just how much ground does it cover now,
and how much ground will it cover in the future?

  Current commercial "network services" include things like:

	GSI - running the NIC for the Internet
	PSI/NEARNet/AltNet - providing regional network connectivity.
	News/mail feeds - extending the Internet farther.

  There are a lot of "free" services out there, too.

	Archie - How to find just about any data repository on the Internet
	WAIS - Public access information searching service/software.
	Gateways - various mail/news gateways.
	Mailing lists, news, email - obvious.

  From a new business point of view, what sorts of services are customers
going to need in the next five years, and are they willing to pay for it?
They're still going to need network design, installation, and management.

How about remote network monitoring and maintenance? The regional networks
do this to some extent. Are corporations or research institutions going
to consider contracting out for someone to run their growing organizational
networks?

How about commercial gateways? Billing is a hassle, to some extent, but
are there any sorts of gateways that people would be willing to pay for
that provided electronic access? How about if you could mail a document
to a Fax gateway?

Are networks to the home ever going to really happen? There are lots of
"mostly one way" networks to the home, and there have been a number of
experimental tries at it, but are the average consumers ready to have
networks in their home? What would those networks do? Home email? Home
automation?

If you're a network customer of some sort, what sorts of services are
you already paying for? What services are you willing to pay for in
the future? What services would you prefer *not* to pay for but you're
afraid you're going to have to?

If you're a network vendor, what services are you already providing? What
are your customers asking for that you're unable, or unwilling, to provide?
What services are you planning to provide in the future?

This is mostly out of professional curiousity. I'm going to have to change
jobs somewhere down the road and I'm wondering what sorts of new jobs,
tasks, and the like are going to show up in the next year or so.

I'd be more than happy to summarize if people will send me email about this.
I'll also keep replies confidential if you don't want to blast your company
secrets all across the net.

TIA.

-- 

-David C. Kovar
	Consultant				Internet: kovar@eclectic.com
	Eclectic Associates			AppleLink: ECLECTIC

-----------[000174][next][prev][last][first]----------------------------------------------------
Date:      19 Nov 91 18:04:05 GMT
From:      greg@duke.cs.unlv.edu (Greg Wohletz)
To:        comp.protocols.ppp,comp.protocols.tcp-ip
Subject:   Re: Point to Point and Terminal Server questions.

In article <91319.080027JFL4@psuvm.psu.edu>, JFL4@psuvm.psu.edu writes:

) I have need to get IP from building-to-building at a small campus.
) started to work on the idea of using PCBridge and line drivers over
) the existing four-wire that exists between these buildings;  however,
) administrative types are *real* uncomfortable with this approach
) (its not high-tech enough and it doesn't cost enough money) :)
) 

Your complaining that they want to spend more money?

Tell them to buy 2 cisco IGS's for about 4k a piece, and you will have
quite a nice setup.

					--Greg

-----------[000175][next][prev][last][first]----------------------------------------------------
Date:      19 Nov 91 18:16:34 GMT
From:      bj@steele.ohsu.edu (Bill Jackson)
To:        comp.sys.hp,comp.protocols.tcp-ip
Subject:   HP Probe and Routers

I am trying to find out something about using HP computers running MPE on a
Wide Area Network using cisco routers.

Our campus network is an Ethernet supporting many types of hosts including two
HP935 machines running MPE.  We access these using WRQ Lan Connect 3000 and
Reflections+ over the network.  We now have a request to access these machines
from a WAN made up of cisco routers.  I have been told that I need to make sure
there is a "probe proxy" on the far side of this network (where the PC clients
are).  I look in the cisco manual and see mention of support of HP Probe, but
thats it!

Can anyone enlighten me as to what this is and what I need to do?  Anyone done
a similar thing.  I have asked cisco and WRQ directly, but am not getting much
help. Information or where to get it on Probe would help!

Please email - I will be glad to post a summary if there are
responses/interest.
-- 
---
William Jackson                     Manager, Network Services, Gaines Hall #113
Oregon Health Sciences University (OHSU), 840 SW Gaines Road, Portland,OR 97201
(503) 494 4535          Internet: bj@ohsu.edu     AT&T Mail: attmail!ohsu3b2!bj

-----------[000176][next][prev][last][first]----------------------------------------------------
Date:      19 Nov 91 20:47:27 GMT
From:      schwartz@latour.cs.colorado.edu (Michael Schwartz)
To:        comp.protocols.tcp-ip
Subject:   announcement of Internet measurement study

Request For Comments 1273 has just been published, about an Internet
measurement study I plan to conduct from January 1992 - November 1992.
Because this study will generate traffic that might be noticed by site
administrators, I am posting this note to help get the word out.

The RFC overviews the goals, experimental design, considerations of
network and remote site load, mechanisms used to control the measurement
collection process, and network appropriate use and privacy issues.

This study is organized in accordance with the guidelines published by
the Internet Activities Board in RFC 1262.

RFCs can be obtained via anonymous FTP from nic.ddn.mil,
ftp.nisc.sri.com, nis.nsf.net, nisc.jvnc.net, venera.isi.edu, or
wuarchive.wustl.edu.  RFCs 1262 and 1273 are also available by anonymous
FTP from latour.cs.colorado.edu, in pub/InetMeasStudy.

 - Mike Schwartz
   Asst. Professor
   Dept. of Computer Science
   Univ. of Colorado - Boulder
   Internet: schwartz@cs.colorado.edu

-----------[000177][next][prev][last][first]----------------------------------------------------
Date:      19 Nov 91 21:00:48 GMT
From:      andyk@hoh.mbl.edu (Andy Kogelnik)
To:        comp.protocols.tcp-ip
Subject:   Mac/NFS


Does anyone know who makes Mac/NFS?  Anyone have any usage comments on
it especially in regard to use in conjunction with PC-NFS.

Thanks

Andy Kogelnik
Network Manager
Marine Biological Laboratory
Woods Hole, MA 02543
andy@mbl.edu

-----------[000178][next][prev][last][first]----------------------------------------------------
Date:      19 Nov 91 21:03:34 GMT
From:      andyk@hoh.mbl.edu (Andy Kogelnik)
To:        comp.protocols.tcp-ip
Subject:   Using TCP/IP from a C program


Does anyone know of any C libraries which facilitate the use of tcp/ip
protocols and/or rlogin features from within a C program?

Thanks


Andy Kogelnik
Network Manager
Marine Biological Laboratory
Woods Hole, MA 02543
andy@mbl.edu

-----------[000179][next][prev][last][first]----------------------------------------------------
Date:      19 Nov 91 21:08:21 GMT
From:      craig@sics.se (Craig Partridge)
To:        comp.protocols.tcp-ip
Subject:   Re: Implementaion of Internet Stream Protocol (RFC 1190)

In <FABIANF.91Nov19132143@pauke.informatik.uni-stuttgart.de> fabianf@informatik.uni-stuttgart.de (Franz Fabian) writes:

>1) Does anyone out there know of an implementation of the experimental
>internet stream protocol as described in RFC 1190?

Yes, there are three.  Bolt Beranek and Newman, The Swedish Institute
of Computer Science, and IBM Heidelberg have all done, or are doing
implementations.  The BBN implementation and the SICS implementation are
available for public FTP (the Swedish one is beta test only).  A
paper on the Swedish implmentation has been submitted to Journal
of Internetworking.

>2) Has anybody used this protocol for audia/video transmisson over
>computer networks?

Yes, both BBN and SICS have done experiments (BBN I believe more extensively).  
Furthermore, ST-I, which RFC-1190 revises, has been in heavy use on the
Internet video conferencing system that BBN runs for DARPA.  Among other
tests SICS and University College London ran a test running ST-II 
inside IP between London and Sweden (which didn't quite work since
IP doesn't support flow reservations of any form but we did get
half-useful communication some of the time, which was intriguing).

I don't know what has been done at IBM.

>3) Are there other protocols for voice/video transmission (on the
>subnetwork-, network- or transport-level)?

Yes, PVP (Packet Video Protocol) and NVP (Network Voice Protocol) are
transport protocols designed to run over ST.  PVP and NVP have been
ported to both the BBN and ICS implementations.  In both cases,
I believe the PVP and NVP code was supplied by Information Sciences
Institute (ISI).

>4) has anybody experience in using UDP (or TCP) as a protocol for
>voice/video transmission (RFC 1257: isochronous protocols are not
>necessary for voice/video transmission, but has anybody done it?).

Experiments are currently underway on the DARPA DARTNET testbed
on just this sort of question.  I'm not in a position to discuss the
work more fully as I'd probably get it wrong.

Craig Partridge
BBN (back from sabbatical at SICS)

-----------[000180][next][prev][last][first]----------------------------------------------------
Date:      20 Nov 91 00:28:26 GMT
From:      Frank@mindlink.bc.ca (Frank I. Reiter)
To:        comp.protocols.tcp-ip
Subject:   gethostby????() use

I am hoping somebody can point out the flaw in the following very simple code.
I have a function thus:

int get_host_entry(s,buffer)        /* Get a host address. */
register char *s;
struct hostent *buffer;
{
    struct hostent *hostentry;
    unsigned long address;

    if ((*s >= '0') && (*s <= '9')) {      /* IP address. */
        address = inet_addr(s);
        if(address == (unsigned long)-1)
            return(ERROR);
        hostentry=gethostbyaddr((char *)&address,sizeof(address),AF_INET);
    } else {                               /* Host name. */
     hostentry = gethostbyname(s);
    }
    if(hostentry == NULL) {
        return(ERROR);
    }
    if(buffer != NULL)
        memcpy(buffer, hostentry, sizeof(buffer));
    return(!ERROR);                          /* Success. */
}

The above appears to me to work but opinions to the contrary are invited.  My
problem is that the scructure it fills always seems to be invalid.  I call it
in the following code fragment:

   /* Show off how much we know about this site */
    if(get_host_entry(*argv,&hostentry)==ERROR) {
        printf("Cannot locate %s",*argv);
        newline();
        return;
    }
    printf("Official name: %s",hostentry.h_name);
    newline();
/*
    if(hostentry.h_aliases[0] && *hostentry.h_aliases[0]) {
        fputs("Aliases:       ",stdout);
        for(i=0;hostentry.h_aliases[i] && *hostentry.h_aliases[i];++i) {
            if(i)
                fputs(", ",stdout);
            fputs(hostentry.h_aliases[i],stdout);
        }
        newline();
    }
*/
    if(hostentry.h_addrtype == AF_INET) {
        fputs("Address(es):  ",stdout);
        for(i=0;hostentry.h_addr_list[i];++i) {
            if(i)
                fputs(", ",stdout);
            fputs(inet_ntoa(hostentry.h_addr_list[i]),stdout);
        }
        newline();
    }

---- end of code fragment -----

With the alias secion commented out as above the code runs and successfully
prints the full host name when given a partial name (for a local host) or a
valid IP address (for any host I've tried.), but does not print any address
information, implying that hostentry.h_addrtype != AF_INET (which I understand
is currently the only legal value).

If the alias section is NOT commented out this code dumps core.

I'm running SCO ODT on an 80386.  Can anyone see what I am doing wrong?

Frank.
--
___________________________________________________________________
Frank I. Reiter                UUCP:  ...!van-bc!rsoft!frank
Reiter Software Inc.       INTERNET: frank@mindlink.bc.ca (prefered)
Surrey, British Columbia             frank@rsoft.bc.ca

-----------[000181][next][prev][last][first]----------------------------------------------------
Date:      20 Nov 91 01:15:58 GMT
From:      Frank@mindlink.bc.ca (Frank I. Reiter)
To:        comp.protocols.tcp-ip
Subject:   struct sockaddr_ns

I'm trying to get the code from Steven's "Unix Network Programming" running
under SCO ODT and have run into a structure type that isn't mentioned in SCO's
include files.

What is a struct sockaddr_ns?  Presumably it would be defined in the non
existant include file <netnss.h> which I have commented out of xns.h.

Frank.
--
___________________________________________________________________
Frank I. Reiter                UUCP:  ...!van-bc!rsoft!frank
Reiter Software Inc.       INTERNET: frank@mindlink.bc.ca (prefered)
Surrey, British Columbia             frank@rsoft.bc.ca

-----------[000182][next][prev][last][first]----------------------------------------------------
Date:      20 Nov 91 02:41:41 GMT
From:      knight@henson.cc.wwu.edu (Matt J. Scott)
To:        comp.protocols.tcp-ip
Subject:   Examples in Stevens book (UNIX Network programming)

I have went out and bought UNIX Network Programming by W. Richard
Stevens (upon the advice that I observed on the net).  The book is great
but I was wondering if there is some place that has the examples in
machine readable form!  There is no mention of the code being availble
(or at least I could find no refference to it).  Any help, even just
portions of the examples, would make my education on this subject that
much easier.

Thanks in advance!


knight@cs.wwu.edu
"Thats the problem with reality--it's taken far to seriously"

-----------[000183][next][prev][last][first]----------------------------------------------------
Date:      20 Nov 91 03:33:26 GMT
From:      jtw@lcs.mit.edu (John Wroclawski)
To:        comp.protocols.tcp-ip
Subject:   Re: Implementaion of Internet Stream Protocol (RFC 1190)

In article <1991Nov19.210821.24248@sics.se> craig@sics.se (Craig Partridge) writes:

   >4) has anybody experience in using UDP (or TCP) as a protocol for
   >voice/video transmission (RFC 1257: isochronous protocols are not
   >necessary for voice/video transmission, but has anybody done it?).

   Experiments are currently underway on the DARPA DARTNET testbed
   on just this sort of question....

Some folks at MIT have built a PC-based box which uses commercially
available components to send video and voice using UDP. This box has
been duplicated at Xerox PARC and the Lawrence Berkeley Lab, and the
three places have used it to communicate over both DARTNet and the
generic internet. As with the work Craig describes, results are
somewhat probabilistic, because neither DARTNet nor the internet
support resource (link bandwidth, etc.) allocation algorithms at the
moment. This box is a first-generation proof of concept; MIT and many
other places are continuing to explore architectures and technology
for packet video.

The key missing element in all of this is a network resource
allocation strategy. It does not matter whether you are using a stream
protocol and managing resources on a per-stream basis or using (say)
UDP and managing resources on a per-class basis; you still need
algorithms for allocating bandwidth, performing admission control
(deciding whether the net can handle a new service request), and
actually managing links and queues to provide the required service
guarantees. Developing these algorithms (and, for that matter,
deciding exactly what service qualities the network can or should
provide and how the client should specify its needs) are extremely
active research topics today. We need the results of this work before
ST, UDP or anything else will really be useful for carting around
video and voice in general-purpose mixed-use networks.

				-john
				 jtw@lcs.mit.edu

-----------[000184][next][prev][last][first]----------------------------------------------------
Date:      20 Nov 91 03:39:50 GMT
From:      yanek@mthvax.cs.miami.edu (Yanek Martinson)
To:        comp.protocols.tcp-ip
Subject:   QUE: internet RFCs available as a mailing list?

schwartz@latour.cs.colorado.edu (Michael Schwartz) writes:

RFCs can be obtained via anonymous FTP from nic.ddn.mil,
>ftp.nisc.sri.com, nis.nsf.net, nisc.jvnc.net, venera.isi.edu, or
>wuarchive.wustl.edu.  RFCs 1262 and 1273 are also available by anonymous
>FTP from latour.cs.colorado.edu, in pub/InetMeasStudy.

Is there any mailing list, so RFCs could automatically be received as
they come out?
-- 
yanek@mthvax.cs.miami.edu
safe0%yanek@mthvax.cs.miami.edu

-----------[000185][next][prev][last][first]----------------------------------------------------
Date:      20 Nov 91 04:05:36 GMT
From:      dan@cedb.UUCP (Dan Busarow)
To:        comp.protocols.tcp-ip,comp.protocols.nfs,comp.unix.sysv386
Subject:   NFS portmapper

Hello all,

I am hoping that someone could provide me with a list
of things to check when portmap won't run.  The system
is running SCO UNIX 3.2.2 , TCP/IP 1.1.3 and NFS 1.1.0o.

There are two systems installed at the site, one works
the other doesn't.  All of the files I know to look at
like hosts, exports, services, the tcp and nfs startup
scripts are identical except for IP addresses in 'tcp'.

When NFS starts up portmap is executed and produces no
messages, however lockd says it cannot register and once
the system is up, ps shows that portmap is not running.
Running portmap manually produces the same results, no
error messages and it has already died when you check
with ps.

Also, when I installed these systems, it was working
since I mounted remote drives to copy files back and
forth to get them in sync (one is used as a backup
processor).

Ideas anyone??

Thanks in advance.

Dan

-- 
Dan Busarow                  dan%cedb@uunet.uu.net        uunet!cedb!dan
DPC Systems                  Laguna Hills, CA
(714) 852-5254 O             (714) 362-8306 H

-----------[000186][next][prev][last][first]----------------------------------------------------
Date:      20 Nov 91 04:51:06 GMT
From:      mizuyama@dixie.sws.cpd.mei.co.jp (Masashige Mizuyama)
To:        comp.protocols.tcp-ip
Subject:   Dynamic Host Configuration Protocol


  I've recently read the INTERNET-DRAFT of Dynamic Host Configuration
Protocol(July, 1991).

1. How is the current status of the draft?  Does anyone know
   when it will become an RFC standard?
2. Can anybody point me to a public software implementing that protocol?
3. I know that there are also some other protocols for the similar
   purpose, such as DRARP of Sun and NIP of MIT.  Which platforms
   can run them?

Thanks.
--
Masashige Mizuyama <mizuyama@sws.cpd.mei.co.jp>
Panasonic, Japan.

-----------[000187][next][prev][last][first]----------------------------------------------------
Date:      20 Nov 91 09:17:47 GMT
From:      edwin@cs.ruu.nl (Edwin Kremer)
To:        comp.protocols.tcp-ip
Subject:   Re: Examples in Stevens book (UNIX Network programming)

In <1991Nov20.024141.23218@henson.cc.wwu.edu> knight@henson.cc.wwu.edu
(Matt J. Scott) writes:

   | I have went out and bought UNIX Network Programming by W. Richard
   | Stevens (upon the advice that I observed on the net).  The book is great
   | but I was wondering if there is some place that has the examples in
   | machine readable form!

Yes, the code is available. Here are the "how to get" details:

________________________________________________________________________
  We, Computer Science department, Utrecht University,  are  running  an
anonymous  FTP  server  on  one  of  our systems. In addition to the FTP
service we're also running a mail  server,  for  those  of  you  without
direct Internet access.


--> How to get 'stevens-book-code' via anonymous FTP:

	Site:		archive.cs.ruu.nl  [131.211.80.5]
	Login:		"anonymous" or "ftp"
	Password:	your own email address (you@your_domain)
	File:		UNIX/stevens-book-code.tar.Z


--> How to get 'stevens-book-code' via e-mail from our mail-server:

    NOTE: In the following I have assumed that your mail address is
	  "fred_flintstone@stone.age.edu"; of course you must substitute
	  your own address for this.
	  ** PLEASE USE VALID DOMAIN ADDRESSES. DO NOT USE ADDRESSES **
	  ** WITH ! and @ MIXED !!!! BITNETTERS USE USER@HOST.BITNET **


    Send the following message to
		mail-server@cs.ruu.nl
    or the old-fashioned path alternative
		uunet!mcsun!hp4nl!ruuinf!mail-server


      begin
      path fred_flintstone@stone.age.edu (SUBSTITUTE *YOUR* ADDRESS)
      send UNIX/stevens-book-code.tar.Z
      end


The path command can be deleted if we receive a valid from address in your
message. If this is the first time you use our mail server, we suggest you
first issue the request:

      send HELP


  A complete "ls-lR" listing of the archive is  kept  in  the  top-level
directory, it will be updated every night. To get it, issue the command:

     send ls-lR.Z



  That's all for now. If you encounter problems using  the  FTP  service
and/or the mail-server, feel free to drop me a line (by e-mail, please).


						--[ Edwin ]--
-- 
Edwin Kremer, systems and network administrator.  [NIC-Whois handle: EHK3]
     Department of Computer Science,  Utrecht University,  The Netherlands
     Email: edwin@cs.ruu.nl  | UUCP to: ...!uunet!mcsun!hp4nl!ruuinf!edwin

-----------[000188][next][prev][last][first]----------------------------------------------------
Date:      20 Nov 91 16:36:24 GMT
From:      mikel@intercon.com (Michael W. Wellman)
To:        comp.protocols.tcp-ip
Subject:   Re: Examples in Stevens book (UNIX Network programming)

In article <1991Nov20.024141.23218@henson.cc.wwu.edu>, 
knight@henson.cc.wwu.edu (Matt J. Scott) writes:
> The book is great
> but I was wondering if there is some place that has the examples in
> machine readable form!

--> Another how to get 'stevens-book-code' via anonymous FTP:
 
 	Site:		uunet.uu.net  [137.39.1.2]
 	Login:		"anonymous" or "ftp"
 	Password:	your own email address (you@your_domain)
 	File:		published/stevens-book-code.tar.Z
 	or File:	published/stevens-netprog.tar.Z

mikel

-----------[000189][next][prev][last][first]----------------------------------------------------
Date:      20 Nov 91 16:39:47 GMT
From:      rick@uunet.uu.net (Rick Adams)
To:        comp.protocols.tcp-ip
Subject:   Re: Examples in Stevens book (UNIX Network programming)


uunet.uu.net:published/stevens-book-code.tar.Z

-----------[000190][next][prev][last][first]----------------------------------------------------
Date:      20 Nov 91 17:00:21 GMT
From:      D.Nash@utexas.edu (Donald L. Nash)
To:        comp.protocols.tcp-ip
Subject:   Re: TGV Multinet traceroute question

In article <1991Nov18.193509.11378@eco.twg.com>, reece@eco.twg.com
(Reece R. Pollack) writes:
>It's not necessary to install traceroute with SYSPRV if it is being
>used
>only by users who have access to the SYSPRV privilege. You may wish to
>leave it uninstalled (or installed without SYSPRV) to prevent abuse by
>non-privileged users. Traceroute, like Ping, can generate a fair
>amount
>of network traffic, and most users have no need for it.

This was all discussed on the info-multinet mailing list back in July. 
Traceroute needs SYSPRV because it needs to open a raw socket.  Ping
also needs SYSPRV because it also opens a raw socket.  MultiNet
automatically installs ping with SYSPRV, but not traceroute, and
traceroute can't generate any more traffic than ping can.  Although this
was never said during the discussion, I think that leaving out
traceroute was just an oversight.  TGV has indicated that it is safe to
install traceroute with SYSPRV (i.e. it won't compromise your system
security).  However, your point about wanting to limit access to
traceroute is nevertheless still valid.  But if you're going to keep
unpriv'ed people from using traceroute then you should un-install ping
as well, since it can generate more traffic than traceroute can (see the
/PRELOAD qualifier).

If we want to pursue this thread any further, we should take it out of
the newsgroup since this is turning into a VMS issue.

				++Don

-----------[000191][next][prev][last][first]----------------------------------------------------
Date:      20 Nov 91 17:09:24 GMT
From:      Frank@mindlink.bc.ca (Frank I. Reiter)
To:        comp.protocols.tcp-ip
Subject:   Re: struct sockaddr_ns

> I wrote:
> 
> I'm trying to get the code from Steven's "Unix Network Programming" running
> under SCO ODT and have run into a structure type that isn't mentioned in
> SCO's
> include files.
> 
> What is a struct sockaddr_ns?  Presumably it would be defined in the non
> existant include file <netnss.h> which I have commented out of xns.h.

The author replied:

> rstevens@noao.edu writes:
> 
> It's a Xerox Network Systems' (XNS) socket address structure.
> SCO probably has nothing to support it, as I doubt SCO supports
> the XNS protocols at all.  You were right to comment the stuff out.
> 
>         Rich Stevens  (rstevens@noao.edu)

Thanks also to several others who answered similarly.

Frank.
--
___________________________________________________________________
Frank I. Reiter                UUCP:  ...!van-bc!rsoft!frank
Reiter Software Inc.       INTERNET: frank@mindlink.bc.ca (prefered)
Surrey, British Columbia             frank@rsoft.bc.ca

-----------[000192][next][prev][last][first]----------------------------------------------------
Date:      20 Nov 91 18:41:44 GMT
From:      sap@crosfield.co.uk (stelios pavlides)
To:        comp.protocols.tcp-ip
Subject:   Basic LLC/SNAP questions

I would appreciate having the simple answers to these simple questions:

1 Are LLC SAP numbers kept in a central registry? By which authority? Is one
  able to get a new number and how?

2 Which document is the official SNAP specification? How can I get it over
  the net?

Thanks.

-- 
Stelios Pavlides			Phone  :  +44 442 230000 ext 3468
Crosfield Electronics Ltd		Fax    :  +44 442 232301
Hemel Hempstead, Herts. HP2 7RH, UK	E-mail :  sap@crosfield.co.uk 
-----------------------------------------------------------------------------

-----------[000193][next][prev][last][first]----------------------------------------------------
Date:      20 Nov 91 23:17:56 GMT
From:      orath@supernet.dallas.haus.com (Omkar Rath)
To:        comp.protocols.tcp-ip
Subject:   Ques. on internet accessibility (nothing to do with SRI)

This question has nothing to do with how to get on the Internet.

I am trying to access (via telnet, ping etc.) a host with IP address say
81.22.33.44 from another with IP address 89.36.24.35 and needless to say it
does not work.  Both the hosts are on the same physical eth-lan.  The reason
it does not work (at least I found that out) is that the ethernet adapters on
the two hosts have "netmasks" set to ff000000 (even with "root" access I could
not succeed to change the netmasks (used the "ifconfig" command) and at this
point my guess is that the netmask is a function of the IP address and cannot
be changed).

No sooner I change the IP address of the first host to 89.22.33.44, of course
everything works.

So am I correct in concluding that hosts on same physical e-lans have to have
same network numbers (i.e the network part of the IP address)?

If I am correct then how do hosts on different e-lans (with diff network
numbers) talk to each other?  My knowledge is poor on bridges, routers etc.
Is there an entity (bridge or router) between the two e-lans making packets
on one net appear as if it was on the second net etc?

I have access to a cross-com bridge with IMS software but at this point I do
not know whether I should follow this avenue?


Thanks a lot in advance,
-- 
Omkar Rath
INTERNET ADDRESS : orath@supernet.haus.com
USENET ADDRESS : uunet!iex!supernet!orath
Work Phone : (214)-386-2245

-----------[000194][next][prev][last][first]----------------------------------------------------
Date:      21 Nov 91 00:11:24 GMT
From:      dyer@spdcc.com (Steve Dyer)
To:        comp.protocols.tcp-ip
Subject:   Re: the slimeballs are at it again

It can be convincingly argued that the appropriate response to a perceived
smear campaign is simply to proceed with what was hopefully the status quo,
namely, your scrupulously clean behavior.  If you have to say something
publically, a simple, discreet message identifying your (lack of a) relation-
ship to all vendors should suffice.  Yelling "I am not on cisco's take"
for several screenfuls just serves to bring attention to something which
doesn't deserve attention.

-- 
Steve Dyer
dyer@ursa-major.spdcc.com aka {ima,harvard,rayssd,linus,m2c}!spdcc!dyer

-----------[000195][next][prev][last][first]----------------------------------------------------
Date:      21 Nov 91 05:59:31 GMT
From:      barmar@think.com (Barry Margolin)
To:        comp.protocols.tcp-ip
Subject:   Re: Ques. on internet accessibility (nothing to do with SRI)

In article <1991Nov20.231756.23667@supernet.dallas.haus.com> orath@supernet.dallas.haus.com (Omkar Rath) writes:
>If I am correct then how do hosts on different e-lans (with diff network
>numbers) talk to each other?  My knowledge is poor on bridges, routers etc.
>Is there an entity (bridge or router) between the two e-lans making packets
>on one net appear as if it was on the second net etc?

There will probably be responses that try to explain some details of how
TCP/IP internetworking works, but I don't really think it's appropriate to
try to teach it in this forum.  Rather, I suggest you get the book
"Internetworking with TCP/IP, Volume I" by Douglas Comer.
-- 
Barry Margolin, Thinking Machines Corp.

barmar@think.com
{uunet,harvard}!think!barmar

-----------[000196][next][prev][last][first]----------------------------------------------------
Date:      21 Nov 91 11:57:49 GMT
From:      hartl@Informatik.TU-Muenchen.DE (Anton Hartl)
To:        comp.protocols.tcp-ip
Subject:   Re: Examples in Stevens book (UNIX Network programming)

knight@henson.cc.wwu.edu (Matt J. Scott) writes:
>but I was wondering if there is some place that has the examples in
>machine readable form! 

As this may be of general interest: the code is available from
    archive.cs.ruu.nl:/pub/UNIX/stevens-book-code.tar.Z
    ftp.informatik.tu-muenchen.de:/pub/programming/lib/stevens-book-code.tar.Z

-Toni
_____________________________________________________________________________
_Anton Hartl___________________________Anton.Hartl@informatik.tu-muenchen.de_

-----------[000197][next][prev][last][first]----------------------------------------------------
Date:      21 Nov 91 12:20:46 GMT
From:      korsberg@kirk (Ed Korsberg)
To:        comp.protocols.tcp-ip
Subject:   Re: Basic LLC/SNAP questions

In article <11926@suns4.crosfield.co.uk> sap@crosfield.co.uk (stelios pavlides) writes:
>I would appreciate having the simple answers to these simple questions:
>
>1 Are LLC SAP numbers kept in a central registry? By which authority? Is one
>  able to get a new number and how?

The IEEE Standards Office is the keeper of the LLC SAP numbers. Talk to
Vincent Condello at IEEE (212-705-7900 ?)


>2 Which document is the official SNAP specification? How can I get it over
>  the net?

Try IEEE Std 802-1990, Overview and Architecture.

Ed Korsberg             E-mail: korsberg@aa.ab.com
Allen Bradley Inc.      phone:  313-998-2470
555 Briarwood Circle
Ann Arbor, Mich 48108
-- 
Ed Korsberg             E-mail: korsberg@aa.ab.com
Allen Bradley Inc.      phone:  313-998-2470
555 Briarwood Circle
Ann Arbor, Mich 48108

-----------[000198][next][prev][last][first]----------------------------------------------------
Date:      21 Nov 91 16:45:35 GMT
From:      ced@bcstec.uucp (Charles Derykus)
To:        comp.protocols.tcp-ip
Subject:   DNS ghost server


Our network has an ex-nameserver, call it "ghost". "ghost" has changed
addresses several times in its non-nameserving life. 

The problem is that "ghost" (and all its newer addresses) lives on in
the mind of BIND. Incidentally, we are running 4.8.1 in a multi-platform 
environment (mostly Sun Sparc 4.1.1). 

"DIG" gives the following:

; <<>> DiG 2.0 <<>> @primary_server sub1.sub0 ns 
;; ->>HEADER<<- opcode: QUERY , status: NOERROR, id: 10
;; flags: qr aa rd ra ; Ques: 1, Ans: 7, Auth: 0, Addit: 13
;; QUESTIONS: 
;;      sub0.com, type = NS, class = IN

;; ANSWERS:
sub1.sub0  120     NS      ghost.sub1.sub0  ; 2 min 

;; ADDITIONAL RECORDS:
ghost.sub1.sub0  542745  A       original_ns_ip  ; 6 days 6 hrs 45 min 45 sec 
ghost.sub1.sub0  468348  A       non_ns_ip1      ; 5 days 10 hrs 5 min 48 sec 
ghost.sub1.sub0  550571  A       non_ns_ip2      ; 6 days 8 hrs 56 min 11 sec 
ghost.sub1.sub0  581952  A       non_ns_ip3      ; 6 days 17 hrs 39 min 1 2 sec 
ghost.sub1.sub0  604800  A       newest_non_ns_ip ; 1 wk 

Various efforts to purge the ghost from the mind of BIND have failed. (Who 
you gonna call?)

We have fine combed all our servers. There are no "ghost" delegations any
more. I don't understand why time doesn't run out on the "ghost" when 
the  "TTL" expires. Are there other explantions to account for the life
of the ghost?  It's a pain because resolvers are picking up the old
addresses for ghost too. 

Any suggestions will be greatly appreciated.

Charles DeRykus				Internet:   ced@carios2.boeing.com
Boeing Computer Services		UUCP:	    ...!uunet!bcstec!ced
Bellevue, WA.  M/S 7M-HA		(206) 957-5326  FAX: (206) 957-5050

-----------[000199][next][prev][last][first]----------------------------------------------------
Date:      21 Nov 91 17:20:35 GMT
From:      shih@maple.eecs.ucdavis.edu (Alan S. Shih)
To:        comp.protocols.tcp-ip
Subject:   Help in SLIP!

Hello netlanders:

Help, I have a unix.386 box and it is running SLIP to connect to the
net.  In the past, I have been using a 2400baud modem which runs 
spot less, except the speed. 

However, I upgrade it to a 9600baud modem and thought is would be a
piece of cake to install it...just swap modems..however, I was wrong.
The modem works fine when I use 'cu' but when I try to runn slip on
top of it, telnet would not properly or simply just won't work. I
cannot telnet more than  once out of my unix.box! Nor any xwin clients
can connect (which will simply hang the connection)...

I am out of ideas..Pls someone Help!!! Any info is greatly
appreciated! (btw I have disabled V42bis and MNP5, still no luck)

-- 
|   ALAN SHIH  
|   UC.Davis EE dept
|   "When I realize my life is full of sh*t, it was good;
|      because I can start cleaning it UP."

-----------[000200][next][prev][last][first]----------------------------------------------------
Date:      21 Nov 91 19:05:17 GMT
From:      rickert@mp.cs.niu.edu (Neil Rickert)
To:        comp.protocols.tcp-ip
Subject:   Re: DNS ghost server

In article <1501@bcstec.boeing.com> ced@bcstec.uucp (Charles Derykus) writes:
>
>The problem is that "ghost" (and all its newer addresses) lives on in
>the mind of BIND. Incidentally, we are running 4.8.1 in a multi-platform 
>environment (mostly Sun Sparc 4.1.1). 
>
>"DIG" gives the following:
>
>; <<>> DiG 2.0 <<>> @primary_server sub1.sub0 ns 
>;; ->>HEADER<<- opcode: QUERY , status: NOERROR, id: 10
>;; flags: qr aa rd ra ; Ques: 1, Ans: 7, Auth: 0, Addit: 13
>;; QUESTIONS: 
>;;      sub0.com, type = NS, class = IN
>
>;; ANSWERS:
>sub1.sub0  120     NS      ghost.sub1.sub0  ; 2 min 
>
>;; ADDITIONAL RECORDS:
>ghost.sub1.sub0  542745  A       original_ns_ip  ; 6 days 6 hrs 45 min 45 sec 
>ghost.sub1.sub0  468348  A       non_ns_ip1      ; 5 days 10 hrs 5 min 48 sec 
>ghost.sub1.sub0  550571  A       non_ns_ip2      ; 6 days 8 hrs 56 min 11 sec 
>ghost.sub1.sub0  581952  A       non_ns_ip3      ; 6 days 17 hrs 39 min 1 2 sec 
>ghost.sub1.sub0  604800  A       newest_non_ns_ip ; 1 wk 

  Well, I just tried: dig ghost.sub1.sub0
and got back NXDOMAIN :-) .

  Seriously, if you want help with these kinds of problems, tell the
truth, the whole truth, and nothing but the truth.  There are plenty of
people on the net who would do some DNS lookups to see what is happening.
But when you insist on giving phony names instead of the real name, they
cannot help you.

  Most likely what happened is that the root nameservers were never asked
to delete the old NS and GLUE records.  Until that problem is corrected the
out of date information will keep circulating.

-- 
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
  Neil W. Rickert, Computer Science               <rickert@cs.niu.edu>
  Northern Illinois Univ.
  DeKalb, IL 60115                                   +1-815-753-6940

-----------[000201][next][prev][last][first]----------------------------------------------------
Date:      21 Nov 91 20:38:24 GMT
From:      johnp@lupulus.ssc.gov (John A. Palkovic)
To:        comp.protocols.tcp-ip
Subject:   PTR woes

The domain in which my workstation lives, ssc.gov, is split into
two class B networks: 134.3.*.* and 143.202.*.*. Earlier this
week, I moved from the first to the second. Now I seem to be a
second class net.citizen. The following script illustrates the
problem:

Script started on Thu Nov 21 13:59:13 1991
bash$ ftp ftp.uu.net
Connected to uunet.UU.NET.
421- Can't map 143.202.80.30 to a valid hostname
    We only allow ftp from recognized sites.  If your host has
    no PTR record have your system administrator fix your name
    servers; otherwise, please try again later.
421 anonymous ftp refused.
ftp> quit
bash$ mail -v nobody@linac.fnal.gov
Subject: test
Hi there
.
EOT
bash$ nobody@linac.fnal.gov... Connecting to linac.fnal.gov (TCP)...
220 linac.fnal.gov 5.65c/IDA-1.4.3 Sendmail open for business at Thu, 21 Nov 1991 13:59:39 -0600
>>> HELO lupulus.ssc.gov
 250 Hello [143.202.80.30], why do you call yourself lupulus.ssc.gov?
>>> MAIL From:<johnp@lupulus.ssc.gov>
 250 <johnp@lupulus.ssc.gov>... Sender ok
>>> RCPT To:<nobody@linac.fnal.gov>
 250 <nobody@linac.fnal.gov>... Recipient ok
>>> DATA
 354 Enter mail, end with "." on a line by itself
>>> .
 250 Ok
>>> QUIT
221 linac.fnal.gov closing connection
nobody@linac.fnal.gov... Sent

bash$ nslookup
Default Server:  lupulus.ssc.gov
Address:  143.202.80.30

> set q=soa
> ssc.gov
Server:  lupulus.ssc.gov
Address:  143.202.80.30

ssc.gov
        origin = sscnet.ssc.gov
        mail addr = dasson.sunova.ssc.gov
        serial = 911122
        refresh = 1800 (30 mins)
        retry   = 300 (5 mins)
        expire  = 604800 (7 days)
        minimum ttl = 21600 (6 hours)
> server sscnet.ssc.gov
Default Server:  sscnet.ssc.gov
Address:  134.3.1.100

> set q=ptr
> 143.202.80.30
Server:  sscnet.ssc.gov
Address:  134.3.1.100

30.80.202.143.in-addr.arpa      name = lupulus.ssc.gov
> set q=a
> lupulus.ssc.gov
Server:  sscnet.ssc.gov
Address:  134.3.1.100

Name:    lupulus.ssc.gov
Address:  143.202.80.30

> bash$ exit

script done on Thu Nov 21 14:00:36 1991

It appears that valid, up to date A and PTR records exist on the
authority servers for my machine. Yet nameservers cannot match
my address with my hostname. What is wrong?

-John
--
hoser@lupulus.ssc.gov
I joined the League for Programming Freedom -- write to league@prep.ai.mit.edu
Not speaking for the SSC, the DOE, or the URA

-----------[000202][next][prev][last][first]----------------------------------------------------
Date:      21 Nov 91 21:49:58 GMT
From:      itexjct!warren%itexjct@uunet.uu.net (Warren Burstein)
To:        comp.protocols.tcp-ip
Subject:   anon ftp setup on Sun 3/60

We have SunOs 4.1.1, I set up anon ftp just like it says in the man
page, but when I log in and run dir, this is what I get

$ ftp there
Connected to there.
220 there FTP server (SunOS 4.1) ready.
Name (there:): anonymous
Password (there:anonymous): 
331 Guest login ok, send ident as password.
230 Guest login ok, access restrictions apply.
ftp> dir
200 PORT command successful.
150 ASCII data connection for /bin/ls (147.161.1.3,3773) (0 bytes).
crt0: no /usr/lib/ld.so
226 ASCII Transfer complete.
25 bytes received in 0.16 seconds (0.15 Kbytes/s)

Here's ftp's password entry:
ftp:*:1193:63:anonymous FTP:/usr/spool/ftp:

And here's what ~ftp looks like
dr-xr-xr-x  2 root          512 Nov 21 17:53 bin
dr-xr-xr-x  2 root          512 Nov 21 20:12 dev
dr-xr-xr-x  2 root          512 Nov 21 17:54 etc
dr-xr-xr-x  3 root          512 Nov 21 17:56 pub
dr-xr-xr-x  3 root          512 Nov 21 17:57 usr

/usr/spool/ftp/bin:
---x--x--x  1 root        11064 Oct 14  1990 ls

/usr/spool/ftp/dev:
crw-rw-rw-  1 root       3,  12 Nov 21 20:12 zero

/usr/spool/ftp/etc:
---x--x--x  1 root           91 Nov 21 17:54 group
---x--x--x  1 root          222 Nov 21 17:55 passwd

/usr/spool/ftp/usr:
dr-xr-xr-x  2 root          512 Nov 21 20:13 lib

/usr/spool/ftp/usr/lib:
---x--x--x  1 root        24576 Oct 14  1990 ld.so
-r-xr-xr-x  1 root       483328 Oct 15  1990 libc.so.0.15

I probably have a file mode wrong or something is missing but can't tell
what it is.
-- 
/|/-\/-\       The entire world
 |__/__/_/     is a very strange carrot
 |warren@      But the farmer
/ worlds.COM   is not worried at all.

-----------[000203][next][prev][last][first]----------------------------------------------------
Date:      21 Nov 91 22:53:00 GMT
From:      gavron@alpha.sunquest.com (Ehud Gavron 602-885-7700x.2546)
To:        comp.protocols.tcp-ip
Subject:   Re: PTR woes

In article <JOHNP.91Nov21143824@lupulus.ssc.gov>, johnp@lupulus.ssc.gov (John A. Palkovic) writes...
#The domain in which my workstation lives, ssc.gov, is split into
#two class B networks: 134.3.*.* and 143.202.*.*. Earlier this
#week, I moved from the first to the second. Now I seem to be a
#second class net.citizen. The following script illustrates the
#problem:

..
# 
#bash$ nslookup
#Default Server:  lupulus.ssc.gov
#Address:  143.202.80.30
# 
#> set q=soa
#> ssc.gov
#Server:  lupulus.ssc.gov
#Address:  143.202.80.30
# 
#> set q=ptr
#> 143.202.80.30
#Server:  sscnet.ssc.gov
#Address:  134.3.1.100
# 
#30.80.202.143.in-addr.arpa      name = lupulus.ssc.gov
#> set q=a
#> lupulus.ssc.gov
#Server:  sscnet.ssc.gov
#Address:  134.3.1.100
# 
#Name:    lupulus.ssc.gov
#Address:  143.202.80.30
# 
#> bash$ exit
# 
#script done on Thu Nov 21 14:00:36 1991
# 
#It appears that valid, up to date A and PTR records exist on the
#authority servers for my machine. Yet nameservers cannot match
#my address with my hostname. What is wrong?

	The 202.143.in-addr.arpa domain is not registered
	in the root nameservers.

	Contact the NIC, have them add the proper NS rrs.	
# 
#-John
#--
#hoser@lupulus.ssc.gov

	Ehud

--
Ehud Gavron        (EG76)     
gavron@vesta.sunquest.com

-----------[000204][next][prev][last][first]----------------------------------------------------
Date:      22 Nov 91 00:45:08 GMT
From:      ced@bcstec.uucp (Charles Derykus)
To:        comp.protocols.tcp-ip
Subject:   DNS ghost story - still haunted

ORIGINAL POST:
*************

>The problem is that "ghost" (and all its newer addresses) lives on in
>the mind of BIND. Incidentally, we are running 4.8.1 in a multi-platform 
>environment (mostly Sun Sparc 4.1.1). 
>
>"DIG" gives the following:
>
>; <<>> DiG 2.0 <<>> @primary_server sub1.sub0 ns 
>;; ->>HEADER<<- opcode: QUERY , status: NOERROR, id: 10
>;; flags: qr aa rd ra ; Ques: 1, Ans: 7, Auth: 0, Addit: 13
>;; QUESTIONS: 
>;;      sub0.com, type = NS, class = IN
>
>;; ANSWERS:
>sub1.sub0  120     NS      ghost.sub1.sub0  ; 2 min 
>
>;; ADDITIONAL RECORDS:
>ghost.sub1.sub0  542745  A       original_ns_ip  ; 6 days 6 hrs 45 min 45 sec 
>ghost.sub1.sub0  468348  A       non_ns_ip1      ; 5 days 10 hrs 5 min 48 sec 
>ghost.sub1.sub0  550571  A       non_ns_ip2      ; 6 days 8 hrs 56 min 11 sec 
>ghost.sub1.sub0  581952  A       non_ns_ip3      ; 6 days 17 hrs 39 min 1 2 sec 
>ghost.sub1.sub0  604800  A       newest_non_ns_ip ; 1 wk 


YOUR RESPONSE:
*************

>>  Well, I just tried: dig ghost.sub1.sub0
>>and got back NXDOMAIN :-) .
 
>>  Seriously, if you want help with these kinds of problems, tell the
>>truth, the whole truth, and nothing but the truth.  There are plenty of
>>people on the net who would do some DNS lookups to see what is happening.
>>But when you insist on giving phony names instead of the real name, they
>>cannot help you.
 
>>  Most likely what happened is that the root nameservers were never asked
>>to delete the old NS and GLUE records.  Until that problem is corrected the
>>out of date information will keep circulating.

Neil Rickert

MY QUESTION ABOUT YOUR RESPONSE:
*******************************

I'm afraid lying has become second nature. They told me it was the
only way to make things really secure :-).

(Boeing has only 1 Internet windown and it doesn't even use internal
 nameservers so not much is visible.)

I'll check out the old NS and GLUE, but root admin swears they're
clean. In fact, they were the first to complain. 

My question is: could somehow there be a bogus delegation from
from some machine mis-configured as root? Could the imposter
get anyone to believe its little lie?

Charles DeRykus				Internet:   ced@bcstec.boeing.com
Boeing Computer Services		UUCP:	    ...!uunet!bcstec!ced
Bellevue, WA.  M/S 7M-HA		(206) 957-5326  FAX: (206) 957-5050

-----------[000205][next][prev][last][first]----------------------------------------------------
Date:      22 Nov 91 02:29:51 GMT
From:      jeff@onion.rain.com (Jeff Beadles)
To:        comp.protocols.tcp-ip
Subject:   Re: anon ftp setup on Sun 3/60

itexjct!warren%itexjct@uunet.uu.net (Warren Burstein) writes:

|>We have SunOs 4.1.1, I set up anon ftp just like it says in the man
|>page, but when I log in and run dir, this is what I get
 
|>$ ftp there
|>Connected to there.
|>220 there FTP server (SunOS 4.1) ready.
|>Name (there:): anonymous
|>Password (there:anonymous): 
|>331 Guest login ok, send ident as password.
|>230 Guest login ok, access restrictions apply.
|>ftp> dir
|>200 PORT command successful.
|>150 ASCII data connection for /bin/ls (147.161.1.3,3773) (0 bytes).
|>crt0: no /usr/lib/ld.so
|>226 ASCII Transfer complete.
|>25 bytes received in 0.16 seconds (0.15 Kbytes/s)
 
|>/usr/spool/ftp/usr/lib:
|>---x--x--x  1 root        24576 Oct 14  1990 ld.so
|>-r-xr-xr-x  1 root       483328 Oct 15  1990 libc.so.0.15
 
|>I probably have a file mode wrong or something is missing but can't tell
|>what it is.

Yup.  Change ld.so to 644 permissions, and all will be well.

	-Jeff
-- 
Jeff Beadles		jeff@onion.rain.com

-----------[000206][next][prev][last][first]----------------------------------------------------
Date:      22 Nov 91 08:03:54 GMT
From:      tli@lager.cisco.com (Tony Li)
To:        comp.sys.hp,comp.protocols.tcp-ip
Subject:   Re: HP Probe and Routers

In article <r62jji@twilight.wpd.sgi.com> coolidge@speaker.wpd.sgi.com (Don Coolidge) writes:
    
    Probe and Probe Proxy are HP-proprietary products. cisco supports Probe
    VNA requests (roughly analogous to ARP requests), but not Probe name
    requests. 
    In any event, HP's the place to go for info, not cisco.

As of 8.3, cisco can serve as a Probe Proxy server.  And getting Probe info
out of HP has been challenging.

I've mailed more info directly to Bill and will let him summarize.

Tony
Tony "Bob" Li 							  tli@cisco.com
		       The nets are not what they seem.

END OF DOCUMENT