home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Columbia Kermit
/
kermit.zip
/
newsgroups
/
misc.19970626-19970929
/
000170_news@newsmaster….columbia.edu _Fri Aug 15 17:42:50 1997.msg
< prev
next >
Wrap
Internet Message Format
|
2020-01-01
|
3KB
Return-Path: <news@newsmaster.cc.columbia.edu>
Received: from newsmaster.cc.columbia.edu (newsmaster.cc.columbia.edu [128.59.35.30])
by watsun.cc.columbia.edu (8.8.5/8.8.5) with ESMTP id RAA20873
for <kermit.misc@watsun.cc.columbia.edu>; Fri, 15 Aug 1997 17:42:49 -0400 (EDT)
Received: (from news@localhost)
by newsmaster.cc.columbia.edu (8.8.5/8.8.5) id RAA23233
for kermit.misc@watsun; Fri, 15 Aug 1997 17:42:49 -0400 (EDT)
Path: news.columbia.edu!panix!howland.erols.net!cs.utexas.edu!news.cs.utah.edu!cc.usu.edu!jrd
From: jrd@cc.usu.edu (Joe Doupnik)
Newsgroups: comp.protocols.kermit.misc
Subject: Re: Kermit cyclic error problem
Message-ID: <hplveE5fxIya@cc.usu.edu>
Date: 14 Aug 97 14:06:13 MDT
References: <u3eodau35.fsf@donal.kintailrd>
Organization: Utah State University
Lines: 52
Xref: news.columbia.edu comp.protocols.kermit.misc:7475
In article <u3eodau35.fsf@donal.kintailrd>, Bruce Cook <BC3-AU@bigfoot.com> writes:
> I have a link between a Winnt4.0SP5 machine with Kermit-95v1.1.12 16550A
> and a DOS machine with MSkermit-3.15 16450.
>
> The link is running a 115200, 32 windows, 360 byte packets, block=3.
32 * 360 = 11.5KB of packet buffer space.
Try 4 window slots; more is not helping at all because the
line delay is still short. Most importantly, ensure hardware flow
control is active across the entire link, because I strongly suspect
the modem link plus serial port buffering within NT is dropping bytes
from overflow.
For others reading along, sliding windows says how many packets
are allowed to be transmitted before an ACK for the oldest arrives. The
time for that ACK to arrive is primarily transmission path delay plus
some delay in the receiving host. It would be an unusual path needing
or even using as many as 8 packet slots, unless those packets were
tiny (a bad idea). Going across the country on a decent Internet link
(25KByte/sec rate, not using modems) uses a max of 6 packet slots in my
experiments, and the faster the link the more packets go out to soak up
the delay time (or, equivalently, to keep the transmit wire full).
Your modem link is NOT running at 115Kbps. The connection between
your modem and the computer might, but the telephone company wiring makes
the data flow at very much slower rates. 33Kbits/sec (say 4KBytes/sec)
is doing well for modems. And there resides the imperative to provide
hardware flow control, to avoid overrunning the slow part of the system.
Joe D.
>
> When receiveing, I get an error every n packets where n is the number
> of windows. (I've adjusted windows from 4 to 32). This is giving
> me an obvious throughput problem at 4 windows (25% error rate).
>
> The packet size is so low, because of the line reliability, which
> realy suprises me. Speed doesn't realy make much difference.
>
> The msk315 client is showing n of 32 windows, and every time n reaches
> 32, a retry happens, and n is reset back to 1. This cycle happens
> no matter how many windows I have set up, obviously 32 windows gives
> me a lower error percentage.
>
> Anyone got any ideas as to what's going on ?
>
>
> --
> ...BRU
>
> Bruce Cook, Synonet Corp.
> E-Mail: bcook@wantree.com.au
> Phone: +061 15 999 330
>
>