home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Columbia Kermit
/
kermit.zip
/
newsgroups
/
misc.19950929-19951130
/
000254_news@columbia.edu_Mon Oct 30 06:14:20 1995.msg
< prev
next >
Wrap
Internet Message Format
|
2020-01-01
|
4KB
Received: from apakabar.cc.columbia.edu by watsun.cc.columbia.edu with SMTP id AA22952
(5.65c+CU/IDA-1.4.4/HLK for <kermit.misc@watsun.cc.columbia.edu>); Tue, 31 Oct 1995 06:05:34 -0500
Received: by apakabar.cc.columbia.edu id AA22770
(5.65c+CU/IDA-1.4.4/HLK for kermit.misc@watsun); Tue, 31 Oct 1995 06:05:33 -0500
Path: news.columbia.edu!panix!ddsw1!news.mcs.net!van-bc!io.org!chi-news.cic.net!newsfeed.internetmci.com!news.mid.net!news.creighton.edu!bluejay.creighton.edu!honge
From: Psychos 'R Us <honge@creighton.edu>
Newsgroups: comp.protocols.kermit.misc
Subject: Re: Maximum transfer speed
Date: Mon, 30 Oct 1995 00:14:20 -0600
Organization: Creighton University, Omaha Nebraska USA
Lines: 64
Message-Id: <Pine.HPP.3.91.951030000148.24557B-100000@bluejay.creighton.edu>
Nntp-Posting-Host: bluejay.creighton.edu
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
In-Reply-To: <1995Oct29.174729.65311@cc.usu.edu>
Apparently-To: kermit.misc@watsun.cc.columbia.edu
: > On the HP-UX end, the only thing I do is "kermit -i -x." On the OS/2
: > end, I have set packet size to 2048, window slots to 24, and
: > everything else is as-is.
: >
: > Is there any other things I can/should do? I tried to use MS-Kermit
: > (MS-DOS version). On that one, I couldn't even raise the packet length
: > beyond 371! Oh well...
: ------------
: Your last paragraph contains a useful piece of information. A
: packet length of 371 bytes is not the kind of number we normally choose
: for packets, but it is the kind of number which arises when a packet is
: dinged and the sending side shrinks them upon retries. That's the "rubber
: packet" heuristic.
: In most cases of this kind our first response is "flow control,
: flow control!" For best flow control use hardware RTS/CTS between your
: machine and the modem, expect the modems to do their modem to modem
: flow control under V.everything, and cross your fingers that your remote
: host does snappy flow control with its modem too. Flow control must be
: carried through from end to end, else traffic jams yield lost bytes.
Per someone else's suggestion, I typed these on the HP-UX kermit before
giving it the command "server" and begin downloading:
set file type binary
set warning on
set buffer 65536 65536
set window 24
set send packet 2048
set block-check 2
set flow none
set flow rts
With OS/2 c-kermit, the transfer rate is around 1350cps. With MS-kermit,
the transfer rate is around 880cps. The effeciency is around 18%. There
are almost no retries. FYI, Y-modem G under same situation will do around
1681cps with Hayes Smartcom (MS-DOS app) or halite that came with OS/2.
I have set both OS/2 ck and MS-DOS kermit to rts/cts.
: The second response is, maybe there is a terminal server or other
: comms box between you and the host and it is unable to handle fast traffic
: without dropping bytes. Terminal servers are often quite wierd boxes.
Yes, I did notice these are quite a weird beast. I can turn off the flow
control but then everything will go to hell (lotza retries under kermit
or z-modem -- y-modem g won't even work).
: The third and usually last blind suggestion is ensure your OS/2
: end has a decent serial port, typically a 16550A UART, so it does not
: drop bytes under load.
The machine has Hayes ESP, which is kinda like 16550A with extended
buffer size. The modem is Hayes Accura 14.4, which I think is also a
decent modem...
------------- clip here with virtual scissors --------------
************************************************************
Looking for roadkills... drop it by honge@creighton.edu...
e-mails are welcome anytime -- but mails are not.
Keyboard stuck failure. Press F1 to continue.
Q: How many Mac users does it take to change a lightbulb?
A: None -- there is no icon for that.
************************************************************