home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Columbia Kermit
/
kermit.zip
/
newsgroups
/
misc.19950929-19951130
/
000197_news@columbia.edu_Wed Oct 25 22:12:59 1995.msg
< prev
next >
Wrap
Internet Message Format
|
2020-01-01
|
3KB
Received: from apakabar.cc.columbia.edu by watsun.cc.columbia.edu with SMTP id AA24360
(5.65c+CU/IDA-1.4.4/HLK for <kermit.misc@watsun.cc.columbia.edu>); Wed, 25 Oct 1995 18:13:08 -0400
Received: by apakabar.cc.columbia.edu id AA18757
(5.65c+CU/IDA-1.4.4/HLK for kermit.misc@watsun); Wed, 25 Oct 1995 18:13:07 -0400
Path: news.columbia.edu!watsun.cc.columbia.edu!fdc
From: fdc@watsun.cc.columbia.edu (Frank da Cruz)
Newsgroups: comp.protocols.kermit.misc
Subject: Re: Kermit vs. FTP speed
Date: 25 Oct 1995 22:12:59 GMT
Organization: Columbia University
Lines: 44
Message-Id: <46mcpb$i9l@apakabar.cc.columbia.edu>
References: <46jq0k$gol@galileo.csun.edu>
Nntp-Posting-Host: watsun.cc.columbia.edu
Apparently-To: kermit.misc@watsun.cc.columbia.edu
In article <46jq0k$gol@galileo.csun.edu>,
Stephen Walton <swalton@galileo.csun.edu> wrote:
>I'm almost embarrassed to be posting this, but here goes: I'm trying
>to transfer binary files from a 386 PC with Optical WORM disk to an HP
>series 700 machine. The PC is running MS-Kermit 3.14 and the HP is
>running C-Kermit 5A(190). The WORM disk is fairly slow, so I don't
>expect miracles, but what I'm actually seeing is transfer speeds with
>Kermit which top out at 12 Kb/s, as compared to 35 Kb/s using NCSA
>FTP. SET WIN 4, SET BLOCK 3, SET FIL TYP BIN, SET REC PACK 2000, SET
>FLOW NONE, SET CONT UNP ALL, SET CONT PRE 0 1 11 13 17 19 are done at
>both ends of the Kermit transfer.
>
There are several potential bottlenecks. First, of course, is the
hardware -- the PC and the WORM disk -- which explains why FTP is so slow.
Second would be the Kermit protocol settings. What happens if you crank
up the window size and packet length? Does it make a difference?
Third is flow control. Tell both Kermit programs to "set flow none".
Let TCP and IP take care of it.
Fourth is the TELNET overhead. When you TELNET from MS-DOS Kermit 3.14
into the HP, you are going through the HP's TELNET server, not to mention
pty drivers and who knows what else. It might be an interesting experiment
to eliminate the TELNET overhead. You can do this as follows:
1. Start MS-DOS Kermit first. Tell it to "set tcp host *" and then
"server".
2. Make a connection from C-Kermit on the HP to the PC: "set host xxx"
where xxx is the IP name or address of the PC.
Then initiate all the file transfers from the HP end. If you use big window
size and packet length and the maximum possible control-character unprefixing:
SET CONT PRE 0 1 13 255
(note the elimination of 17 and 19 because now you aren't doing Xon/Xoff
any more.)
I'm sure all of our readers would be interested in further reports on how
this goes.
- Frank