Path: utzoo!attcan!uunet!lll-winken!ncis.llnl.gov!ncis!helios.ee.lbl.gov!pasteur!ucbvax!inmos.co.UK!rob
From: rob@inmos.co.UK (Robin Pickering)
Newsgroups: comp.protocols.tcp-ip
Subject: Ethernet throughput
Message-ID: <5083.8901161055@brwc.inmos.co.uk>
Date: 16 Jan 89 10:55:44 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 21


I realise that this is probably an often asked question, but I am trying to
build up some relative performance figures for LAN hardware/software
configurations.

Has anyone got a sensible (preferably experimentaly derived) figure
for performance of TCP/IP over Ethernet.

This is beginning to sound like a 'how long is a piece of string question',
so lets define it a bit better and say:
    Single ethernet segment, no bridges or routers
    10 Workstations running TCP/IP,UDP and NFS (Sun 3's say)
    Thick wire ethernet (if that makes a difference)

If anyone has any figures for another environment, I am just as interested
but please specify the environment.

Rob Pickering
JANET(UK): ROB@UK.CO.INMOS
ARPA:      rob@inmos.co.uk ( or rob%inmos.co.uk@nss.cs.ucl.ac.uk)
UUCP:      rob@inmos.co.uk (...uunet!mcvax!ukc!inmos!rob)