Newsgroups: comp.sys.sun.hardware,comp.dcom.lans.ethernet
From: vjs@calcite.rhyolite.com (Vernon Schryver)
Subject: Re: Ethernet chip bugs (slow network performance)
Keywords: Ethernet,bugs,slow
References:  <1994Feb8.175343.2943@newstand.syr.edu> <2jc6iv$g99@redwood.cs.scarolina.edu>
Organization: Rhyolite Software
Date: Thu, 10 Feb 1994 16:58:49 GMT
Message-ID: 
Lines: 32
Xref: geraldo.cc.utexas.edu comp.sys.sun.hardware:10222 comp.dcom.lans.ethernet:5736

In article <2jc6iv$g99@redwood.cs.scarolina.edu> Tim White  writes:
>ppmorris@rodan.syr.edu (Peter P Morrissey) writes:
>
>>We have been grappling with the issue of just how many collisions are considered
>>normal. I have heard estimates of from 1 to as high as 10%.  ...

>   I've read many times that if collisions get up above 3 or 4 percent
>  performance will fall off rapidly. ....

You have the decimal point wrong.
3% or 4% collisions means the Ethernet is idle.
10% is means the Ethernet is lightly loaded.
30% or 40% is something to worry about.

Do the arithmetic yourself to see that a 4% collision rate will have
an insignificant effect on throughput either between any pair of
stations or for the entire network.

Do not believe the switching hub salescritters who, for the sake of their
commission checks, would have you believe that collisions are bad things.
A moderate collision rate is just fine.

>   ...        I know a site with thick wire with 90 degree angles in the 
>  cable. Cable runners were low bid on a state contract.  Unbelievable
>  collision rates. 60-70 percent some times. ....

A sick cable plant and network load as indicated by collision rates
are two entirely different things, although each can affect the
effects of the other.


Vernon Schryver    vjs@rhyolite.com