Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!cmcl2!rutgers!princeton!udel!garrett
From: garrett@udel.EDU (Joel Garrett)
Newsgroups: comp.protocols.appletalk
Subject: Re: appletalk error rates
Message-ID: <638@louie.udel.EDU>
Date: Tue, 27-Oct-87 23:56:41 EST
Article-I.D.: louie.638
Posted: Tue Oct 27 23:56:41 1987
Date-Received: Sat, 31-Oct-87 03:30:47 EST
References: <9083*sample@cs.ubc.cdn>
Reply-To: garrett@udel.EDU (Joel Garrett)
Organization: University of Delaware
Lines: 23
Summary: Zilog SCC in Mac isn't buffered deeply...

In article <9083*sample@cs.ubc.cdn> sample@ubc.CSNET (Rick Sample) writes:
>Running Peek gives three types of errors: overruns, CRC errors,
>and time outs.  Are collisions included in these errors?  Is an
>error rate of 5% much more than normal?  Seems rather high to me.
>What might cause such a high rate?
>
I'm not sure if this has a lot to do with it, but I think I remember reading
somewhere (maybe in the BYTE article on Appletalk on Ethernet a while back)
that the Appletalk interfaces on the Mac aren't buffered very deeply (ie
either one packet or none at all, not sure which) so depending on what is
going on on a receiving mac, it would be very easy to overrun its buffer (or
the lack thereof) and also have timeouts.  CRC errors are another story,
no idea about that.  Also, I've never measured the error rate on a
"non-flaky" network (or for that matter on a "flaky one" ;-) so I wouldn't
know what a "typical" error rate in either of these situations would be,
though I'd think that 5% might not be out of line, especially if it is
a busy net.

					Joel Garrett
					Research Associate
					University of Delaware
					Center for Composite Materials
					arpa: garrett@udel-ccm.arpa