Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site wdl1.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!bellcore!decvax!decwrl!sun!wdl1!jbn
From: jbn@wdl1.UUCP
Newsgroups: net.lan
Subject: Token rings, error reporting in
Message-ID: <952@wdl1.UUCP>
Date: Thu, 23-Jan-86 01:17:20 EST
Article-I.D.: wdl1.952
Posted: Thu Jan 23 01:17:20 1986
Date-Received: Fri, 24-Jan-86 21:56:17 EST
Sender: notes@wdl1.UUCP
Organization: Ford Aerospace, Western Development Laboratories
Lines: 19
Nf-ID: #N:wdl1:87900001:000:1026
Nf-From: wdl1!jbn    Jan 22 14:14:00 1986


     One annoying problem with Ethernet-type systems is that you send
blind; there's no link-level indication that nobody is listening to you.
This causes some problems in adaptive routing that I won't go into here.
     This ought not to be a problem with token rings, since the receiver
usually takes the message off the ring and replaces it with the token.
(You could cheap out and let the sender take it off the ring, but it
halves the ring bandwidth if you do, of course.)
If the sender sees a message with its own source address go by, it has
to remove it to prevent it going around forever, and this indicates a
failure to deliver the message.  (You could really cheap out and let the
ring time out and regenerate the token whenever this happens, but then
sending to a dead node will tie down the ring for some time.)
     Which token rings actually provide indication of non-delivery to the 
sender?  I've been told that the Cambridge Ring does and Proteon does not.
Who has hard information on this?

					John Nagle