Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site rlgvax.UUCP
Path: utzoo!linus!gatech!seismo!rlgvax!dennis
From: dennis@rlgvax.UUCP (Dennis Bednar)
Newsgroups: net.dcom
Subject: Re: Request for Information on Standards Documents/Diagnostic Modems
Message-ID: <842@rlgvax.UUCP>
Date: Thu, 14-Nov-85 19:39:54 EST
Article-I.D.: rlgvax.842
Posted: Thu Nov 14 19:39:54 1985
Date-Received: Fri, 15-Nov-85 20:25:34 EST
References: <547@ttidcb.UUCP>
Organization: CCI Office Systems Group, Reston, VA
Lines: 44
Keywords: IEEE Token Bus Protocol, CSMA/CD, Deterministic, Non-Deterministic

> 
> 
> 
> Fellow D-commers, I humbly request your assistance. I would really appreciate
> info on the following subjects:
> 
> 1. Contact at IEEE for asking questions about, and ordering, standards
> documents (802 committee stuff).
> 

I don't have the name of IEEE contact, but I have some related news
which fellow dcom'ers might find interesting.

IEEE has published the IEEE 802.x (where x = 1 to 5, I think)
standards in a series of books, available from IEEE.

Personally, the one book which I want to read most, is the
Token Bus spec, because of the fact that a distributed algorithm
is needed to pass the token.  I would like to learn more about
the startup algorithm which is responsible for deciding which
node on the LAN gets to hold the token first.

Also, I remember about 2 years ago reading some articles that debated
which was better: CSMA/CD or token-bus?  The battle was between
non-deterministic (csma/cd) and deterministic (token-bus).
The claim was that token-bus is superior to csma/cd when there is
a lot of traffic on your LAN, because the law of probabilities
may be against your LAN-node being able to send your packet out
over the LAN to another LAN-node.

I personally wonder how well token-bus would work in real life,
if there was a heavy load on the LAN, when a new LAN-node suddenly
was turned on, and wanted to have itself admitted to the "logical
ring".  This next idea may not be well-formulated, but is is possible
that the powered on LAN-nodes admission to the "logical ring" will be
non-deterministic? That is, is there a drawback that
you cannot predict how long it will take in the worst case, for
the new LAN-node to get the token in the first place????
I cannot answer this question, because it depends on how the
startup initialization is done in the first place.

-- 
Dennis Bednar	Computer Consoles Inc.	Reston VA	703-648-3300
{decvax,ihnp4,harpo,allegra}!seismo!rlgvax!dennis