Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!cornell!rochester!udel!pcpond!farber
From: farber@pcpond.UUCP (David J. Farber)
Newsgroups: comp.protocols.tcp-ip
Subject: Re: Token Ring Patent
Message-ID: <40@pcpond.UUCP>
Date: 19 May 89 01:30:56 GMT
References: <8905171829.AA13865@monk.proteon.com>
Reply-To: farber@.UUCP (David J. Farber)
Organization: PCPond
Lines: 31


Howard,

I just wanted to elaborate a  bit  on  the  early  days  of  ring
technology.  The Newhall/Farmer Ring developed at Bell Labs was a
central station  ring  where  the  control  station  generated  a
bipolar violation to act as a control token. When we realized (at
Irvine during the DCS Project) that rings were the thing, we  did
not  like either the control station or the hack of using bipolor
violations. We thus developed the notion of the fully distributed
ring control which has been the feature of all modern token rings
as well as the message format that is essentially  that  used  in
current   rings.   I   should  add  that  we  actually  used  the
"destination" address as a process name of processes that  should
receive  the message. That is why we decided to allow the message
to be removed by only the source ring  interface  (in  retrospect
that was a great idea for high speed ).

We did the original token ring for dcs at 2.7 megabits per second
and  the  received  a  contract from Darpa to design a modernized
version -- called the LNI. We did a prototype that almost  worked
(it  was  big!!) and then passed it on to MIT. MIT hired you guys
to continue the development and thus the Proteon token  ring  was
born.  (You  did  a  lot more that just finish the testing -- you
redesigned a lot of it -- wish you  had  left  the  process  name
structure in though).

Big credit is owed to proteon for keeping the faith  about  rings
before IBM "discovered the token ring".

Dave