Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 (Tek) 9/28/84 based on 9/17/84; site hammer.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!genrad!decvax!tektronix!orca!hammer!steveg
From: steveg@hammer.UUCP (Steve Glaser)
Newsgroups: net.lan,net.dcom
Subject: Re: ETHERNET on Broadband
Message-ID: <1372@hammer.UUCP>
Date: Wed, 10-Jul-85 02:21:29 EDT
Article-I.D.: hammer.1372
Posted: Wed Jul 10 02:21:29 1985
Date-Received: Fri, 12-Jul-85 03:30:25 EDT
References: <2926@decwrl.UUCP> <449@bu-cs.UUCP> <1263@opus.UUCP> <454@bu-cs.UUCP>
Reply-To: steveg@hammer.UUCP (Steve Glaser)
Organization: Tektronix, Wilsonville OR
Lines: 32
Xref: watmath net.lan:911 net.dcom:1095
Summary: 

>Yes, I would like a black box that just snapped onto an ethernet cable
>and broadband tap and magically forwarded to another, remote box with
>the same connections, but I don't like the idea of 2 or 3 dedicated
>channels to do it. I also wonder exactly what gets forwarded (every
>packet? the DEC box certainly doesn't read IP packets.)  I guess for now
>I am using a couple of Vaxes as this 'black box'.

I think Bridge makes a box that connects 2 ethernets over a point to
point link (V.35 or whatever at any data rate you care about).  It
works directly on the Ethernet layer, making no assumptions about
what's on top of that (IP, DECNET, XNS, etc.).  With a simple RF modem,
you can easily get this "point-to-point" link onto a broadband cable.

It essentially adapts itself to the ethernet addresses of the stuff on
either end and uses that to help it decide what needs forwarding across
the link.  Given that there are only about 100 hosts per Ethernet the
tables aren't that big a deal, especially if you don't need to generate
them.

I suppose this only really works on protocols like TCP where
retransmission times adapt to the network rather than "knowing" what an
Ethernet can do (most TCP implementations seem to do this).  Also, I
don't know what this box does with broadcast packets - I suppose they
get forwarded (otherwise ARP etc. wouldn't work).  I expect that this
solution may break (or at least get bogged down) when folks really
start using multicast packets on Ethernet.  It certainly will have
problems if there is a large speed disparity between 10Mbps and the
point-to-point link.

	Steve Glaser
	Tektronix Inc.
	/* usual disclaimer goes here */