Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!samsung!zaphod.mps.ohio-state.edu!usc!apple!sun-barr!newstop!texsun!smunews!csvax.seas.smu.edu!merlin
From: merlin@csvax.seas.smu.edu (David Hayes)
Newsgroups: comp.protocols.tcp-ip
Subject: Re: Broadcasting Adreess and Subnets
Message-ID: <16240@smunews.UUCP>
Date: 11 Apr 90 21:55:17 GMT
References: <9003270339.AA02962@WLV.IMSD.CONTEL.COM> <9003262150.AA14999@asylum.sf.ca.us> <362@jove.dec.com>
Sender: root@smunews.UUCP
Reply-To: merlin@csvax.seas.smu.edu (David Hayes)
Organization: Southern Methodist University, CSE Dept. Dallas, TX
Lines: 26

In article <362@jove.dec.com> mogul@decwrl.dec.com (Jeffrey Mogul)
writes:
>Since I wrote RFC922, I've had a change of heart; I agree that multi-
>subnet broadcasts are a bad idea.  I've heard from some people that
>RFC1009 can be read to require their support, and also to require that
>if supported that one can configure them to be blocked; my reading of
>RFC1009 is that gateways are not required to support them.  At any rate,
>this is clearly an issue for the Router Requirements Working Group, and
>I hope that they rectify my error.

I'm working on a printer accounting program that runs throughout
our (subnetted) campus network.  It uses a client-server
architecture.  There are multiple, redundant servers.  I wanted
our system to be able to find the servers by broadcast, but our
cisco router doesn't forward them.  

That's at least one application for which multi-subnet broadcast
was the correct solution.  I don't think Jeff made a mistake
originally.  While any broadcast should be used sparingly, totally
prohibiting multi-subnet broadcasts needlessly deprives us of a
clean solution to my class of problem.

David Hayes	School of Engineering	Southern Methodist University
merlin@smu.edu	uunet!smu!merlin
"Here's a test to see if your job here on Earth is finished:  If you're
still here, it isn't."  -- Richard Bach, _Illusions_