Re: Unusual use of IP Broadcast Address

Steve Deering <deering@parc.xerox.com> Thu, 13 January 1994 19:07 UTC

Received: from ietf.nri.reston.va.us by IETF.CNRI.Reston.VA.US id aa09532; 13 Jan 94 14:07 EST
Received: from CNRI.RESTON.VA.US by IETF.CNRI.Reston.VA.US id aa09528; 13 Jan 94 14:07 EST
Received: from venera.isi.edu by CNRI.Reston.VA.US id aa14653; 13 Jan 94 14:07 EST
Received: from alpha.Xerox.COM by venera.isi.edu (5.65c/5.61+local-14) id <AA17582>; Thu, 13 Jan 1994 10:57:21 -0800
Received: from skylark.parc.xerox.com ([13.2.116.7]) by alpha.xerox.com with SMTP id <14512(4)>; Thu, 13 Jan 1994 10:57:00 PST
Received: from localhost by skylark.parc.xerox.com with SMTP id <12171>; Thu, 13 Jan 1994 10:56:53 -0800
To: Bob Braden <braden@isi.edu>
Cc: ietf-hosts@isi.edu, etkind@stowe.mitre.org
Cc: deering@parc.xerox.com
Subject: Re: Unusual use of IP Broadcast Address
In-Reply-To: braden's message of Thu, 13 Jan 94 10:24:41 -0800. <199401131824.AA07354@zephyr.isi.edu>
Date: Thu, 13 Jan 1994 10:56:51 -0800
X-Orig-Sender: Steve Deering <deering@parc.xerox.com>
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: Steve Deering <deering@parc.xerox.com>
Message-Id: <94Jan13.105653pst.12171@skylark.parc.xerox.com>

> I don't think I fully comprehend the situation, but it seems to me
> that if you want to send to a host whose address you don't know
> but which is known to be on subnet 129.83.14, you should use the
> directed broadcast address 129.83.14.255.  Using 255.255.255.255
> does not seem correct.

Bob,

Wendy said that her PCs don't necessarily know that they are on subnet
129.83.14, so they wouldn't be able to recognize 129.83.14.255 as a
broadcast address they should accept.

Upon re-reading her message, I realize that I may also have misunderstood
what she was proposing when I wrote my response.  The references to a
modified version of routed seem to imply that the "broadcast" packets will
be originated from machines other than the Sun that has the serial links to
the PCs.  If that's true, what she is proposing is significantly at odds
with the IP architecture (treating an all-ones broadcast as a directed
broadcast to a remote subnet known as a special case by the routers).  IP
multicast would satisfy her requirement without bending the architecture.

Steve