Re: Concerns about ripv2 "subsumption"
Dennis Ferguson <dennis@ans.net> Fri, 13 November 1992 15:13 UTC
Received: from ietf.nri.reston.va.us by IETF.CNRI.Reston.VA.US id aa05110;
13 Nov 92 10:13 EST
Received: from CNRI.RESTON.VA.US by IETF.CNRI.Reston.VA.US id aa05106;
13 Nov 92 10:13 EST
Received: from atlas.xylogics.com by CNRI.Reston.VA.US id aa10248;
13 Nov 92 10:14 EST
Received: by atlas.xylogics.com id AA29968 (5.65c/UK-2.1-921001);
Fri, 13 Nov 1992 10:14:29 -0500
Received: from interlock.ans.net by atlas.xylogics.com with SMTP
id AA24371 (5.65c/UK-2.1-921001); Fri, 13 Nov 1992 10:14:23 -0500
Received: by interlock.ans.net id AA28369
(InterLock SMTP Gateway 1.1); Fri, 13 Nov 1992 10:12:33 -0500
Received: by interlock.ans.net (Internal Mail Agent-2);
Fri, 13 Nov 1992 10:12:33 -0500
Received: by interlock.ans.net (Internal Mail Agent-1);
Fri, 13 Nov 1992 10:12:33 -0500
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: Dennis Ferguson <dennis@ans.net>
Message-Id: <199211131510.AA06130@foo.ans.net>
To: Gary Scott Malkin <gmalkin@xylogics.com>
Cc: jnc@ginger.lcs.mit.edu, ietf-rip@xylogics.com
Subject: Re: Concerns about ripv2 "subsumption"
In-Reply-To: (Your message of Fri, 13 Nov 92 09:20:43 EST.)
<27110.199211131420@atlas.xylogics.com>
Date: Fri, 13 Nov 92 10:10:25 -0500
Gary, > I take it that the problem is that non-contiguous subnets will cause > things to break, which is why they were illegal in RIP-1. Well, > they're illegal in RIP-2 by inheritance. RIP-2 changes only what > RIP-2 changes, everything else is the same. I can put in some > text explicitly forbidding partitioning when the RFC goes to Draft. No, the problem is the subsumption rules in section 3.4, item 1, don't work. By default you need to advertise routes as you received them, you can't in general combine several routes into a single route with a shorter mask automatically without running the risk of breaking something. I think what needs to be done is to remove section 3.4, item 1, and instead indicate that some routers will require the ability to have subsumption configured for advertisements on a per-interface basis. This is consistent with OSPF's handling of the issue at area boundaries, and allows you to configure the same behaviour as RIP-1 at network boundaries if you want it while allowing you to do other things if you are all RIP-2+OSPF. Dennis Ferguson
- Concerns about ripv2 "subsumption" Dennis Ferguson
- Re: Concerns about ripv2 "subsumption" Noel Chiappa
- Re: Concerns about ripv2 "subsumption" Fred Baker
- Concerns about ripv2 "subsumption" Gary Scott Malkin
- Re: Concerns about ripv2 "subsumption" Dennis Ferguson