Certainly in principle it can, though that is not in the current proposal. The basic idea is to suppress installing routes into the FIB when there is a "virtual aggregate" that you can tunnel to instead. I remember we discussed this in San Jose NANOG, but I forget the details. Can you remind me? PF
-----Original Message----- From: Alain Durand [mailto:alain_durand@cable.comcast.com] Sent: Sunday, July 20, 2008 9:13 AM To: Paul Francis; nanog@nanog.org Subject: Re: virtual aggregation in IETF
Paul,
Can this proposal be applied to IS-IS (or other IGP) as well as BGP?
- Alain.
On 7/20/08 8:46 AM, "Paul Francis" <francis@cs.cornell.edu> wrote:
Gang,
I have submitted an internet-draft to the IDR group on virtual aggregation (VA) (http://www.ietf.org/internet-drafts/draft-francis-idr-intra-va- 00.txt). This draft suggests a few changes to routers that allow operators to control the size of their FIBs, shrinking them by 5x or 10x quite easily. This would extend the lifetime of routers that are constrained by FIB size.
There has been a lively discussion of this on the IDR mailing list, including a suggestion that FIB reduction is more important for lower tier ISPs (tier 2, tier 3...) than for tier 1 ISPs. Unfortunately I don't think that people from smaller ISPs pay much attention to the IDR mailing list, so they are not being represented in this discussion.
So I'm looking for input from folks who think that FIB reduction helps them, so as to better understand their requirements.
Any help is much appreciated.
Thanks,
PF