Re: Last call for bgp-redistribution

On Thu, 20 Jun 2002, Mark Knopper wrote:
> This is a Last Call for draft-ietf-ptomaine-bgp-redistribution-00.txt to be
> moved to Proposed Standard. The immediate reason is to cause the IANA to
> assign extended community type codes/values as specified in the draft.

Quickie comment: the CIDR/prefix length filter appears to be incompatible
with IPv6, and cannot really be expanded to handle it.


And to a more generic comment..

I have a cornern whether this would really solve the needs which route 
redistribution now has; basically the impression is that this is a way of 
standardizing non-transitive redistribution community values.  Good.

But the question is, are transitive communities (or at least, ones that 
are passed downs at least one more hop from your peer) being used for 
controlling redistribution?

I could imagine that if we would consider a simple model:

 Small-ISP --> Large-ISP --> International-Transit --> Some-ISP -->

(The reality is surely more multilayered.)

Currently the proposed mechanism allows to influence how Large-ISP 
redistributes the routes.

However, I could imagine some, even many (especially if a goal is 
standardize redistribution values) might want to be able to influence the 
redistribution e.g. one or two more hops upstream.

So in conclusion, my worry is whether this addresses significant enough 
part of the redistribution problem.

-- 
Pekka Savola                 "Tell me of difficulties surmounted,
Netcore Oy                   not those you stumble over and fall"
Systems. Networks. Security.  -- Robert Jordan: A Crown of Swords