Re: [Idr] WGLC on draft-ietf-idr-as-private-reservation-00

[Jared Mauch <[e-mail]>]

On Dec 5, 2012, at 12:00 PM, Jay Borkenhagen <[e-mail]> wrote:

> Yes, this sort of problem with incremental deployment happens all the
> time.  But we should limit the occasions where it happens to those
> where the win is big enough.  In this case I don't think there is
> enough of a win.
> 
> 
> BTW, if this draft does proceed, it will provide me with more reasons
> to encourage network designers I talk to to avoid using private ASNs
> altogether.  Them: "But private ASNs must be great.  Look, they just
> allocated a bunch more."  Me: "All the more reason never to use them."
> :)

This represents my concerns as well with incremental deployment.  Most BGP ASes aren't as smart as the people on this list, and I am concerned for them as they upgrade and the vendors update the code and implement this change.  (I say this, because I feel like i'm hearing consensus on supporting this).

I think perhaps there should be some recommendations about implementing this, perhaps:

routes should not go eBGP without explicit configuration

Then again, I wish that vendors wouldn't send ANY routes without explicit policy configuration.. would solve the millions of transient routing leaks seen all day long.  (see work in grow-wg).

- Jared