Re: additional requirements: capneg and policy discovery

[Joe Abley <jabley@nlri.org>]

On Mon, Nov 12, 2001 at 09:08:39AM -0800, Michel Py wrote:
> > - - policy - Multihoming solutions should provide a means for peers
> > to discover the routing policies applied to their routes by peer
> networks.
> 
> I missed the email thread on that one; can we have it again? I am not
> sure everyone would agree to do that.

It was a private message, which I am copying below, since it contains
some elaboration on that point (I should have included it before).

----- Forwarded message from Tony_Hiller <thiller@enteract.com> -----

Delivered-To: jabley@automagic.org
Date: Tue, 26 Jun 2001 17:49:41 -0500
From: Tony_Hiller <thiller@enteract.com>
To: jabley@automagic.org
Subject: suggestions for multi6 draft
User-Agent: Mutt/1.2.5i

Hi jabley,

I'd like to make some suggestions for your multihoming draft. They're probably pretty specific, and may be beyond the scope of your doc, but here goes:

- - capabilities negotiation - Multihoming solutions must provide a means for peers to communicate capabilities to each other, in order to support new capabilities that may or may not be included in all implementations. 

- - policy - Multihoming solutions should provide a means for peers to discover the routing policies applied to their routes by peer networks.

Uh, I don't like the way that sounds. Suffice it to say, I want some way to tell what filters my peers have in place at any given time. I realize this is asking for denial of service attacks, but it seems that such things could be minimized through proper authentication of peers. There are similiar drafts in idr right now - but I have no idea how close they are to running code. 

I'm not sure if these are outside your scope, but I'd see both (especially the first) as valuable capabilities. Just my thoughts...

hiller 


----- End forwarded message -----