Re: [Idr] draft-ymbk-sidrops-rov-no-rr

[Job Snijders <[e-mail]>]

Hi Robert,

On Mon, Nov 15, 2021 at 07:32:33PM +0100, Robert Raszuk wrote:
> It sure can. RIR can adjust ownership of the prefix, RIR can have bugs, ASN
> ownership can change etc ... It may or may not need to trigger an
> actual BGP update and those two worlds are decoupled.
> 
> IMHO we should just mandate to keep shadow of the Adj-RIB-In pre
> policy or offload the validation to local controller(s) fed by RPKI
> and BMP data.

I think you and the authors are aligned in terms of desired outcome.

When you say "shadow of the Adj-RIB-In pre policy", it makes me wonder
if there is some kind of disconnect in our mutual understanding of the
definition what "Adj-RIB-In" means. As far as I understand the term,
Adj-RIB-In is "pre-policy". https://datatracker.ietf.org/doc/html/rfc4271#section-3.2

I'm not sure I a point to explicitly describe 'offload to controller',
because from a protocol description perspective in this context there is
no difference whether a 'traditional BGP router' or 'thing + controller'
is used.

Kind regards,

Job