On Tue, Nov 01, 2016 at 11:27:20AM +0000, Sriram, Kotikalapudi (Fed) wrote:
> Thanks, Job, for the responses.
> 
> >I'd like to propose a different strategy and describe some background.
> >Both of you mention per-peer flags or new per-route state flags, which
> >in essence are updates to rfc4271, this feels like too heavy of a
> >hammer.
> 
> The flags/markings are local to the router. Does not require any rfc4271 
> updates, IMO.
> E.g. Origin validation (RFC 6811) marking is already implemented in commercial
> routers and did not require rfc4271 update.
> 
> >NEW2:
> >   "Software MUST discard any routes from an EBGP peer, if no import
> >    policy was configured."
> 
> Sounds good to me.

I'll update the doc and incorporate this in -03. Thanks.

> Additionally, should there be an alert generated for the network
> operator.  The discard of routes happening quietly (while operator
> knows nothing about it) is not good.

An alert is overkill imho, can't have a system spew 600,000 alerts
because it rejected a full table feed.

I think this will be very hard to capture in a spec. Some
implementations remain silent (yeah not ideal), other implementations
show a default policy (example: "REJECT") applied on a neighbor in "show
neighbor xxx"-style commands if there is no policy configured.

Kind regards,

Job

_______________________________________________
GROW mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/grow

Reply via email to