On Wed, Oct 11, 2017 at 11:41:11AM +0000, Wolfgang Tremmel wrote: > > On 10. Oct 2017, at 05:46, Job Snijders <[email protected]> wrote: > > > > Have you considered just updating RFC 7947 to resolve the described > > ambiguity by stating that a route server SHOULD pass the NO_EXPORT > > community unaltered, rather than interpret it or block it? > > From my reading of RFC1997 and RFC7947 the ambiguity is not really one: > > RFC1997 states that any community-aware BGP speaker MUT NOT advertise > prefixes received with NO_EXPORT > --> a route server is a BGP speaker > --> it is community aware > > RFC7947 uses wording SHOULD NOT and MAY which IMHO are weaker.
OK, but you are not explaining to me why the current set of RFC's can't be updated to encourage different NO_EXPORT behaviour on route servers (compared to the rest of the BGP speakers), but instead a new community is required. > > Why is there no consensus amongst route server operators on what the > > correct behavior is? Can you provide a citation? > > all that aside, DE-CIX already has all that functionality using our > own communities (we actually have two, one for adding NO_EXPORT and > one for adding NO_ADVERTISE, and also allow selective adding using > Large Communities): > https://www.de-cix.net/en/locations/united-states/dallas/routeserver-guide This doesn't tell me why there is no consensus :) Kind regards, Job _______________________________________________ GROW mailing list [email protected] https://www.ietf.org/mailman/listinfo/grow
