Hi Maria, There is one more change your draft is bringing in ... could be not that insignificant.
Today with dual v4/v6 IXP fabric each member (if peering to RS) establishes two eBGP sessions IPv4 and IPv6. What you are proposing is transition to a single session - unless you run v4 & v6 over separate v6 sessions. That would be clearly much more fragile as any session reset will jeopardise the fun for both v4 and v6 reachable destinations. I am not sure if this would be a good thing. /* Maybe one day we will depart from session based BGP state bindings but this moment does not seem to be happening in the near time. */ Cheers, Robert On Thu, Jul 23, 2026 at 10:21 AM Maria Matejka <[email protected]> wrote: > On Thu, Jul 23, 2026 at 09:35:20AM +0200, Robert Raszuk wrote: > > That has been long rejected, see > https://www.ripe.net/community/policies/proposals/2011-05/ > > It is not possible to use unannounced or RFC1918 address space for IXP > peering lans, as this breaks URPF and ICMP communication, and breaks the > opportunity for debugging of connections over the Exchange for connected > and non-connected networks due to the IX path missing from traceroutes, and > peering routers not getting the ability to set reverse DNS. > > Well in respect to your proposal I bet there would be folks asking to use > link-local v6 addresses as next hops: > > https://datatracker.ietf.org/doc/draft-ietf-idr-linklocal-capability/ > > And how would that help to accomplish the above quoted reasons ? > > That is outside the scope of this document which clearly expects that the > network is addressed by IPv6 GUA. > > Also if I have v4 only clients (yes large enterprises still use v4 and not > even planning transition to v6) how would that debugging work over v6 only > IXP fabric ? > > What about this draft which is in the process of finalization? > > https://datatracker.ietf.org/doc/draft-ietf-intarea-extended-icmp-nodeid/ > > Also, literally this, i.e. large enterprises never transitioning to v6, is > the major reason for specifying a transitional mechanism which would allow > IXPs to get rid of IPv4 before all their customers do. > > – > Maria Matejka (she/her) | BIRD Team Leader | CZ.NIC, z.s.p.o. >
_______________________________________________ GROW mailing list -- [email protected] To unsubscribe send an email to [email protected]
