At Wed, 6 Apr 2005 13:22:58 -0400, Brian Haberman wrote: > > There is a proposal in draft-jabley-v6-anycast-clarify-00.txt to change > the following text in the addressing architecture: > > o An anycast address must not be used as the source address of an > IPv6 packet. > > o An anycast address must not be assigned to an IPv6 host, that > is, it may be assigned to an IPv6 router only. > > to > > o An anycast address MAY be used as the source address of an IPv6 > packet. > o An anycast address MAY be assigned to an IPv6 host. > > This change will allow users to operate IPv6 anycast services in the > same manner in which they do today with IPv4 anycast.
I support making this change. As far as I can tell, the current restrictions on IPv6 anycast, while written with the best of intentions, serve no purpose other than to forbid every real current use of anycast addresses, and are honored entirely in the breach. Removing these restrictions is just recognizing that we have learned something from experience since the restrictions were written. Use of anycast with TCP (or any protocol in which state crosses packet boundaries) is a very tricky subject in which black and white answers are suspect and everything is some shade of grey. However, I believe that the GROW WG is already addressing this, in the correct way, by attempting to detail the considerations for use of such protocols with anycast. As far as I can tell, the issues are identical for IPv4 and IPv6, and informed discussion of them depends heavily on operational experience with routing protocols, so I think that the GROW WG is the right place for that work. -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
