On Mon, 29 Jul 2002 [EMAIL PROTECTED] wrote: > >> > I'd like to see an ICMPv6 error message value assigned for the case where > >> > a node would need to send a packet from a specific address, but cannot > >> > because that address is anycast (since the other end has no other good way > >> > of knowing this is the case). One example of this is if a node receives a > >> > TCP SYN on an anycast address, today it apparently has to silently drop > >> > the SYN, and the other end can't tell why. No existing ICMPv6 error > >> > message applies to this case. > >> there is an easier fix for this. un-break sourcing anycast packets from > >> their proper address. > >Exactly, plus e.g. "addr unreach" is good enough if you can't. > > pitfall alert: > > the term "anycast" is used with very vague definition here. > Dave is talking about RFC2460 anycast, Randy is talking about "unicast > address appear in multicast location" anycast (pseudo-anycast in > draft-ietf-ipngwg-ipv6-anycast-analysis-01.txt). > > read draft-ietf-ipngwg-ipv6-anycast-analysis-01.txt for more on this > topic.
Please elaborate. I don't think the draft discusses this at all. With RFC2460 anycast, you can always assume there is another unicast address of at least same scope assigned on the node. As the node cannot use RFC2460 address as a source address, it will use unicast address. Problem solved(?). -- Pekka Savola "Tell me of difficulties surmounted, Netcore Oy not those you stumble over and fall" Systems. Networks. Security. -- Robert Jordan: A Crown of Swords -------------------------------------------------------------------- IETF IPng Working Group Mailing List IPng Home Page: http://playground.sun.com/ipng FTP archive: ftp://playground.sun.com/pub/ipng Direct all administrative requests to [EMAIL PROTECTED] --------------------------------------------------------------------
