On Jan 30, 2008, at 7:46 PM, Brian E Carpenter wrote:
On 2008-01-31 13:33, Kevin Day wrote:When a 6to4 relay encapsulates v6 traffic and sends it to a 6to4 host over v4, should the source address be 192.88.99.1 or the relay's v4 unicast address?This kind of makes my head hurt because...RFC3056 and RFC3068 are both silent on this decision, as far as I can tell.Naturally, 3056 is silent because the notions of host-based 6to4 and of an anycast address for the relay were both absent at that time. The idea was router-based 6to4 with the relay's IPv4 addressbeing just another router's address (i.e. very definitely not anycast).
Sorry, I didn't mean that to sound like 3056 should have had anything to say about this specific issue, just that it didn't mention v4 source address selection at all. 3068 also didn't mention it, which it maybe should have.
Regardless of which is right, I think the current RFCs are ambiguous onthis point, and that it should probably be clarified - even if the clarification is to say that both are acceptable.I think that's the only viable choice, because there will be deployments where one or the other is better.
Agreed. But if that is the general consensus, I believe it should be documented somehow.
My take on the situation was that the only thing that made sense was to use 192.88.99.1 - like just about everything else on the internet, if packets are sent to that IP, replies should come from that IP. Others believed the exact opposite. If both are right, documenting that somewhere makes the most sense.
Additionally, isanyone strongly for or against adding a PTR record for 192.88.99.1 thatmight help document its use better?Well, yes, but unless it can point to rfc3964.tools.ietf.org I'm not sure it will help people much.
I'm not exactly sure what, but I was thinking something along the lines of "anycast.6to4.relay.see.rfc3068.net" which I think would give people enough to find out more on their own.
-- Kevin