On Mon, Mar 31, 2008 at 02:21:06PM +0200, Arnt Gulbrandsen wrote: > >>>This amounts to retro-active re-design on-the-fly. > >> > >>We can't escape redesigning this. IPv6 forces us to. > >I still don't understand why IPv6 forces a redesign. I don't consider > >an MX record in the DNS system strongly tied to a transport. > > Indeed. But this isn't about the MX, it's about what to do when the > domain in question has no MX.
Correct. > >... > > > >>I like falling back to A only. That imposes no new requirements on > >>existing code or IPv4-only hosts, and it doesn't make unrealistic > >>presumptions about what software runs on future IPv6-only devices. Notice: lookup A only > >In spirit I agree. > > > >If something is trying to deliver e-mail, they have a choice between > >looking up an A or AAAA record or looking up MX records. Notice: lookup A and/or AAAA That's no agreement at all. It's what this discussion is about. > RFC 2821 page 60 says quite clearly that if MX records exist, they must > be used, and that an A record is used only if there aren't any MXes. > 974/1123 say much the same (some details are different). That's the current status. What people are trying to do, is to do the same with AAAA records. Basically I see these points of view right now: a) leave as is. Only lookup A record where no MX record exists b) leave as is. Like looking up A record, one should lookup AAAA record if no MX record exists (details A or AAAA vs. A and AAAA can be discussed separately once consensus on (b) is achieved) c) change the design Notice how (a) and (b) disagree and both say 'no change'. IMHO (c) shouldn't happen. For what it's worth: I'm for (a).
