On Thu, 2009-12-10 at 14:43 -0500, Dale Worley wrote: > On Thu, 2009-12-10 at 13:49 -0500, Scott Lawrence wrote: > > I did the selection/sort-weighting on names rather than addresses > > deliberately... it seemed to me to be more in line with the conceptual > > structure of the SRV mechanism. An SRV record points to a name, not to > > an IP address. That name may have many addresses, and an address may > > have many names. Using the name to bias the weighting seemed less > > likely to cause problems than trying to do something at the address > > level (it was also _much_ easier to implement). > > The essential difference is that this mechanism acts only when the > presented domain name is an SRV name, one of whose targets is the same > as the host's name. (As opposed to when the presented domain name has > several A records with different addresses, where one address is the > host's address.) The latter case is unlikely in systems that are > configured according to our recommendations, but we should make sure > that we make this choice consciously.
Correct... that and many of the almost infinite number of ways to incorrectly configure things won't be affected, but it also shouldn't be broken (any more badly than it is anyway). Multiple A records should work now if all the alternatives are equivalent, but I don't think that we do anything special for that case. Since that's not how we recommend setting things up, I'm comfortable with explicitly choosing to not optimize for it. _______________________________________________ sipx-dev mailing list [email protected] List Archive: http://list.sipfoundry.org/archive/sipx-dev Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-dev sipXecs IP PBX -- http://www.sipfoundry.org/
