I know this is a bit late but:
Section 5.1's sythesis of MX records based on the presence of
AAAA records is a bad idea.
If no MX records are found, but an address
RR (i.e., either an IPv4 A RR or an IPv6 AAAA RR, or their
successors) is found, the address RR is treated as if it was
associated with an implicit MX RR, with a preference of 0, pointing
to that host.
Synthesizing a MX record on NODATA to a MX lookup and a
subsequent successful AAAA lookup is bad engineering decision.
It will work reasonably well for IPv4 only + dual stack
envirionment. It will not work well for IPv4 only + dual
stack + IPv6 only envirionment.
The reason it is a bad engineering decision is that:
* the IPv4 only world needs a MX RRSet to find a dual stack
MTA to relay into the IPv6 network.
* the IPv6 world has a raft of solutions which will allow it
to initiate a connection to a IPv4 only MTA without having
to find a dual stack MX for the target mail domain.
* it changes the definition of what it means to exist in
the mail domain and you will have different MTA/MSA making
different existance decisions. Some will say that AAAA +
no MX exist but others will say that the site does not
exist.
e.g.
a new (IPv6 aware) MSA which is configured to relay through
a old (non-AAAA aware) MTA on its outward bound path.
Do you really thing we should be trying to force a upgrade
of all MTA's on the planet to support MX synthesis from
AAAA when there is no engineering need to to this?
MX from A was a transition strategy. IPv6 only sites have a
transition strategy that doesn't require synthesis. It is
advertise a dual stack MX. At some point in the future sites
will stop having a dual stack MX, the same way they stopped
adding A records for mail only domains back in the 90's.
Mark
--
Mark Andrews, ISC
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742 INTERNET: Mark_Andrews@xxxxxxx
_______________________________________________
Ietf@xxxxxxxx
https://www1.ietf.org/mailman/listinfo/ietf