At 13:46 +0200 on 04/04/2008, Michael Storz wrote about Re: current
usage of AAAA implicit MX?:
On Thu, 3 Apr 2008, Tony Hansen wrote:
I'm looking for help answering these questions:
* How are existing IPv6-enabled mail sites actually dealing with
missing MX records?
* Are the IPv6-enabled email sites mostly following the
recommendations found in RFC 3974?
* How widespread is the use of implicit MX with AAAA records? That
is, how prevalent are email sites that do not have MX records
but do have AAAA records?
* How big is the existing IPv6-enabled email world?
* What problems, if any, has been experienced by following RFC
3974?
From my standpoint of view, the existing IPv6-enabled email world is
really small.
At the moment we are in the transition to dual-stack MTAs. Therefore I
have only data about some of our domains which are reachable by IPv6 from
the Internet. We do not send emails via IPv6 to the Internet.
- number of MTAs from which we accepted emails over IPv6: 255
- number of MTAs in addition to above, from which we rejected emails: 303
(mainly backscatter DSNs to unknown recipients)
- number of MTAs from which we accepted emails over IPv4: 44793
The MTAs, from which the data comes, are responsible for a major part of
the Munich Scientific Network with 4 universities and several colleges and
scientific institutions.
[snip]
I hope this helps and is not to much data.
Michael Storz
Your numbers are interesting but are missing the most important statistics:
- Number of IPv6 MTA Servers you are running (ie: Those that are
receiving email from those 255 IPv6 MTAs and rejecting from those 303
IPv6 MTAs) and how many different FQDNs these MTAs are servicing.
- Number of your IPv6 MTAs which are located via MX records.
- Number of your IPv6 MTAs which must be located via AAAA (ie: Which
belong to a FQDN with no IPv6 MX).
The only REALLY important numbers are the IPv6 MTAs whose FQDN maps
to an MX vs those who only get mapped via an AAAA record.
Having 500 or 50,000 FQDNs mapped to incoming IPv6 servers is only
important if you know how many must be reached/located via AAAA
records due to the lack of a MX record set for the FQDN.
This same measure for IPv4 reachable FQDNs would be interesting but
only to show how many of these FQDNs still lack a MX 15-20 years
after the need for A Lookup was made unneeded by the introduction of
the MX record (except for those MTAs that are still running using a
MTA package that does not support MX lookup or has this ability
disabled).