Wildcards in MX Record Domain Names

Jim Reid jim at rfc1035.com
Fri Dec 17 04:19:27 UTC 1999


>>>>> "Joseph" == Joseph S D Yao <jsdy at cospo.osis.gov> writes:

    Joseph> The "better" one would be whichever one better fits the
    Joseph> "truth".  It's always easier to maintain a consistent
    Joseph> story if you're telling the truth.  ;-) And, "Say What You
    Joseph> Mean" is Joe's First Law of Software Engineering.  My
    Joseph> first impression was that the MX trick, above, violated
    Joseph> this.  But within its domain, as Kevin has pointed out, it
    Joseph> does NOT.  It is true.

Not really. At best, it's a controlled lie.

The point is that an internal root implies there is no need to resolve
external names. (Save for things like firewalls and proxy servers at
the perimiter net.) So if someone has decided to implement this, why
should they re-introduce external names to the internal name space?

    Joseph> I will continue to use my sendmail configuration, since it
    Joseph> suits my needs better.  [Everybody thinks they can meddle
    Joseph> with DNS at their whim.  Everybody is afraid to touch
    Joseph> sendmail.cf.  ;-)]

Hmmm. Maybe DNS zone files should have a more complex and obscure
syntax so that people don't meddle with them so readily either? :-)

    Joseph> I do note that one advantage to the MX
    Joseph> solution is that one can specify failover firewalls with
    Joseph> that, but not with the sendmail relay.

Nope. Mail relays can have their own MX records too. And with smart
relays, much simpler configurations on "satellite" mail systems are
possible. "This mail is not local so punt it at smart-mail-relay which
has a bunch of MX records for load balancing, redundancy and all that."


More information about the bind-users mailing list