[prev in list] [next in list] [prev in thread] [next in thread] 

List:       namedroppers
Subject:    X.400 and DNS (was: Re: resent-message)
From:       John C Klensin <KLENSIN () INFOODS ! MIT ! EDU>
Date:       1993-03-02 11:25:56
[Download RAW message or body]

>This can be done
>either with a new record type, or, if you dont want to touch he code at all, by
>adopting a simple naming convention for x.400 mtas -- e.g. if the host name
>terminates by tokens such has ".mta", ".mta-84" or ".mta-88".
  Or, since this is supposed to be transitional anyway, doing a little 
registration and then insisting that the equivalent of MTA-84 or MTA-88
appear in the WKS record.  The argument against testing WKS records for
old-line TCP/IP services is clear, but would presumably not apply to a
transitional service that only a subset of hosts were offering. 
Certainly creating WKS records for those hosts is easier than creating a
new DNS tree in IN (or a new class) for them, and an MTA that would
otherwise be prepared to deal with funny names, etc., could make such a
test with little additional trouble.
   Ordinary, SMTP-only MTAs would omit or ignore the WKS test (per HR)
and try to open SMTP connections on the official port (or as dictated by
MX records).  Those connections would either succeed or fail, which is
the desirable behavior from an installed base standpoint.
   The other thing that hasn't been said explicitly here is that, at
least within the IN class, it is important to preserve the "one host,
one primary name" principle.  Different host names for different
services, e.g., foo.domain for SMTP mail and foo-mta-84.domain for X.400
mail strikes me as a slippery slope with really nasty things at the
bottom.

    --john

[prev in list] [next in list] [prev in thread] [next in thread]