ietf-smtp
[Top] [All Lists]

Re: draft-daboo-srv-email and 'creative use' of DNS SRV

2010-01-10 14:54:40

Ned Freed wrote in the message archived at
  <http://www.IMC.ORG/ietf-smtp/mail-archive/msg05917.html>:

...

But even this isn't really necessary to specify precedence among
protocols.  The specification could say that pop/pops/imap/imaps
records should be considered as a group, and absent client
configuration to the contrary should be used in the order specified
by their individual priority fields.  For example, an imaps record
with a higher priority than any imap record would indicate imaps
is preferred over imap.

RFC 2782 appears to allow schemes like this.  But whether it's worth
defining such a thing is another matter.  My personal opinion is that
it's just a bit outside my comfort zone in terms of complexity, but if
others feel strongly that servers should be able to specify protocol
preferences this is a way to do with it with no additional DNS records
or supplementary protocols.

Folks,
please take a look at the draft Olafur Gudmundsson and I recently
have posted to clarify RFC 2782 (and hence intended for PS),
  http://tools.ietf.org/html/draft-gudmundsson-dnsext-srv-clarify-00

According to these clarifications (if approved), such specific rules,
*if and only if unambiguously specified* would IMO well be possible.
Absent such application-specific documentation, the standard SRV
naming and usage rules apply.  Cyrus Daboo's draft would likely be
the proper place to incorporate such specific rules, if desired.

Kind regards
  Alfred.

-- 

+------------------------+--------------------------------------------+
| TR-Sys Alfred Hoenes   |  Alfred Hoenes   Dipl.-Math., Dipl.-Phys.  |
| Gerlinger Strasse 12   |  Phone: (+49)7156/9635-0, Fax: -18         |
| D-71254  Ditzingen     |  E-Mail:  ah(_at_)TR-Sys(_dot_)de                    
 |
+------------------------+--------------------------------------------+

<Prev in Thread] Current Thread [Next in Thread>
  • Re: draft-daboo-srv-email and 'creative use' of DNS SRV, Alfred Hönes <=