I think it would also be useful to have as part of the discussion some
real world folk (presumably big ISPs, big DNS outsourcers, or the like)
who's life would actually really and truly become unbearable were we to
lack wildcard support at all.
It is my impression so far that heretofore we've only had anecdotes
regarding the LIKELY and POSSIBLE pain points for these folk. I'd find
it instructive to hear some first hand information on the topic.
Bob
-----Original Message-----
From: owner-ietf-mxcomp(_at_)mail(_dot_)imc(_dot_)org [mailto:owner-ietf-
mxcomp(_at_)mail(_dot_)imc(_dot_)org] On Behalf Of Andrew Newton
Sent: Thursday, May 27, 2004 1:40 PM
To: IETF MARID WG
Subject: Distilling Wildcards for MARID Monday
On Monday, May 31, we will have another XMPP chat session.
Jabber room: marid(_at_)ietf(_dot_)xmpp(_dot_)org
Date: Every other Monday starting April 19
Time: 20:00 UTC
Your Time:
http://www.timeanddate.com/worldclock/fixedtime.html?
day=31&month=5&year=2004&hour=20&min=0&sec=0&p1=0
We are aware that this is a U.S. holiday.
We would like to discuss the topic of wildcards.
On the topic of wildcards, here is food for thought:
1) Should we make a distinction between functionality delivered via
DNS
wildcard vs. tree-walking by the application (either using SOA data or
label iteration)? Are the advantages and disadvantages of a DNS
wildcard vs. tree-walking by the application?
2) Given that DNS wildcards only match on non-existent names (not name
and type), what is the functional difference between covering a
non-existent domain name with a MARID record vs. no record at all?
3) Is there a need to have a MARID record for non-existent domain
names
when there is a need for an MX record for non-existent domain names?
What are the use cases for an MX record for non-existent domain names?
4) If a wildcard-like function is desired, is iteratively dropping the
left-most label by the application an acceptable solution? What are
the advantages/disadvantages to this compared with using the SOA data?
-andy