[jig] IETF DNSEXT work of variants and resolution

[Avri Doria <avri@xxxxxxx>]

Hi,

During today's meting I agreed to send several bits of information to this list 
on what was going on in the IETF in relation to the technical requirements on 
variants.  Here are the first bits.  I also agreed to write up a synthesis of 
some of the ongoing discussion on the issue.  That will take me until sometime 
in January since I have some day-job work to deliver first.



Excerpt from the rechartering of the DNS EXT WG in the IETF announced on 7 Dec 
2010

* Mechanisms to alias DNS trees or parts thereof

While the DNS offers two mechanisms for aliasing DNS labels -- CNAME
and DNAME -- neither of these provides the support necessary to alias
completely one part of the DNS tree as another part.  There are claims
that the restriction has proven to be too great in practice,
particularly with burgeoning deployment of IDNA and the need to
provide domain name variants.  The issue is made more complex by
DNSSEC

The DNSEXT WG will evaluate ways to provide such aliasing, to add
metadata to zones to allow easier operation of zones when such
aliasing is needed, or both.  The WG will also provide an
informational document outlining the various strategies available,
what they might be used for, and what their limitations are.  It is
possible that the WG will conclude no aliasing or metadata support is
possible, or that none of the proposals so far made are adequate.

Before formal adoption of any work item at least 5 working group
participants must publicly state that the item is within charter and
is a worthwhile item for further study.


Relevant Milestone

Nov 2011       Document on uses and limitations of different alias
                   techniques to IESG


The curent draft discussing the issue is:
http://datatracker.ietf.org/doc/draft-yao-dnsext-identical-resolution/

Abstract

   This document attempts to describe a set of issues that arises from
   the desire to treat a set or group of names as "aliases" of each
   other, "bundled," "variants," or "the same," which is problematic in
   terms of corresponding behavior for DNS labels.

   With the emergence of internationalized domain names, among other
   potential use cases, two or more names that users will regard as
   having identical meaning will sometimes require corresponding
   behavior in the DNS.  It's not clear how to accommodate these
   requirements for behavior of such names in DNS resolution; in
   particular, it's not clear when they are best accommodated in
   registry practices for generating names for lookup in the DNS,
   existing DNS protocol elements and behavior, or some set of protocol
   elements or behavior not yet defined.  This document attempts to
   describe some of these cases and the behavior of some of the possible
   solutions discussed to date.

   NOTE: Even more than usual, version -02 of this document is a "work
   in progress".  Additional updates may be expected between the date of
   this document and the DNSEXT meeting in Beijing, and can be found at
   http://users.isc.org/~woolf



 I am checking with Suzanne to see if there have been any updates since this 
was released.  This is the latest copy in the IETF draft repository.

cheers,

a.