Internet Resource Name Search Service BOF (irnss)

Monday, December 10 at 1030-1130
==================================

CHAIRS:	John Klensin <klensin@jck.com> 

DESCRIPTION:

Abstract


This BOF will examine some of the "above DNS" options and 
alternatives for solving problems with Internationalization and
Internationalized access to DNS names and related Internet 
resources.   It will use a collection of existing documents as
the starting point for any subsequent working group or other
efforts.


Introduction


Examination of Internationalization and other issues with the DNS has
lead many members of the Internet community to the conclusion that
the DNS is being expected to serve roles and support functions for
which it is, by design, profoundly unsuitable [1]. While not the only
issue, the two of these desired functions that have received the most
attention to date is the effort to internationalize domain names [2]
and the demand that the Internet's primary naming systems be
sensitive to the requirements of national and international trademark
law and practices [3].  In that context, strong requirements have
been expressed that require sensitivity to language issues (not
merely the sequences of characters/ code points on which the DNS is
based), ability to differentiate names by locale and area of
applicability (such as line of business).  Variations in spelling,
case matching, or character representations or similarities also call
for a system with some "search and match" capability, rather than
merely the strict lookup capability of the DNS.  And, while the
requirement for global uniqueness of network identifiers (such as is
the case with the DNS) remains critical to a seamless,
interconnected, Internet, there are clearly requirements for named
resource searches that are completely sensitive to local (or national
or regional or topical) issues and requirements.


A general architectural model for such a system, using a layered
model and incorporating the existing DNS for the functions for which
it was designed and works well was outlined in [4].  This model
posits two new search mechanisms for finding resources:


* A multihierarchical, faceted, global search system


* A collection of localized search systems (corresponding roughly to
  yellow pages services that may be separated by topic as well as by
  location)


These are described as "sublayer 2" and "sublayer 3" in [5].  [6] is
a specific proposal for implementing the sublayer 2 service using
CNRP [7].


This BOF will consist of:


* A brief review of the three [sub]layer model and the relationship
  among the layers.


* Exploration of sublayer 2 options, using [6] as a base.


* Discussion of whether or not to standardize sublayer 3 and, if so,
  what elements of it.


* Planning a schedule for fleshing out sublayer 2 and developing
  further documents.


It is assumed that the framework document (now [5]) and its
successors will remain a design team effort, eventually progressing
into an Informational or BCP document by four week last call.  It is
also assumed that, if standardization or further framework efforts
are needed within sublayer 2 or 3, these will be done as separate,
per-sublayer, working groups.


Candidate tasks at sublayer 2:


         (i) Determination of the minimal facet set and determination
         mechanisms / registries for the controlled vocabulary facets.


         (ii) Decisions about search mechanisms (e.g., CNRP or otherwise)
         and appropriate details.


         (iii) Understanding and details of distribution, mirroring, and
         caching of data.


Agenda:
(Sharing slot with the Applications Area meeting at 0900 on
Monday) 


         Agenda review and bashing (5 minutes)
         Introduction and overview (Klensin, circa 10 minutes)
         SLS (Mealling and Daigle, circa 15 minutes)
         Where keywords fit in the system (Arrouye et al, circa
             15 minutes) 
         Location of directory services (Hoffman, five minutes)


This leaves a lot of time for discussion, which is the
general idea.  As mentioned on a few mailing lists already,
the intent is to minimize "presentations" and focus on
discussion; the BOF is likely to be incomprehensible to
anyone who is not reasonably familiar with the documents and
mailing list discussion and _no_ time will be allocated to
explaining things to those who are unprepared.



Notes:


[1] See draft-klensin-dns-role-01.txt for a more extended discussion
and RFC 2825 for some additional discussion.


[2] E.g., the efforts of the IDN WG and a number of efforts outside
the IETF, including MINC, an ICANN Committee, and
internationalization or localization efforts in a number of
countries. 


[3] See, e.g., the discussions of Domain Name trademark issues and
dispute resolution policies at http://ecommerce.wipo.org/.


[4] See RFC 2826.


[5] draft-klensin-dns-search-03.txt (forthcoming, RSN)


[6] draft-mealling-sls-00.txt


[7] Popp, N., M.  Mealling, L. Masinter, K. Sollins. "Context and
Goals for Common Name Resolution", RFC 2972. October 2000.