CURRENT_MEETING_REPORT_
Reported by John Veizades/FTP
Minutes of the Service Location Protocol Working Group (SVRLOC)
The Service Location Working Group held two sessions at this IETF
meeting. The first session focussed on the Internet-Draft submitted by
the chairs, ``Service Location Protocol.''
This discussion was divided into several areas:
o The Base Protocol Specification
o Authentication
o Predicate Language
o Directory Agents
The base protocol specification had several issues in the size of
several fields. The locale field was extended to a 16-bit value with
the need to look at any international specifications that define
language specifiers. (The chair found that ISO 639 defines many
language specifiers and that the mime working group is working on
extending this work to include dialects of spoken languages increasing
the size of this field to four characters.)
The group made the recommendation that the specification writers look at
the work that the Common Authentication Technology Working Group (CAT)
is doing as a way of defining the authentication types and also to look
at the GSS API work. The authentication length field was increased to
16 bits.
The predicate language was hashed out, and the addition of wild card
methods on strings was added. The conditional operators that were
accidentally left out of the specification were also added.
The afternoon session was devoted to the discussion of the directory
agent interaction in the protocol. The issues that needed to be
resolved were that the directory agent is the point of scaling in the
protocol and that it is necessary that the directory agents need to
solve all the scaling issues.
The following protocol was defined for directory agents:
o Directory agents have a concept of a scope that they are
responsible to support. The scope is a text string. Samples of a
scope would be ``engineering'' or ``marketing.'' During a
directory agent discovery, the directory agent passes to the user
agent the scopes that are available on the network and the scope(s)
that they are acting for. The user agent needs to find the
directory agent that supports the scope they are searching in.
This is done by sending a service query with the directory agent
and the scope that is interested in the appropriate directory agent
responds.
o Directory agents need to find themselves and exchange the scopes
that they know of.
o Service agents need to register with all directory agents that
support the scope they have chosen to be in.
The next question is, ``How do directory agents and sites advertise
themselves in the Internet at large?'' The proposals that were
considered were:
o Advertising in a directory service like DNS or X.500, and
o Advertising in a resource discovery service like gopher or WWW.
The discussion continued with the statement that sites may want to
advertise their service in several taxonomy domains, these include a
white pages system indexed by the public name of the advertiser and/or a
yellow pages type service indexed by service provided, geographical
location, etc.
The chairs will post an updated draft by the end of the year.
Implementation of this protocol is proceeding at FTP Software and other
implementors are being sought out.
Attendees
Andy Adams ala@merit.edu
Steve Alexander stevea@lachman.com
James Allard jallard@microsoft.com
Frank Ciotti frankc@telxon.com
Chuck de Sostoa chuckd@cup.hp.com
Mei-Jean Goh goh@mpr.ca
Roland Hedberg Roland.Hedberg@rc.tudelft.nl
Kevin Jackson kjackson@concord.com
Scott Kaplan scott@ftp.com
Gordon Lee gordon@ftp.com
Glenn Mansfield glenn@aic.co.jp
Matt Mathis mathis@psc.edu
Mike Ritter mwritter@applelink.apple.com
Michael Safly mike.safly@msfc.nasa.gov
Martin Schulman schulman@smtp.sprint.com
Robert Stevens robs@join.com
Larry Tepper ltepper@compatible.com
John Veizades veizades@ftp.com
Walter Wimer walter.wimer@andrew.cmu.edu