CURRENT_MEETING_REPORT_

Reported by Marcus Leech/Bell Northern Research

Minutes of the Authenticated Firewall Traversal Working Group (AFT)

The AFT Working Group meeting was held on Tuesday, 18 July.  There were
about 50 attendees---less than in Danvers, but Stockholm attendance in
general was less than Danvers.

The issues brought to the table were:


   o Extending the METHOD negotiation model to include exchange of
     arbitrary octet strings as METHOD descriptors.  There was general
     consensus at the meeting that staying with the current model is a
     good thing (that is, no one objected).  The chair's position is
     that ``better is the enemy of good enough.''

   o Feedback from implementors on UDP DESTROY has indicated that it is
     not a terribly useful construct.  There was general consensus that
     removing UDP DESTROY was a good thing.  The new proposed model is
     that the connection that the UDP ASSOCIATE arrives on defines the
     lifetime of the UDP association it creates.

   o It is expected that interopability testing may commence soon,
     perhaps as early as mid August, it appeared that three independent
     implementations are already in progress, with the BNR-based one
     being the basis for the ICL/SESAME implementation.  Volunteers are
     needed to step up to the plate.

   o IMAP was brought up as a possible framework for plugging SOCKS
     into.  The chair (having reviewed the IMAP documents) feels that it
     belongs underneath SOCKS as a METHOD, rather than as something that
     SOCKS gets buried in.


The chair generally feels that (modulo some clarifications, and
incorporation of the technicals discussed in Stockholm) the document is
close to being ready to go through Last Calls and ``Proposed.''  The
chair expects to issue a new Internet-Draft on the base protocol by
28 July.