CURRENT_MEETING_REPORT_


Reported by Peter Ford/LANL and John Scudder/Merit

Minutes of the joint session of TUBA, NOOP and RARE CLNS Working Groups

Marcel Wiget reported on the RARE CLNS Pilot.  CLNS was run over a
variety of media across Europe.  The objectives of this project were
routing coordination and testing of CLNS implementations to feedback to
vendors and implementors.  Systems were deployed on a wide variety of
platforms including cisco, DEC, Sun, BSD systems, Novell and others.
Applications tested included X.400, X.500, FTAM, DECnet, VT, TUBA, OSI
Ping, and traceroute.

RARE developed databases for the distribution of addresses (NSAP, NSEL,
TSEL, etc.)  and routing topology.  Tools used for managing the network
included clnp_traceroute, tcpdump (INRIA), bis_ping, and proxy managers
for sunnet manager.  They also have extensions to the Internet rover.
There are also tools for managing consistency between DNS and host
files.  All of these are available via anonymous FTP from ftp.rare.nl or
anonymous FTAM/FTP from nic.switch.cs:/network/clns.  To be added to the
discussion list, send a request to wg-llt-clns-request@rare.nl.

Current activities include a task force for CLNS routing coordination.
They are discussing how to take advantage of the RIPE routing registry
which will require developing a schema for representing this
information.  There is also discussion of forming TUBA and IDRP task
forces but there is a sense that perhaps it is best to do that sort of
work in the IETF.

In discussing the scale of the CLNS pilot, it was interesting to note
that the UK has now turned on Phase V DECnet, and that the Italian high
energy physics lab has over 100 CLNS routers at this time.  It is
estimated that there are over 200 CLNS routers in Europe at this time.

Dave Katz reported that the IDRP specification is on its way to Geneva,
and that it will become a full International Standard at the October ISO
meeting in Seoul.

A spirited discussion was held on the use of IS-IS for routing the
global CLNS network.  Many felt that the size of the current CLNS system
was tool large to be managed with IS-IS. There are also issues of
whether or not certain networks (e.g.  CERN) would want to absorb the
routing of the entire world into their own world.  Dave Katz suggest
building a ``virtual core'' for doing some of this.  Further discussion
of this will be done on the NOOP mailing list.

There are now many TUBA implementations:


   o PC/MS-DOS by Richard Colella.  Source code is available via
     anonymous FTP from osi.ncsl.nist.gov.

   o SunOS - kernel mode TUBA by Francis DuPont.  Source not available
     unless you have a BSD 4.3 license.

   o Francis Dupont reported on the user mode network layer translator
     implemented in user space on SunOS that currently runs on top of
     Sun's OSI implementation.  The sources are available; send mail to
     Francis.

   o Peter Ford reported on further work on top of Keith Sklower's
     implementation of TUBA on the BSDI platform.  If you have a BSD 4.3
     source code license you can get these sources for BSDI (send mail
     to Peter).  There are now implementations of TELNET, FTP (using
     Dave Piscitello's FUBAR specification), finger, and there is a
     version of inetd which manages CLNP sockets for Unix
     implementations.

   o Cyndi Jung reported that 3Com routers continue to support TUBA.

   o Dino Faranacci reported cisco routers also support TUBA. Images can
     be obtained from ftp.cisco.com:/beta921-dir/*.931_O.14.Z where * is
     replaced with the cisco model you have.  You need to send mail to
     clns-beta-request@cisco.com to get on the mailing list if you pick
     up one of these images.


Cyndi Jung reported on CLNS efforts for the next INTEROP in San
Francisco (August 21 to August 26).  She noted that this would be a good
opportunity to demonstrate TUBA; George Chang is working on getting some
spots around the show floor to demonstrate TUBA. The mailing list for
discussion is osiig@spitfire.interop.com.

The meeting ended with a discussion of where to discuss TUBA deployment.
It was decided that this should happen on the NOOP mailing list.