CURRENT_MEETING_REPORT_


Reported by Mark Knopper/Merit Network

Minutes of the Extensions to OSI for use in the Internet BOF (OSIEXTND)


Dave Katz presented the following slides to the group.  Comments interjected
by the recording secretary are [bracketed].

Motivation
    Plenty of technical work to do
    Expertise is here
    Work will benefit from IETF process
    Liaison relationship likely
    Liaison relationship likely to take awhile
    Experience shows that work transfers

Goals
    Functional, tested protocols
    Further deployment of OSI suite
    SINGLE technical specification
    Progression into ISO/ITU communities
    Backward compatibility
    Concentrate on Network Layer
    Open specifications on line

[There should be significant liaison with TUBA working group
for application and transport layers over CLNP.]

Mechanism
    Zero or more new working groups
    Maintain unofficial relationships with ISO, national bodies
    Utilize official relationships when they are in place
    Forward work through these channels 
    [There are good channels even with unofficial relationships.]
    Cooperation between IETF, IS bodies likely

Liaison Status
    Memorandum of understanding between ISOC, ISO to be written
    Liaison class (A or C) unclear (does it matter?)

[Many people agreed that the relationship status does not matter.
Pragmatic technical interactions are most important.  What role would
other standards organizations take in this arrangement?]

[NOOP, TUBA and ANSI X3S3.3 committees are doing work.]

[Jack Houldsworth was present, representing ISO/JTC1. He works for ICL. JTC1 
has endorsed the idea of relationship. Jack can carry documents from IETF to 
ISO using informal liaison. ISO is very keen to get this work done.]

Standards Status [(for network layer)]
    Multicast
       Addressing to IS in Seoul [will include group addressing semantics]
       CLNP, IS-IS PDAM ballot closing
       Early work on scope control [including automated methods of 
             constraining recipient sets for multicast]
       No work on routing [yet]
       [Multicast IS-IS needed. The intent is to leverage off of
        ongoing work in IETF.]

    IDRP to IS in Seoul

[The rate at which documents progress is proportional to the amount of
work the authors and their organizations are willing to undertake to
manage the process.  At IETF there is more implementation up front.
That is also beginning to happen with ISO standards. There is a
timeline convergence happening between ISO and IETF.]

Possible areas of work

Multicast (network layer)
    anycast
    multicast routing
    structured addressing
    ES-IS

CLNP extensions
    QoS fixes [How useful are QoS features in existing spec?  2nd edition
	       CLNP ballot to be discussed in Seoul]
    provider loose source route
    fix loose source route
    header compression
    MTU discovery
    flows, resource reservation

IS-IS extensions
    Multiprotocol [Part of original IS-IS WG charter]
    Metric expansion

IDRP extensions
    New attributes [IDRP has the capability of tunneling new attributes
	through routers that don't understand them, if appropriate]

ES-IS extensions
    System ID in address administration [for dynamic address assignment.
	Allows host to drive address assignment process.]

IDRP/IS-IS interaction

PPP LCP/NCP for OSI network layer

Host configuration protocols

EON

Mobility

Large network support
    CLNP
    routing
[Radia Perlman of DEC has written documents on CLNP over 
SMDS. This is an internet draft. There is another document that is SMDS-
specific available from SMDS Interest Group.]


The group agreed that a letter should be written to ISO, based on the
presentation and discussion, stating that the IETF has sufficient
expertise to contribute to these areas.  Dave Katz and Dave Piscitello
will draft a document.

This work spans about six IETF working groups.

BGP and IPIDRP Working Groups are meeting jointly (IDRP is BGP5).  IDRP
for SIP is being considered.

Phill Gross expressed concern over the apparent danger of having
different IETF and ISO standards for protocols.  ISO has more precedence
for accepting standards from other groups without change.  IETF tends to
change or rewrite protocols before acceptance.  Peter Furniss said that
the Internet-Draft process will be appreciated by ISO.

Mark Knopper has created a discussion list, osiextnd@merit.edu.  Those
interested in being added to the list should send a request to
osiextnd-request@merit.edu.

Phill Gross suggested to identify work that could be done by existing
IETF working groups, as well as that which could be done by this group
if it is to become a working group.

CLNP over Large Public Data Networks (LPDN) is an area which needs
consideration.  Much of the work is done.  ES-IS and IS-IS protocols
over LPDNs needs further work.  Note that as of this IETF, the IPLPDN
Working Group has ended their work.  Perhaps CLNPLPDN could be handled
as a BOF with identified base documents.

The consensus of the BOF attendees was that a working group should be
formed from these ideas, and relationships should be pursued with ISO.
Dave Piscitello and Dave Katz have drafted a letter and will send it to
the IESG, IAB, and ISO (through Jack Houldsworth).


Attendees

Nick Alfano              alfano@mpr.ca
Bernt Allonen            bal@tip.net
Rebecca Bostwick         bostwick@es.net
Jim Bound                bound@zk3.dec.com
Ross Callon              rcallon@wellfleet.com
George Chang             gkc@ctt.bellcore.com
A. Lyman Chapin          lyman@bbn.com
Richard Colella          colella@nist.gov
Dave Cullerot            cullerot@ctron.com
Toerless Eckert          Toerless.Eckert@informatik.uni-erlangen.de
Dino Farinacci           dino@cisco.com
Peter Ford               peter@goshawk.lanl.gov
Peter Furniss            p.furniss@ulcc.ac.uk
Phillip Gross            pgross@ans.net
Chris Gunner             gunner@dsmail.lkg.dec.com
Susan Hares              skh@merit.edu
Denise Heagerty          denise@dxcoms.cern.ch
Jack Houldsworth         J.Houldsworth@ste0906.wins.icl.co.uk
Chris Howard             chris_howard@inmarsat.org
Geoff Huston             g.huston@aarnet.edu.au
David Jacobson           dnjake@vnet.ibm.com
Philip Jones             p.jones@jnt.ac.uk
Cyndi Jung               cmj@3com.com
Anders Karlsson          sak@cdg.chalmers.se
Dave Katz                dkatz@cisco.com
Sean Kennedy             liam@nic.near.net
Mark Knopper             mak@merit.edu
Rajeev Kochhar           rajeev_kochhar@3com.com
Ton Koelman              koelman@stc.nato.int
John Krawczyk            jkrawczy@wellfleet.com
Robin Littlefield        robin@wellfleet.com
David Marlow             dmarlow@relay.nswc.navy.mil
David O'Leary            doleary@cisco.com
Christian Panigl         christian.panigl@cc.univie.ac.at
Alex Reijnierse          a.a.reijnierse@research.ptt.nl
Victor Reijs             reijs@surfnet.nl
Georg Richter            richter@uni-muenster.de
John Scudder             jgs@merit.edu
Keith Sklower            sklower@cs.berkeley.edu
John Stewart             john@bunter.fdc.iaf.nl
Kamlesh Tewani           ktt@arch2.att.com
Richard Thomas           rjthomas@bnr.ca
Marcel Wiget             wiget@switch.ch
Douglas Williams         dougw@ralvmg.vnet.ibm.com
Rachel Willmer           rachelw@spider.co.uk



Attachment:  Letter to the IESG and IAB



To:   IESG, IAB
From: David Piscitello (Bellcore), David Katz (Cisco)
RE:   Recommendations from"Extensions to OSI for use in the 
      Internet" (osiextnd) BOF regarding future CLNP activities in the 
      IETF.

During the "Extensions to OSI for use in the Internet (OSIEXTND)" 
BOF, attendees identified a number of areas where the IETF might 
apply its experience and expertise to complement and enhance the 
ongoing work within ISO/IEC and the ITU-TS relating to ISO/IEC 
8473 (CLNP). David Katz and David Piscitello were chartered by the 
BOF to annotate the list with Internet activities and standards that 
are or might be relevant to these areas. The results and 
recommendations are as follows.

1)  Multicasting (network layer)

ISO/IEC JTC1/SC6/WG2 is currently working on a multicast 
architecture, addressing scheme and protocol for CLNP. The 
IETF has made considerable progress in the area of multicast 
(see in particular RFCs 1112, 1458) within several of its 
working groups. A working group is likely to be formed within 
the IETF to address "anycast" (the ability to deliver traffic to 
one member of a group), multicast routing, the use of 
structured multicast addresses, and multicast extensions to the 
ES-IS protocol for CLNP.

2)  Extensions to ISO/IEC 8473 and its routing architecture:

The IETF has acquired considerable CLNP deployment 
experience. The coordinating body for developing a CLNP 
infrastructure for the Internet is the Network OSI Operations 
working group (NOOP). Based on the deployment of CLNP and 
more recently, TCP/UDP atop CLNP (TUBA), the IETF has 
identified several extensions to CLNP (improvements to quality 
of service support, provider loose source routing, amendments 
to the current partial source routing parameter); Intradomain 
IS-IS protocol (multiprotocol support, routing metric 
expansion); Interdomain routing protocol (new attributes, 
IDRP/IS-IS interaction), and ES-IS (provision of system 
identification in address administration). OSIEXTND will seek 
working group status within the IETF to continue this work.

3)  Link and network layer control protocols for operation of CLNP 
    over point-to-point subnetworks

The IETF has developed a set of protocols that enable link 
negotiation, authentication, and operation of multiple network 
layer protocols (IP, IPX, CLNP, etc.) over point-to-point 
subnetworks. This work exists either as Internet standards or 
standards in progress (see in particular RFCs 1331 and 1337). 
Attendees to OSIEXTND recommend that the IESG/IAB/ISOC 
encourage ISO/IEC and ITU-TS to study these standards as 
potential future joint standards between the ISOC, ISO/IEC, and 
ITU-TS. 

4)  CLNP header compression and MTU discovery

The IETF has developed a method of compressing IP headers 
for low-speed serial links to maximize throughput across such 
links (RFC 1144); equivalent methods need to be developed for 
CLNP based on the experience acquired by the IETF.  Similarly, 
the IETF has developed a method for determining the 
maximum transmission unit size that may be used between IP 
hosts connected across a multi-hop internetwork route (RFC 
1191); an equivalent means should be developed for deriving 
the maximum subnetwork service data unit size for CLNP.  
Attendees to OSIEXTND expect that a working group will be 
formed to address these mechanisms and solicit contributions 
in these areas.

5)  CLNP "flows", resource reservation

The IETF is examining the notions of flows and resource 
reservation (packet sequencing, allocation of bandwidth, 
processing, etc. to source-destination pairs across an internet).  
Attendees to OSIEXTND believe that the work developed for IP 
is very likely to be directly applicable to CLNP, and expect to 
participate actively in this work.

6)  Host configuration protocols

The IETF has standards in progress that allow hosts to be 
installed and configured with a minimum of manual 
intervention.  Similar work is being pursued in ISO/IEC.  
Attendees to OSIEXTND believe that aspects of the work 
developed for IP are likely to be applicable to CLNP, and 
expects that a working group will be formed to enhance 
functionality in this area. 

7)  Mobility

The IETF is currently exploring methods of supporting host 
mobility.  Attendes to OSIEXTND believes that the approaches 
being developed can be directly applied to CLNP, and expect to 
participate actively in this work.

Attendees to OSIEXTND believe that the work outlined in items (4) - 
(7)  is very likely to be directly applicable to CLNP, and recommend 
that the IESG/IAB/ISOC encourage ISO/IEC and ITU-TS to study and 
possibly participate in this work. 


Respectfully,

________________                    ________________
David Piscitello                         David Katz