CURRENT MEETING REPORT


Minutes of the IP over ATM Working Group

Reported by Mark Laubach, Com21, Inc.


Morning Session

The meeting was opened by Mark Laubach and covered the usual 
agenda bashing.  The charter was recently updated and approved by 
our Area Directors.  IPv6 over ATM treatment was added and some 
statements on how the ipatm groups relates to the ROLC, RSVP, and 
Integrated Services Working Groups.  The two server synchronization 
presentations were moved together in the morning session to permit 
easier comparison.

Mark Laubach and Andy Malis (the ROLC chair) presented a chair's 
statement regarding rolc and ipatm coordination of the transition from 
ATMARP to ATMARP + NHRP to NHRP.  A mention was made that 
the AD's stated that a step-wise transition plan is required for any 
evolution of a widely deployed proposed standard. Interoperability 
must be maintained so that existing implementations don't break and 
that within the LIS, all clients must be able to resolve the address of 
any other client by required default behavior regardless of 
implementation or deployment.

Andy Malis presented ROLC status. The packet formats for NHRP and 
MARS have been "harmonized".  The ROLC Working Group consensus 
was to remove NHRP mention from the RFC1577 update Internet-Draft.  
As this originally was done as part of the Danvers joint IPATM/ROLC 
meeting session, ROLC and the IPATM chairs felt that ROLC had the 
OK to remove NHRP from the work in progress.

Andy Malis presented a proposed ROLC/IPATM Liaison to the ATM 
Forum.  The liaison included two sections, one suggesting leveraging the 
LANE Configuration Server (LEEKS) for IETF uses, the second asking 
for LAN Emulation Configuration Server redundancy.  The consensus 
was that the liaison be submitted as is and that it just represents the 
first communication on the subject.  There was much discussion on this 
topic. It was clear that more discussion on both auto-configuration and 
server synchronization needs to be discussed in order to proceed with 
any formal definition of how to incorporate LECS's or other 
configuration servers into the classic IPATM standards development.

The ATM Forum Liaison status was given as an informal report from the 
ATM Forum member attendees.  George Swallow reported that the 
Multiprotocol over ATM (MPOA) effort is work in progress will be 
relying on the ROLC NHRP and IPATM mars work.  The IETF <> ATM 
Forum mailing lists, as reported at the Stockholm ATM Forum BOF, 
still have not been set up by the ATM Forum.  Keith McCloghrie 
presented an informational overview of an ATM Forum contribution for 
providing an ATMARP server address via the ILMI.  This was 
presented as a simple mechanism for the short term.  An issue was 
raised if the mechanism can support hosts that are in multiple LISs.  
The consensus was that Keith and others raise this issue at the ATM 
Forum meeting next week.

Ken White and Maria Greene presented an update of the RFC1577+ 
MIB.  A question was raised about multiple addresses.  The MIB creates 
a virtual interface per IP address and handled in the net to media 
table.  This work will continue and be presented for real review by the 
next IETF meeting.  The NHRP MIB editors and the IPATM MIB editors 
are sharing information for coordination opportunities.

The IPATM chair noted that there will likely at a joint session at the 
next IETF meeting for the integrated-services, RSVP, and IPATM 
working groups.

Steven Berson presented an information and discussion treatment of an 
approach to RSVP over IP over ATM. Issues were raised about the 
transition from currently active Virtual Channels (VCs) to new VCs 
with the new reservation. RSVP over the ATM model has some 
problematic areas. RSVP defaults to best effort if the reservation fails, 
which seems to indicate that a multicast group would need to maintain 
a best effort VC along with a Quality of Service (QOS) VC. Sources do 
not get reservation error messages.  Comments were made about all this 
needs to work with the routing pieces.  Comment was also raised that 
the aggregation model appears to be very complicated.  A much closer 
work for IP over ATM might be to take the integrated services API and 
map that over VCs.

Andy Malis chaired the remainder of the morning session.

Mark Laubach presented the RFC1577 update Internet-draft model 
draft-ietf-ipatm-classic2-00 of the epidemic database distribution 
model.  Questions are raised about how this relates to flooding models.  
The observation was raised as to why not use point-to-multipoint 
capabilities and the answer was that the Epidemic algorithms are 
based on point to point connections with feedback.  A question was 
raised with regards to patent and licensing issues.  Mark will explicitly 
check with Xerox PARC.  [Just afterthe IETF meeting ended and before 
these minutes were produced, Bryan Lyles from Xerox PARC reported 
back via email that there is no known patent issues regarding the 
epidemic method referenced in the classic2 draft.]

Carl Marcinik presented the Distributed ATMARP model based on the 
work in draft-marcinik-ipatm-dist-atmarp-00.  The work is based on a 
set of servers connected by a full mesh of point-to-multipoint VCs.


Evening Session

Between sessions the ROLC and IPATM chairs, along with the Routing 
Area Director (Joel Halpern) and the Internet area directors (Susan 
Thomson and Frank Kastenholz) met to discuss the transition plan and 
direction of the working groups.  The recommendations to the working 
groups from this meeting are that the 1) classic2 draft should proceed 
with ATMARP and without reference to NHRP, 2) the classic2 draft 
should proceed and include ATMARP server synchronization, 3) that 
IPATM and ROLC working groups should coordinate MIA development, 
explore leveraging the same server synchronization methods, and keep 
in mind that future IAB/IETF directions may require authenticated 
address registration (i.e., this is a heads up and don't preclude a 
straightforward evolution to this as part of the synchronization work), 
and 4) a future ROLC RFC will be used to specify how to substitute 
NHRP for ATMARP; i.e., ipatm will continue to develop ATMARP in 
its work at this time.

Andy Malis acted as working group chair for the continued server 
synchronization presentations and discussions.

A discussion was held comparing the two ATMARP synchronization 
schemes, draft-ietf-ipatm-classic2-00 and draft-marcinik-ipatm-dist-
atmarp-00.  Among the issues discussed were ju st how many ATMARP 
servers does the classic LIS need to support?  Will the server protocol 
need to be ATMARP only or multiprotocol?  There was much discussion 
relating to scaling, number of clients, servers, and sites to support, time 
to market, usability by rolc, etc.  There was no clear consensus on these 
items, therefore the acting chairman (Andy Malis) presented a set of 
goals for the synchronization issue that would be used as proposals. 
Mark Laubach will distribute these out on the mailing list for comment 
and refinement.

Tim Smith presented draft-smith-ipatm-bcast-01 for discussion.  The 
consensus was to put this on the work plan and keep as a separate 
document with target towards the proposed standards track.  Tim will 
release an update to the document based on the new ipmc-10 update.

Grenville Armitage presented draft-ietf-ipatm-ipv6nd-00.  Time was 
short and the presentation covered an overview of the draft.  Due to 
time, no issues were raised.  The chair recommended that people read 
the draft and send comments to the list.

Peter Schulter presented draft-schulter-ipv6atm-framework-00.  Due 
to time constraints, Peter could only present an overview.  The issue was 
raised that the method presented in the draft needs to be compared to 
rolc's nhrp model.  The chair recommended that people read the draft 
and send comments to the list.

Carl Marcinik presented draft-marcinik-ipatm-auto-arp-00.  Due to 
time constraints, Carl could only present a brief overview of the draft.  
The chair recommended that people read the draft and send comments 
to the list.

There was not enough time to present and discuss the work plan agenda 
item.  The refinement of the 1996 work plan will be discussed on the 
list.