Internet Info 1997 December/Internet_Info_CD-ROM_Walnut_Creek_December_1997.iso/ietf/93mar/atmmib-minutes-93mar.txt

Text File | 1995-03-01 | 8KB | 204 lines

CURRENT_MEETING_REPORT_



Reported by Kaj Tesink/Bellcore

Minutes of the Managing ATM with SNMP BOF (ATMMIB)

The goal of the ATMMIB BOF was to determine if there was sufficient
interest to create a working group to develop managed objects for ATM.
Editor's Note (md):  Following the IETF meeting this Group became a
working group and took the name AToM MIB Working Group (ATOMMIB).

Agenda


   o Overview, Proposed Charter.
   o Proposals.

      -  ATM MIB Strawperson Proposal (Masuma Ahmed)
      -  ATM ILMI (Keith McCloghrie)
      -  SONET MIB (Kaj Tesink)


Overview - Kaj Tesink

Following unanimous agreement that a working group is necessary to
develop managed objects related to ATM, Kaj identified the following
``terms of reference'' as the basis for the proposed Charter of the ATM
MIB work effort:


   o Priority on ATM PVCs.

   o Keep the MIB small.

   o Priority is creation of an SNMPv1 managed objects.  Support of
     SNMPv2 may be a future effort.

   o Difference between interface and end-to-end view of NM information
     for ATM.


Kaj discussed the relationship of the proposed ATM MIB with physical
layer MIBs.  He suggested that the Working Group include work on the
SONET MIB which was posted by Tracy Cox.  The Trunk MIB Working Group
which had developed MIBs for DS1/E1 and DS3/E3 no longer exists.  Ayal
Opher cautioned against assuming that the SONET MIB is ``done''.  Since
there are more implementations of SONET equipment, folks may pay more
attention and have more comments on the proposed SONET MIB module.

There was discussion on the different perspectives on ATM-related
network management information:


   o Local DTE view of ATM-based communication over the local interface.
   o Network view of the local interface.
   o End-to-end view of the communication path.
   o ATM switch management - both private and public.
   o ATM-based service information - both private and public.


These different perspectives drive identification of different sets of
managed objects.

Based on this discussion, a Charter was agreed upon that directs the
proposed Working Group to undertake work on ATM and SONET management.

Proposals


   o ATMMIB Proposal - Masuma Ahmed

     Masuma gave an overview of a Bellcore strawperson proposal for an
     ATM MIB. See the copy of the viewgraphs for a detailed list of
     proposed management information.  The strawperson MIB is based on
     the ILMI, and includes configuration information about the ATM
     interface, ATM layer statistics, but (unlike the ILMI) also VPL/VCL
     level performance information.  Virtual Channel Links are the
     interfaces between two adjacent ATM devices.  A Virtual Channel
     Connection is the set of VCLs used end-point to end-point.  A VCL
     is identified by a Virtual Channel Identifier (VCI). Similarly, a
     Virtual Path Connection consists of VPLs, identified by VPIs.
     Issues identified during the discussion included:

      -  The relation to the existing ILMI MIB (see Keith's talk).
      -  The need for end-to-end PVC configuration information.
      -  The modelling of ATM in relation to MIB-II interfaces table.
      -  The need to keep the MIB small.
      -  The need to coordinate modeling of PVCs with FR MIB effort.
      -  The need for AAL performance statistics.
      -  The need for VPC/VCC level performance statistics.


   o ATM Forum ILMI - Keith McCloghrie

     Keith gave a presentation on the ATM Forum's Interim Local
     Management Interface ILMI, stressing that even though the ILMI uses
     SNMP packet formats, the purpose of the ILMI is not network
     management, but rather ``interface management''.  However, since
     the management information needed for interface management is a
     subset of the information needed for network management, the ILMI's
     MIB should be appropriate for use by network management as well as
     for interface management.

     The ATM Forum's ILMI MIB was designed to be symmetric -- that is,
     to have the same MIB on each side of the interface.  It currently
     contains read-only objects.  The ILMI MIB contains information on
     the physical port (e.g.,operStatus), ATM layer, ATM cell, and
     VPC/VCCs.  There are no VCL or VPL level statistics in the ILMI.

     Keith recommended that the proposed IETF Working Group use the ILMI
     MIB as a base for the development of IETF ATM MIB(s).  The
     management of the local interface should use a common MIB
     module/table/group.

   o SONET MIB - Kaj Tesink

     Kaj gave a short presentation of the SONET MIB. The SONET MIB has
     been posted as an Internet-Draft for a long time.  It is consistent
     with other trunk MIBs.  A minor change may be needed on table
     indexing.  Interested people are encouraged to review this MIB.
     Internet-Draft:  ``Definitions of Managed Objects for the SONET
     Interface Type'', Jan.1993, Tracy A. Cox, Kaj Tesink,
     <draft-ietf-cox-sonetmib-01.txt>.


Conclusions of the BOF


   o There is interest in forming a Working Group to develop standard
     managed objects for ATM and SONET.

   o Deirdre Kostick suggested calling it the AToMMIB Working Group,
     with the ``o'' signifying SONET, and the ``AToM'' as a reminder to
     keep MIBs small(!).

   o It is likely that there will be multiple MIB modules for ATM
     developed to support the different management perspectives
     identified during the BOF.

   o The MIB module(s) for management for the local interface should
     build on the ATM Forum ILMI MIB.


Attendees

Masuma Ahmed             [e-mail]
David Arneson            [e-mail]
David Battle             [e-mail]
Nutan Behki              [e-mail]
Caralyn Brown            [e-mail]
Jack Brown               [e-mail]
Theodore Brunner         [e-mail]
John Chang               [e-mail]
Anthony Chow             [e-mail]
George Clapp             [e-mail]
Michael Collins          [e-mail]
Tracy Cox                [e-mail]
James Davin              [e-mail]
M.J. Dixon               [e-mail]
Kurt Dobbins             [e-mail]
David Engel              [e-mail]
Kenneth Giusti           kgiusti.chipcom.com
Patrick Hanel            [e-mail]
Juha Heinanen            [e-mail]
Don Hofacker             [e-mail]
Mark Kepke               [e-mail]
Moshe Kochinski          [e-mail]
Deirdre Kostick          [e-mail]
David Lovering           [e-mail]
Andrew Malis             [e-mail]
Jun Matsukata            [e-mail]
Keith McCloghrie         [e-mail]
Evan McGinnis            [e-mail]
George Mouradian         [e-mail]
Tom Nisbet               [e-mail]
Bill Norton              [e-mail]
Zbigniew Opalka          [e-mail]
Ayal Opher               [e-mail]
Joe Pagan                [e-mail]
Michael Patton           [e-mail]
Drew Perkins             [e-mail]
Venkat Rangan            venkat@.metrix.com
Kenneth Rodemann         [e-mail]
Dan Romascanu            [e-mail]
Marshall Rose            [e-mail]
Jon Saperia              [e-mail]
Andrew Schmidt           [e-mail]
Kanan Shah               [e-mail]
Andrew Smith             [e-mail]
Subbu Subramaniam        [e-mail]
Terry Sullivan           [e-mail]
Sally Tarquinio          [e-mail]
Kaj Tesink               [e-mail]
Stephen Tsun             [e-mail]
Mario Vecchi             [e-mail]
James Watt               [e-mail]
Peter Wilson             [e-mail]
Yow-Wei Yao              [e-mail]
Kiho Yum                 [e-mail]