The IETF SNA NAU Services MIB WG and the AIW APPN MIBs SIG held a
joint meeting in Durham, NC at AIW 17 (9/15/98). Bob Moore chaired
the meeting. The following people were present:
- Bob Moore IBM remoore@us.ibm.com
- Ralph Case IBM caser@us.ibm.com
- Michael Sherman Bay Networks msherman@baynetworks.com
- Camille Bijjani Nortel cbijjani@nortel.ca
- Gary Dudley IBM dudleyg@us.ibm.com
- Bill Douglas IBM bdouglas@us.ibm.com
- Marcia Peters IBM mpeters@us.ibm.com
- Sudhakar Chellam IBM svelkant@us.ibm.com
- Matthew Finlayson DCL mcf@datcon.co.uk
- Mark Mitchell DCL mm@datcon.co.uk
- Andy Rogers DCL ar@datcon.co.uk
- Bob Clouston Cisco rclousto@cisco.com
- Jim Cobban Nortel jcobban@nortel.ca
- Don McGinley Nortel/Bay mcginley@baynetworks.com
There were three items on the meeting agenda:
1. Status report on the group's current MIBs.
Bob Moore reported that the EBN MIB, the APPN-TRAP MIB, and
version 2 of the APPN MIB are all in IESG review. These MIBs
are expected to receive final IESG approval, and be issued as
Proposed-Standard RFCs, in the next month or two.
Bob also reported that the HPR-IP MIB has had WG / SIG consensus
for some time, and thus is ready to move on to Area expert review
and then to IESG review. Bob will initiate this process.
2. Possible extensions to the HPR MIB for HPR performance monitoring.
Ralph Case led this discussion, based on his recent experience
with responsive mode ARB, an extension to the HPR architecture
that IBM introduced to the AIW at this meeting. Ralph identified
a number of possible additions to the hprRtpTable, and some other
candidates were mentioned in the meeting. Possible additions to
this table include:
- enumeration identifying ARB mode
- count of ARB cold starts
- counts of various ARB control messages
- count of bytes retransmitted
- count of times the Competer acted
- objects related to the ARB leaky bucket
- enumeration identifying the row in the Link Speed table
currently being used
There was some discussion of traps for Link Speed table changes,
but the consensus was that adding these traps would not be a
good idea.
The group decided that extensions to the HPR MIB in this area
look promising enough to warrant updates to the two charters,
committing to work on them. The IETF WG charter will be updated
to reflect a first Internet-Draft of a revised HPR MIB by
February, 1999. In addition to new objects for performance
monitoring, the scope of the revision will also include
possible removal (technically, deprecation) of objects in the
current MIB that no longer seem useful. The plan is for the
document resulting from this work to be issued as a second
Proposed-Standard RFC, obsoleting the current HPR MIB (RFC 2238).
Since we ordinarily get some level of AIW consensus on a document
before we issue an Internet-Draft, the AIW MIBs SIG charter will
be updated to reflect AIW AP approval for the updated HPR MIB by
the end of 1998.
3. Subsetting the APPC MIB for session monitoring.
We initially discussed this topic at AIW 16, and in fact added it
to our WG charter. The inputs to the discussion were independently-
derived subsets of the RFC 2051 APPC MIB from Bob Moore and Bob
Clouston. While we didn't review these proposals object by object,
it was clear that they were substantially in agreement with each
other.
BobM and BobC agreed to post the two subset proposals to the
mailing list, and then, after suitable discussion, to produce a
revised Internet-Draft of the APPC MIB. The target date for this
Internet-Draft is December 1998. In addition to deprecating
objects from the RFC 2051 APPC MIB, the new document will have to
move from the 1996 "standard" for MIB documents to the 1998 level.
As with the APPN and HPR MIBs, the plan for the subsetted APPC
MIB is to request approval as a second Proposed-Standard RFC,
obsoleting RFC 2051.
One topic not on the published agenda was also discussed at the
meeting: planning how to demonstrate compatibility among independent
implementations (of any of the WG's MIBs), for eventual progression
of the MIBs to Draft Standard. Bob Clouston agreed to re-post to the
mailing list his proposal from last winter for a "reference" APPN
network that vendors could configure independently. With identically
configured networks, multiple vendors' implementations should return
not only the same MIB objects, but also the same values for these
objects. The group agreed that if this could be demonstrated, it
would constitute the proof of consistent independent implementations
required for advancement to Draft Standard.
The next planned meeting for the SIG / WG will be at AIW 18, which
will occur some time in 1999. Meanwhile, documents will be
distributed and discussed on the mailing lists.
Minutes submitted by Bob Moore.
Regards,
Bob
Bob Moore
IBM Networking Software
+1-919-254-4436
remoore@us.ibm.com