Editor's Note: Minutes received 7/30
CURRENT_MEETING_REPORT_
Reported by Ted Brunner/Bellcore
Minutes of the SNMP over a Multi-protocol Internet
Working Group (MPSNMP)
Four drafts were considered in turn:
1. ``SNMP over IPX''
2. ``SNMP over Appletalk''
3. ``SNMP over OSI''
4. ``Guidelines for the Specification of Protocol Support for the
SNMP''.
The first three are descriptions of how to carry SNMP over particular
transport domains, and are meant to be uniform in their directives.
Thus many of the issues decided in one draft are similarly decided in
the other three. In particular, much of the boilerplate at the
beginning of each draft is the same. All three drafts have
implementations. The issues considered in each come from a check list,
which is enumerated in the final draft.snmp-lists@bir.com.
``SNMP over IPX''
Several issues had come up on the mailing list and were considered in
turn. The most time consuming was the issue of OID assignments for
Transport Domain and initial Party ID used by Party mechanism in
rfc1351-rfc1353. The first plan had been to assign OIDs within the
enterprise subtree. This proprietary technique was dropped and a second
emerged: assign OIDs under the transportDomains and partyAdmin subtrees
of the SNMP Party MIB (rfc1353.) Several draft had decided on this
form. The problem is that independent authors may have conflicting
assignments. The final choice was that the IANA assign these OIDs thus
avoiding conflict. Thus the draft is written with the OID assignments
under the experimental subtree (choice of leaf is left blank e.g., xxx)
and the IANA will re-position them under the mib-2 subtree at the time
the draft becomes RFC.
The wording of the maximum message size had been discussed on the
mailing list, and the wording of the recommendation for a 546 byte
packet was deemed acceptable.
The wording of the recommendation that an agent support one transport
mapping and a manager support as many as necessary was deemed
acceptable.
Citations will be fixed.
The Working Group voiced its support that with these minor changes the
1
draft will be posted as Internet Draft and given the usual two week,
last call before recommendation for promotion to Proposed Standard.
SNMP over Appletalk''
The same decisions regarding OID assignment were applied here.
On the mailing list was a discussion of the convention for displaying
the Appletalk name, as used by the Name Binding Protocol when binding to
a DDP address. There are three fields in an Appletalk name, expressed
in the form: ``object:type@zone''. The delimiter is the issue: either
a special character, or a length field. Although special characters are
commonly used in expressing an Appletalk name, it was felt that some
utility may be gained with a length field. No manager should break when
displaying non-printable ASCII characters. So the convention was
adopted whereby the octet preceding each of the three fields contains
the length of the field (1-255). (e.g., 6object4type4zone) That length
will commonly be an un-printable ASCII character.
The Working Group voiced its support that with these minor changes the
draft will be posted as Internet Draft and given the usual two week,
last call before recommendation for promotion to Proposed Standard.
``SNMP over OSI''
The OID assignment and language on supporting one transport mapping were
brought into alignment with the other drafts.
At the last meeting of the Working Group, the decision to run the SNMP
exclusively over CLTP was made. Subsequent feedback from other OSI
groups has been positive, according to the OSI Integration Area
Director.
A question was raised about GOSIP compliance. Some history was
recalled. The original impetus for this draft came from an OSI
Integration Working Group (NOOP), a router vendor, and their management
needs for CLNP pilot projects. That Group is happy with the form of
this draft.
The Working Group voiced its support that with these minor changes the
draft will be posted as Internet Draft and given the usual two week,
last call before recommendation for promotion to Proposed Standard.
``Guidelines for the Specification of Protocol Support of the SNMP''
The OID assignment will be brought into alignment.
Because of its similarity with the other three drafts, it was proposed
that this draft receive similar treatment, with respect to timing and
promotion, even though the Working Group had little time to read it
before the IETF meeting.
2
The Working Group voiced its support that with these minor changes the
draft will be posted as Internet Draft and given the usual two week,
last call before recommendation that it be an informational RFC.
Having completed its Charter - and in the absence of new work showing up
soon to force re-examination of the decision - the Working Group agreed
to disband itself.
Attendees
Jonathan Biggar jon@netlabs.com
Steve Bostock steveb@novell.com
Theodore Brunner tob@thumper.bellcore.com
Philip Budne phil@shiva.com
James Davin jrd@ptt.lcs.mit.edu
James Halpin halpin@turkey.sw.stratus.com
Bob Jeckell rrj@3com.com
Greg Minshall minshall@wc.novell.com
David Piscitello dave@sabre.bellcore.com
Mike Ritter mwritter@applelink.apple.com
Marshall Rose mrose@dbc.mtview.ca.us
David Waitzman djw@bbn.com
Steven Wong wong@took.enet.dec.com
3