CURRENT_MEETING_REPORT_
Reported by Michael Erlinger/Micro Technology
Remote LAN Monitoring Minutes
o Copies of ``How to Write SNMP MIB,'' the Novell LANtern MIB
(available on-line), and preliminary MIBs from Spider, NAT, and
Frontier were distributed.
o Wednesday evening meeting was scheduled.
o Other Working Groups involved in similar activities were discussed:
Accounting Working Group (accounting-wg@bbn.com), Operational
Statistics (new group), and Benchmarking Methodology
(bmwg@harvisr.harvard.edu).
o The Working Group Charter was quickly reviewed and it was noted
that the effort is correct, but that various milestone dates were
now changed.
o The Chair wanted make it clear that writing assignments would be
made prior to the close of the IETF meeting.
o Remote LAN monitoring could be accomplished in a number of ways:
dedicated devices (e.g., LANtern), devices with other tasks (e.g.,
hubs), and software running on a workstation (e.g., SGIs systems).
o Currently there are two SNMP products that seem to fall into the
Remote Lan Monitoring arena: Novell's LANtern and FTP's LanWatch.
Novell`s MIB is the only one available in the MIB directory on
venera.
o Spider, NAT, and Frontier have all announced products, or the
intention to produce a product. They each provided very
preliminary MIBs to the Working Group (hardcopy only).
o The remainder of the meeting was spent reviewing the Spider, NAT,
and Frontier MIBs with the idea of using these MIBs for development
of a common MIB the Working Group goal.
- Spider: Anne Ambler Of Spider
- While the SNMP philosophy is to reduce agent processing effort,
Spider chose to increase the complexity of the agent because it
is a dedicated agent.
- Spider has support for both Ethernet and TokenRing.
- Spider provides out-of-band support for probe access.
1
The Spider discussion was long and detailed as the document is one
hundred pages. Discussion was spent on the problems of packet capture,
packet return to the NMS, counter wrap around, and other issues. Steve
Waldbusser was asked to present some of these issues to the SNMP
Steering Group. Spider will post the MIB as soon as it is finalized.
NAT - Mike Erlinger: No one from NAT was at the IETF and thus only a
short summary of the available document was attempted.
Frontier - Steve Waldbusser: The discussion centered on filters and
packet capture. Steve believes that he has an algorithm that would
allow efficient transfer of bulk data from a probe to an NMS. He talked
about the algorithm and will present his findings via the mail list.
HP - Gary Ellis: A short discussion on the HP LanProbe and its
incorporation of SNMP was presented.
Wednesday Evening Meeting
Attendees represented CMU, Concord, Contel, David Systems,
Hewlett-Packard, MTI, and Spider Systems.
A ``segment'' is defined as ``everything a probe can see'' (this seemed
to be necessary to get some agreement on MIB group names).
It was reiterated that the SMI states that while implementation of a MIB
Group is optional, if that group is implemented, all objects in that
group are mandatory; also, a MIB should have only a single level of
groups, each of which contains objects (but not groups).
Traffic Generation was controversial; it was agreed that any support in
a standard MIB will be for simple capabilities (e.g., a single defined
packet that can be sent a number of times with a specified interframe
period); we will call the group SendPackets instead of Traffic
Generation to emphasize the simplicity.
The Administration groups will be difficult to define; although many of
the objects that might go here are vendor-specific, there is some subset
of objects that are common to all probes; we will need to identify this
``least common denominator'' subset for inclusion in the MIB.
It was agreed that it is a goal to get a proposed standard MIB out of
the March IETF; in support of this, the first RLAN MIB will be built to
reflect capabilities in currently available probes; later versions can
add features for which there are not currently any implementations.
The next meeting of the group will be during the first week of February;
notice will be sent to the rlanmib mailing list.
First Pass at an rlanmib MIB organization:
o MIB-groups:Ethernet Segment Counters
- Ethernet Segment Log
- Ethernet Station Counters
- Ethernet Segment Log
- Ethernet Traffic Matrix Counters
- Ethernet Traffic Matrix Log
2
- Thresholds Notifications
- Protocol Event Notifications
- Filters
- Triggers
- Packet Capture
- Test -- TDR
- Test -- Echo Protocols
- Test -- Traceroute
- Test -- SendPackets
- Administration -- Out of Band Access
- Administration -- Program Download
- Administration -- Trap Tables
- Administration -- Probe Status
- Administration -- Authentication
Steve Waldbusser will edit the Ethernet side of the document, Anne
Ambler will edit the Token Ring side and Mike Erlinger will coordinate
the document development.
The Chair wants to thank Gary Ellis and Sudhanshu Verma for providing
meeting notes.
Attendees
Anne Ambler anne@spider.co.uk
Karl Auerbach karl@eng.sun.com
Scott Bradner sob@harvard.edu
Ken Brinkerhoff
Theodore Brunner tob@thumper.bellcore.com
Jeffrey Case case@cs.utk.edu
James (Chuck) Davin jrd@ptt.lcs.mit.edu
Kurt Dobbins dobbins@ctron.com
Gary Ellis garye@hpspd.spd.hp.com
Fred Engel
Robert Enger enger@seka.scc.com
Mike Erlinger mike@mti.com
Richard Fox sytek!rfox@sun.com
Brian Handspicker bd@vines.enet.dec.com
Ken Jones uunet!konkord!ksj
Christopher Kolb kolb@psi.com
William Kutz Kutz@dockmaster.ncsc.mil
Mark Leon leon@nsipo.arc.nasa.gov
John Lunny jlunny@twg.com
Donna McMaster mcmaster@davidsys.com
Lynn Monsanto monsanto@sun.com
Bahaa Moukadam
David Perkins dave_perkins@3com.com
Robert Reschly reschly@brl.mil
Kary Robertson
3
Bill Rust wjr@ftp.com
Ray Samora rvs@proteon.com
Jon Saperia saperia@tcpjon.enet.dec.com
Lance Sprung
Ron Strich ssds!rons@uunet.uu.net
Glenn Trewitt trewitt@nsl.pa.dec.com
Sudhanshu Verma verma@hpindbu.cup.hp.com
David Waitzman djw@bbn.com
Steven Waldbusser waldbusser@andrew.cmu.edu
4