24 NMRG Meeting, March 14, 2008
		 71st IETF Meeting, Philadelphia, USA

Participants:

  The meeting was attended by about 20 people. The names were recorded
  on the rooster which went to the IETF secretariat.

Agenda:

  09:00	Agenda bashing and administrivia
	Juergen Schoenwaelder (Jacobs University)
  09:05 Status of the "SNMP Traffic Measurements" draft
	Bert Wijnen
  09:10 Report from the IAB review
	Juergen Schoenwaelder (Jacobs University)
  09:15 Report from the CANMOD BOF
	Randy Presuhn
  09:25 Research classification report
	Georgios Karagiannis (University of Twente)
  09:35 Discussion of the "SNMP Trace Analysis Definitions" draft
	Gijs van den Broek (University of Twente)
  11:10 Future of the "SNMP Trace Analysis Definitions" draft
	Juergen Schoenwaelder (Jacobs University)
  11:15 AOB
	Everybody
  11:30 End of the meeting

General Information:

  The Network Management Research Group) NMRG session on Friday, March
  14 was chaired by Juergen Schoenwaelder. Randy Presuhn took notes for
  these minutes and Wes Hardacker listened on the Jabber channel.

  The meeting materials were posted on the NMRG website at
  http://www.ibr.cs.tu-bs.de/projects/nmrg/ rather than the IETF
  meeting materials site.

Status of the "SNMP Traffic Measurements" draft:

  Bert Wijnen summarized the status of the "SNMP Traffic Measurements"
  draft.  The previous round of comments has been responded to.  The
  next step is to formally ask the IRTF chair to ask for review in the
  IRSG.  It was noted that David Harrington's SMI datatypes document
  which is currently under discussion in the Operations and Management
  Area is not quite in synch with one that Juergen Schoenwaelder had
  produced.  The main issue is that the regular expression for IP
  addresses in the ops area draft is stricter.  Juergen thinks it
  makes sense to align with the stricter form.

Report from the IAB review:

  Juergen Schoenwaelder presented a report on the IAB review.  He
  noted that there is a desire to increase participation by operators.
  To that end, co-location of meetings with nanog might help.  Another
  concern is that some of the NMRG's work is showing up in academic
  publications rather than RFCs, limiting the visibility of the work.
  Consequently there is a desire to consider republication of some
  papers as RFCs.  Finally, Juergen noted that he hopes to step down
  as chair, and that consequently there is a need for new co-chairs.

Report from the CANMOD BOF:

  Randy Presuhn gave a report on the CANMOD BOF and its aftermath.

Research classification report:

  Georgios Karagiannis gave a presentation on the work on a network
  management research classification.  The goal is to define a
  taxonomy for organizing network and systems management research
  topics.  The plan is to incorporate taxonomy in the Journal and
  Event Management System as well as future conferences (NOMS, IM,
  DSOM) and use it to classify research efforts in IRTF.
  Subsequently, the intent is to collect results and review the
  taxonomy.

  There were numerous comments on this presentation.  It was observed
  that taxonomies are rarely perfect, but can be useful nonetheless.
  It was suggested that the access segment (e.g.  core, non-core,
  metro) might be another dimension to consider.  It was clarified
  that the "p2p" category was about peer-to-peer from a system
  perspective, rather than the protocol design perspective.  It was
  also noted that the current hierarchy doesn't appear to have a niche
  for database and storage management issues.  Lastly, there was a
  concern that a hierarchy might not be the best way of looking at a
  multi-dimensionsal classification scheme.  Work was summarized as
  being useful, and that the challenge is to limit its size.

Discussion of the "SNMP Trace Analysis Definitions" draft:

  Gijs van den Broek presented the work on SNMP trace analysis
  definitions.  He emphasized that feedback is desired.  There were
  several questions about how the definitions work with some of the
  less common things that have been seen in management traffic, such
  as responses appearing on a different interface from the
  corresponding request, interleaved table walks, and how the slice
  definition excludes event-directed polling.

  The chair then asked the group about the future of the trace
  analysis definitions.  Several speakers liked work, but were not
  willing to actually do work on this topic.  Consequently, the chair
  will take this discussion back to the mailing list.

AOB:

  There were no open mike issues.

The meeting adjourned at 10:25.