IPFIX meeting at IETF76, Hiroshima
Wednesday, 11 November 09 at 1300-1500
Notes by Nevil Brownlee (attending via Jabber + audio stream)

WG status: Current (or close) milestones
   Oct 09: submit Mediators PS
   Jan 10: submit PSAMP MIB

IPFIX MIB (Benoit)
   ObservationDomainID added in ObservationPointTable
     (is read-only, will probably also be read-only in config rfc)
   -> Make new version, pass back to Dan/IESG

Config Data Model
   Changes in -04: more info in UML class diagram, PSAMP parameters,
     SCTP parameters, options templates, IDs (e.g. selectorID).
   No IPFIX changes needed now, Juergen proposes IPFIX WGLC,
     plus send it to netmod WG and ask for reviews.
   Agreed to start WGLC with -04 doc immediately.
     (Brian Trammell to post comments on Unordered SCTP to list)

SCTP export per-stream
   Dan comments that Transport WG adopted sctp-reset draft, which
   puts some pressure on IPFIX to keep it as a normative reference
   and change draft status to STD.
   Dan says 'post change to list, ask for comments (i.e. informal
   WGLC).  Juergen asks Benoit to post that email, will wait one
   week then pass -04 to Dan for IETF LC (needed for STD status).

Structured Data (SD) export
   Improvements over earlier draft.  One open issue, logical operations
   in semantic lists.  Needs further discussion on list, need a clear
   use case, and must decide whether structured data is the best place
   to support semantics.  Brian and Nevil will comment on list.
   Will need new version before WGLC.
   Comment: could one use SD to export detailed info about DPI,
            e.g. list of people running BitTorrent?
   Benoit: yes, would need agreement on IE values to describe applications.

Mediation Problem Statement
   Kobayashi-san presenting - summary, history.
   New version by end November, then submit to IESG (Juergen will
   ask for objections on list).

Mediation Framework
   Kobayashi-san again.  03->04 adds several new mediation processes,
   intermediate conversion, bidirectional flow composition.
   Needs to address some issues raised in mediation protocol draft.

Anonymisation Support
   00 draft published, have comments from list, will make a new version.
   Aim to have it ready to submit by Maastricht meeting.

New Drafts
   Data-link Layer Monitoring, Shingo Kashima.
     Proposes to make 3 new IEs?
     dataLinkFrameType - why not use ifType (from RMON) for values.
     Juergen says "since Ethernet is so common, this should be
     documented in an RFC."  Could be an Independent Submission.
     Dan agrees we need a more detailed draft, discussed on WG.
     Nevil will send some comments to the authors.

   Mediation Protocol (Benoit)
     A set of conventions to be used in defining the Mediation Framework.
     Makes the point that there are lots of things that aren't spelled
     out in the Framework.  What's the best way to specify all these?
     Does the framework depend on it, or could it be published first?
     Problem is, we don't want to delay the Framework unduly.
     -- lost audio at this point

- - - -