Minutes of the IPFIX meeting at IETF 81
About 25 people present
Scribes: Andrew Feren, Chris Innacio & Nevil Brownlee

Juergen Quittek opened the meeting, pointing out that all but two
drafts from our current charter have now been approved.  Two remain:
 - Flow Selection (Benoit to check IE values, Nevil to submit write-up)
 - PSAMP MIB
We expect that these two will be finished shortly.

Juergen presented the PSAMP MIB.  The IPFIX MIB (RFC 5815) was
intended to use a registry for ipfixSelectorFunctions, but instead
IANA created an IPFIX SELECTOR MIB, which would need to be changed
each time we add a new selector function.  After some discussion, we
decided to stop maintaining the IPFIX SELECTOR MIB and create new OID
registry for ipfixSelectorFunctions.  That will require changes to the
PSAMP MIB and to RFC 5815; since that requires AD approval, these will
be new WG work items.

Benoit Claise presented "Exporting MIB Variables" (now at -02),
pointing out that this has issues with exporting variables from
different SNMP contexts, but that would clearly be useful.  Clear
consensus as new WG item.

Benoit presented the "IPFIX Mediation Protocol" draft (now at -04), a
method of exporting information about mediation processes.  Clear
consensus as new WG item.

Brian Trammell presented the "IE Doctors" draft (now at -01).  This
documents the process for establishing new IPFIX Information Elements.
Clear consensus as new WG item.

Brian presented the "IPFIX Aggregation" draft (now at -03), which
would establish an Intermediate Process to perform spatial and
temporal aggregation in IPFIX Mediators.  Clear consensus as new WG
item.

We have one other draft that was presented in earlier meetings - "Link
Layer IEs."  We discussed whether this could be a "trial run for the
IE Doctors," or a WG item.  We reached consensus for it as a WG item,
with the proviso that it will need views from Layer 2 domain experts,
e.g. IEEE 802.1.

Nevil led a discussion on "updating the IPFIX base standards," at
least 5101 and 5102.  These would include changes made by errata, and
some editorial work to clarify text in the light of experience.  The
DTLS question was raised - should this be included in a new version of
5153 (Implementation Guidelines)?  Consensus reached to update 5101
and 5102, and to EITHER revise 5153 OR to create a new document from
the DTLS draft [the latter to be decided on the mailing list].

Andrew Yourtchenko presented a new draft that will document the Cisco
IEs, 1-127.  This could be an Independent Submission, or it could
remain as an Internet Draft, updated over time then allowed to expire.
Benoit explained that it will make information about these NetFlow v9
IEs available, thus easing the transition to IPFIX. [In other words,
the range 1-127 will eventually be filled in, and the distinction
between below and above 127 will then disappear. Some of the IEs will
be deprecated, some will be reused].

The WG chairs will write a new charter covering all the 'consensus'
drafts above and their milestones, and send it to the list.  Once
consensus is reached on the list, we will our Area Director to take it
to IESG for approval.

The meeting finished at 1435.

- - - - -

One-para Summary

The IPFIX WG met on Wednesday afternoon.  The WG has two current
work items, these should be completed by 1 September.  We discussed
eight individual drafts, reaching consensus that - bearing in mind
that five of these have had several revisions - we wish to revise
our charter to include all of them.

- - - - -