[IPFIX] DRAFT minutes for IPFIX meeting in Orlando

[Nevil Brownlee <[e-mail]>]

Hi all:

Here are the DRAFT minutes, please send me any corrections or improvements.

Cheers, Nevil


One-para summary:  IPFIX WG meeting in Orlando

Our three current chartered drafts, Link layer IEs, MIB Variable
Export and Mediation Protocol are close to completion.  Several
possible new work items were discussed:
+ Extended Field Specifier Fields (EFSF) as a way to handle
    'decorators' in Information Elements, e.g. the layer of an
    mplsLabelSection.
+ A textual representation for Information Elements, making
   it easier to use the IPFIX Information Model for flows in
   text-based environments.
+ Mechanisms for moving Private Enterprise IEs into the IANA
   IPFIX IE Registry, so that Collectors can recognise them
   more easily.
These will be discussed further on the IPFIX list.


Minutes of the IPFIX meeting at IETF 86 in Orlando
About 16 people present, plus 4 on jabber
Scribes: Chris Inacio, Nevil Brownlee
Meeting chair: Nevil Brownlee

Nevil opened the meeting, and gave the IPFIX status update:
- Flow Selection Techniques now in AD Follow-up,
   -14 version published 10 Mar 13
- IE Doctors
   Approved 4 Dec 12, in RFC Editor queue, MISSREF on 5101bis
   Initial ie-doctors team is Brian Trammell, Paul Aitken,
     Juergen Quittek and Nevil Brownlee
- IPFIX Information Model, 5102bis Approved 13 Feb 13,
   MISSREF on 5101bis
- IPFIX Protocol, 5101bis
   Submitted to IESG, 27 Feb 13

Paul Aitken presented (via Meetecho) the Link Layer IEs draft:
- Revised taking account of IEEE 802 feedback,
   second WGLC finished 10 Mar 13
- Pat Thaler (IEEE 802 chair) reviewed the draft during WGLC.
   She repeated her comments for the meeting:
   . E-TAGs are internal to 802.1 bridges, therefore they
     should not be included in this draft (or in an IPFIX IE).
   . Similarly, the S-Channel is used for the links between
     bridges; this doesn't seem to be too helpful.
- Authors will publish another revision, the draft should then
   be ready to submit.  Pat agreed to review the new version.

Benoit Claise asked how the IE-Doctors will work?  IANA need to be
asked for advice during WGLC so that they can ask the Designated
Experts; that way IESG will know that any IANA problems are resolved
before a draft is submitted to IESG.

Paul presented the MIB Variable Export draft:
- Issues raised by version -02 have been resolved, but there are
   still many issues to be resolved.
- The draft proposes to introduce Extended Field Specifier Fields
   (EFSF), with 'decorators' that could, for example, be used to
   handle Unobserved Fields.
- EFSF generated a long discussion - it could be used for many
   other things; this concept could provide a new, elegant and concise
   way of handling attributes of particular IEs.
- Consensus in the meeting was that
   . EFSF is a new direction for IPFIX - it's really IPFIXv2.
   . We need to finish the MIB Variable Export draft as soon as possible.
   . Paul will remove it from this draft, making sure the draft
     will still work in IPFIXv2.
   . The WG should adopt this as a work item.  That will require a
     new charter; meanwhile work on it can proceed on the IPFIX list.
- Please send your comments on this to the list!

Brian Trammell presented the IPFIX Mediation Protocol draft:
- -03 and -04 revisions have been published since IETF 84 in Vancouver.
- All open issues now have a proposed solution.
- Should be ready for WGLC after one more round of reviews;
   Paul will review it (Brian will review MIB Export).

Ram Krishnan presented his draft on Flow-Aware Packet Sampling
- A vigorous discussion focused on two aspects:
   . There's a lot of IPR for sampling techniques, need to make sureIPFIX 
one-para summary

     that any IPR related to this draft is declared.
   . Although it's an interesting technique, it doesn't put forward
     a clear use case, so it doesn't seem to require work within the
     IPFIX WG.
- Consensus in the room was that it needs more work to make it clear
   why IPFIX should work on it.

Brian Trammell presented his draft on a Textual Representations for
     Abstract Data Types:
- IPFIX is not text-based, this would provide a way to describe
   Information Elements by name, much like YAML or JSON.
- This seems an interesting idea; please continue discussing it
   on the IPFIX list.

Chris Inacio presented his draft on Private Enterprise Information
     Elements (PENIE)
- This proposes a simple mechanism for setting up IE Registries
- No consensus on it was reached.

Paul presented his Equivalent Information Elements draft:
- This draft addresses the question "how to change a PENIE into
   an IANA-registered IE?
- It proposes that Exporters should send Equivalence Information
   to Collectors.
- Brian and Andrew Feren commented that we now have two solutions
   to this problem, 'PENIE' and 'Equivalent IEs.'  Perhaps we should
   implement both?
- Benoit suggested asking Collector implementors which would work
   best for them.  We need a brief summary of the pros and cons of
   each; Paul's slides do that for 'Equivalent IEs,' Chris will
   make a similar list for PENIE.
- This could be an item in a new WG charter.

There was no Other Business.

The meeting finished at 1105

-- 
---------------------------------------------------------------------
  Nevil Brownlee                    Computer Science Department | ITS
  Phone: [phone] x88941             The University of Auckland
  FAX: [phone]   Private Bag 92019, Auckland 1142, New Zealand