[IPFIX] DRAFT IPFIX minutes for Paris meeting

Hi all:

Here are my DRAFT minutes for the meeting - please send corrections,
suggestions for improvements, etc to me.

Cheers, Nevil

Minutes of the IPFIX meeting at IETF 83
About 27 people present
Scribes: Chris Innacio, Paul Aitken & Nevil Brownlee

Juergen opened the meeting, pointing out that
  - IPFIX Export per Stream is now RFC 6526.
  - The revised IPFIX MIB (5815bis) has been approved by IESG,
      this will clear the PSAMP MIB and IPFIX Configuration drafts.
  - Flow Selection Techniques is now in IETF Last Call.

Brian Trammell presented our 5101bis draft.  If we are to progress
the protocol to Internet Standard, we need to test interoperation
of all its 'must implement' features.  In the draft
  - template withdrawal is simplified.  For UDP, problems with this
    need to be fixed; Brian proposes a switch that would allow
    5101bis to interoperate with 5101 implementations.  Our AD
    will check whether that would be acceptable.
  - sections 8-10 have been edited to make them easier to understand.
    Please read these and comment on the list.
  - Security when using SCTP implementations needs more testing.
    This could be done using DTLS over SCTP (as in 5101), or using
    SCTP over DTLS (draft presented in tsvwg).  Brian proposed that
    users will mostly use SCTP in a Data Centre environment, where
    physical security is sufficient; over a wide area the data rates
    needed are low, TLS over TCP would be sufficient.
    We need to discuss this in more detail on the list.

Brian presented our 5102bis draft.
  - IE length limits: Brian proposes we use section 4.2 of our
    MIB Doctors draft, comments on the list please.
  - Allowable semantics for each type: there's an AI for WG to help
    IANA clean that up - it doesn't need to be in 5102bis.
    The chairs will discuss the cleanup with IANA.
  - We will run this draft's WGLC before IETF-84 in Vancouver.

Brian presented our IE Doctors (Guidelines for Information Elements)
draft, he believes it is ready for WGLC.
  - We will start a WGLC for this draft early in May, further
    comments on it before May are welcome.

Brian presented our Mediation Protocol draft, it currently has five
open issues.  There was considerable discussion, particularly of how
a mediator should handle re-ordering of IEs.
  - discussion needs to continue on the list

Brian presented the Aggregation draft.  The chairs reported that we
now know and understand the IPR related to this draft.  Our Area
Directors pointed out that since we now have IPR disclosures for it,
we need to do another WGLC; Nevil will start this next week.

Paul Aitken presented the Data Link Monitoring draft, item 7 in
our current charter. The next version will be published as a -00
WG draft.  Two issues were discussed:
  - Some implementors want to use these IEs; can we simply ask IANA to
    allocate them?  Our ADs pointed out that that reviews may affect
    this draft; we need agreement on them at WGLC.
  - When it's published, 00 draft will need reviews from the WG, we
    will also ask IEEE 802.1q to review it.

Paul Aitken presented the Exporting MIB Objects draft; this is item 5
in our current charter.
  - The draft proposes introducing a new Set ID.  Brian commented
    that in doing so, we should make it a little more general by
    including a Set ID code point (maintained by IANA), with 'MIB
    Object' as its first entry.  Brian will post to the list.
Paul asked "could this be a WG item?"  The chairs response was
"it needs a lot more discussion on the list."

Paul Aitken presented the 'Unobserved Fields' draft, which considers
how IPFIX should report on IE's for which no value is available.
Chris White (Riverbed Technology) agreed that this is definitely
needed.  Discussion will continue on the list.

Paul Aitken presented a few slides on "better IEs for TCP Window
size."  More discussion on the list is encouraged.

Hendrik Scholtz presented his 'VoIP Information elements in IPFIX'
draft.  There was considerable discussion:
  - Could be merged with Aamer Akhter's work with SIP.
  - Useful work, should be done, but IPFIX is not the right place
    to define new metrics - we simply report them!
  - Can be reviewed by PMOL directorate, but needs to be done
    in a current WG
The chairs closed the discussion, leaving it to be continued
on the list.

Andrew Yourtchenko presented his 'Cisco Specific Information Elements
reused in IPFIX' draft.
  - Benoit Claise (our AD) commented that this draft should not clash
    with the Data Link Monitoring draft.
  - Juergen Qittek suggested that this should be integrated into that
    draft.
Discussion will continue on the list.

The meeting finished at 1941.

-- 
---------------------------------------------------------------------
  Nevil Brownlee                    Computer Science Department | ITS
  Phone: +64 9 373 7599 x88941             The University of Auckland
  FAX: +64 9 373 7453   Private Bag 92019, Auckland 1142, New Zealand

_______________________________________________
IPFIX mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/ipfix