Minutes for IPPM at IETF-90
minutes-90-ippm-3

IPPM Minutes, IETF 90, Toronto - Tuesday, 16:40 - 18:40
=======================================================
Many thanks to Sarah Banks for taking minutes!


Comments on the Rate Problem WGLC
---------------------------------

The final question in WGLC is whether to require asymmetric packet
size support in protocols for rate measurement.

- Al Morton: all tools today send symmetrical packets; if you don't want to
solve this problem, you can choose to ignore it.
- Al/Yaakov Stein, there was clarification last year around this requirement
that it was symmetrical, and asymmetrical.
- Al: this is a requirement for test protocol dev.
- Greg Mirsky: I wouldn't support buying a product that wasn't asymmetrical, but
I see that some folks would want it this way, so I support SHOULD as
a requirement.
- Al quotes Shakespeare
- Greg: How about MUST on rate, SHOULD on packet size? 
(general agreement in the room)
- Al: then it's SHOULD for BOTH symmetrical and asymmetrical size
- Brian Trammell: there should be a statement of applicability on the SHOULD 

Call for adoption on checksum trailer draft
-------------------------------------------

Hum of support with no objection for adoption, but need to take 
the question back to the list.


Registry for performance metrics
-------------------------------
Marcelo Bagnulo presenting

- Presentation
- How do you determine the process for defining the new sub registry?
- Proposing that we revert back to the flat registry - for some of the
columns, you'd put "not applicable" - this allows you to skip the problem
of categorizing things.
    * Brian, no problem, 
    * Al, no problem, but he liked the categories
- Names, identifiers, and URIs
    * Barbara Stark clarifies the need,
    * Debate/discussion around roles, assignments of the role 
      from the controller to the measurement agent
    * Brian agrees with Al, this belongs with measurement
    * Brian believes this falls between the cracks of LMAP and IPPM;
    * Marcelo believes that LMAP protocol spec will have to cover its 
      own parameters


Model based metrics
-------------------
Matt Mathis presenting

- Presentation
- Outline of the problem
- Outline of some of the problems with some solutions

- Brian (chair hat off) thinks this is a good stress test to the registry
- Matt to provide a new milestone date/update


RFC 2679-bis and RFC 2680-bis
-----------------------------
Al Morton Presenting

- Presentation
- Former authors unreachable
- WG call for adoption by Brian - hum to adopt,
  together with previous support on the list. 
  Brian calls the docs together as adopted.


DSCP/ECN Monitoring in TWAMP
----------------------------
Greg Mirsky presenting

- Discussion about deployment stats for ECN
- Adoption question will be taken to the list
- 4 people have indicated they'll read the draft


IKEv2-based shared secret key for O/TWAMP
-----------------------------------------
Kostas Pentikousis presenting

- Active discussion on the list to respond to WGLC
- Brian to review the diffs, perhaps to send the doc through another WGLC.


Use cases for passive measurement in wireless networks
------------------------------------------------------
Lingli Deng presenting

- Presentation
- Feedback from audience - one request for more detail around the
use cases
- Active vs Passive discussion was shut down by the chair :)


Passive Framework draft-zheng
-----------------------------
Nalini Elkins presenting

- Presentation - mostly about active vs passive
- Q from audience: What's the purpose of this doc? 
- (Al + Matt) - if we don't tackle the hybrid model, we might not be 
  doing anything useful.
- Nalini: We purposely stayed away from the hybrid model.
- Al: Keep working on the individual drafts, they'll inform the larger
  framework


IP Flow Performance Measurement (IPFPM) Framework & Report
----------------------------------------------------------
Vero Zheng presenting

- Presentation
- Wants more feedback on the draft, and wants to call for WG Adoption
between IETF-90 and 91.
- Brian: work going on in LMAP that's possibly applicable to this; it'd be
nice if what's being done here were compatible with the work being done in
LMAP
- Greg: its broadly in line
- Al: Anything on passive measurement needs a very thorough privacy 
  considerations sections, or it won't make it through the IESG

- [Vero will not be the last presenter in Honolulu by order of the chairs]