IETF-89 BMWG Meeting Notes
Friday March 7, 2014
1150-1320 Afternoon Session I
Richmond/Chelsea/Tower

Co-chairs Al Morton & Sarah Banks
Minutes taken by Vijay Gurbani & Bill Cerveny

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Session started at 11:50
0. Agenda bashed, no changes

1/2. WG Status: 
- SIP drafts back on the table.
- BGP dataplane adds more WG items
- Strong proposals for recharter
- VNF work starting to appear

3. SIP Device Benchmarking
- Vijay Gurbani Presenting
- Vijay indicates the belief to move this forward, asks the chairs for recommendations.
- Bill Cerveny & Barry Constantine to review document.
- Al suggests Vijay solicit feedback from someone within the SIP community; Vijay agrees.

4. Basic BGP Convergence Benchmarking Methodology
- Bhavani Parise presenting
- Current status of draft: new version addresses one comment from last IETF.
- 3-4 people in the room had read the draft.
- Have had a vendor run the tests in the draft, with input from said vendor included in the latest version.
- Authors asked for WGLC
- Chairs agree; will ask for WGLC and authors to solicit feedback from  an expert within the Routing area.

5. Milestone review and New Work Proposal
- Al reviewed the current milestones and displayed the proposal summary matrix.
- A few work items have lapsed due to personal/author issues.
- VNF and infrastructure drafts are appearing with proposals and they appear to be in scope of the charter.

6. Traffic Management Benchmarking
- Ram Krishnan presenting
- Ram provided a quick overview, and discussed the traffic shaper tests discussed in Vancouver.
- Saw some vendor variability in shaping traffic.
- Tested the shaper test method in a mobile operator lab. All metrics were viewed as advantageous by the operator.
- Q1 from Ilya V: In section 6.1, 6.2.2, what exactly are you trying to measure?
- A2 from Ram: Ram will clarify in the draft
- Q2 from Benoit: What is the relationship to LMAP
- A2 from Al: This is a lab measurement
- A2 from Ilya: But is it black box?
- A2 from Ram: It is supposed to be. You think itÕs a black box and it will behave the same, but it does not.
- A2 from Al: There was some discussion on the list.


7. IPv6 Neighbor Discovery
- Bill Cerveny presenting.
- Provides an overview of the draft, based on work suggested by Ron Bonica in IETF-85.
- Draft-04 is the latest.
- 3-4 people in the room read the draft.
- Bill reviews the problem the draft is trying to solve: scanning a v4 network is doable, with 510 hosts, but with IPv6 and the advent of 2^64 hosts, scanning could be attempted, but is difficult and not scalable. 
- This doc builds a network and runs tests to illustrate neighbor discovery.
- It creates measurements on the behavior of the DUTs.
- No comments from the room.

8a. NFV and Infrastructure Benchmarking Considerations
- Al Morton presenting
- New proposal for work
- Al discussed the NFV vision and why we should be interested in benchmarking this work.
- Comment from Curtis: Important consideration is functionality per watt, because cooling is an important issue. Those are all capital considerations and you are really constrained by how much power is being consumed.
- Comment from Dave McDyson seconded CurtisÕs feedback as the main objective to minimize

8b. Future Draft: Virtual Router Benchmarking
- Muhammed Durrani presenting
- Presented slides to cover problem statement of future draft
- 3 main topologies to benchmark: VNIV, SRIOV, and PCI Bypass.
- Need to ensure that the hardware is not the bottleneck, and presented a potential methodology to expose the software limits
- Q1 from Ilya V: It looks like a lot of this is recreating methodology created in an earlier BMWG doc and re-applying it to virtual routers
- A1 from Muhammed: There are some gaps in RFC 2544 that need to be filled in.
- A1 from Ram: Things are fundamentally different and you need to look at it in a different way
- Q2 from Ilya V: We need to first document what is different and then work on methodology . I donÕt yet know what is different.
-  A2 from Al: that is a good point
- Comment from David McDysan: Potential interaction between BMWG and NFV?

Closing Minutes
- Al: New charter paragraph presentations. He believes we have a good base charter; traffic management might need to be wordsmithed; it should be part of the next proposal charter.
- Al: IPv6 ND has support as well, as does the ISSU draft.
- Al: We thought about combining the DC and VNF/Infrastructure work, although Lucien A indicated he'd like them to remain separate. More discussion needed. 
- Q from Al: Any comments?
- A: No comments from the room
- Al: Chartering will hopefully be done in a few months
- Sarah will sent out new paragraphs to the list; letÕs have this discussion proactively and now.