About the Working Group (GREEN)

Materials, Charter, Documents

GREEN Chairs & Area Director

AD: Mahesh Jethanandani
Chairs: Diego Lopez, Robert Wilton

Session link, minutes, jabber

IETF 124 GREEN Session Times

GREEN has two sessions:

1 hour, Monday November 3, 12:00-13:00 Montreal (UTC-5) 17:00-18:00
UTC
2 hours, Friday November 7, 14:30-16:30 Montreal (UTC-5) 19:30-21:30
UTC

GREEN Session 1 Agenda (12:00-13:00, Monday 20251103)

Intro by chairs

10 minutes
- NOTE WELL
- Explicit request for replies on the open adoption calls (use
  cases and terminology)

Base Drafts

Update on draft-bclp-green-terminology
- Presenter: Gen Chen (onsite)
- Rob: Is not too early to include intents as part of this
  document
  - Gen: The document does indeed need to be aligned wit the
    other documents
- Mahesh: A similar concern on the definition for SLEE. Be aware
  that we may not use the definitions immediately or in the near
  term.
  - Rob: This is a document that requires a long evolution along
    the work in the WG
- Jefferson: There is no clear classification of intents in any
  RFC
- Benoit: Agree on the comments from Rob and Mahesh. The terms in
  NMOP followed a two-phase approach (first, general terms, and
  then, collect terms in other documents)
  - Gen: A further review of terms would certainly be necessary
Update on draft-stephan-green-use-cases
- Presenter: Emile Stephan (onsite)
- Mahesh: Is there a plan to identify the use cases the WG is
  interested in working on?
  - Rob: Once we have the use cases document, we would need a
    consensus call on which ones to address. I hope most of them
    would be suitable to be addressed by the core model
  - Emile: We could do a cross-check with the termin
- Diego: There is a independent draft on ISAC, recently updated.
  How does it connect to the use case dociment?
  - Carlos: The draft on ISAC is going to converge with the use
    cases on ISAC matters
  - Muhammad: The update is for the YANG model, not the use
    case.
- Benoit: I see the use cases converging with the framework, and
  there should be a clear correlation with the core data model
  - Mahesh: This must be the case. Just to be clear that asking
    for what use-cases will be considered is to not limit the
    work, but for folks to understand which use-cases are being
    worked upon.
  - Rob: What we have to avoid is use-case or requirement
    creeping
Update on draft-belmq-green-framework
- Presenter: Benoit Claise (onsite)
- Rob: What you are proposing is to use sort of an EMAN
  translation into YANG as baseline?
  - Benoit: Yes
  - Benoit: We must consider, as a basic principle, monitoring
    before control, and there is a clear need of source
    identification
- Benoit: The draft on YANG provenance could be applied here. Four
  enclosing methods defined there.
  - Diego: For a new module, I would go for the first method,
    adding an optional leaf holding the proveannce signature
  - Benoit: What about datasheets, that are not YANG
  - Diego: we could consider a new draft on this. Let me talk
    with the implementers of our signing library
  - Rob: In any case, I believe most of those data should be
    tranformed into YANG
- Rob: I agree we need a framework document and that's why we
  started an adoption call. We need the community to help the
  discussion and break the chicken-and-egg dilemma.
- SoH on the adoption of draft-belmq-green-framework
  - Y: 23
  - N: 0
  - DK: 6
- This seems to be a clear indication the community is in favor of
  the adoption call

Open Discussion

No comments

GREEN Session 2 Agenda (14:30-16:30, Friday 20251107)

Intro by chairs

10 minutes
- NOTE WELL
- Chairs slides & status update
  - Reminder on the ongoing adoption polls
  - A new adoption poll, on the framework draft, will be
    launched after IETF124
- Scribe selection
- Agenda bashing

Models

Update on draft-petra-green-api
- 10 min
- Presenter: Luis Contreras (onsite)
- Rob: Are you looking for adoption?
  - Luis: Not yet. We could wait till the core model
    consolidates
  - Med: The work can progress, including the alignment with the
    core model
  - Luis/Rob: This is the intent
- Diego: Please elaborate on the choice of units
  - Luis: Based on the kind of use cases we are considering, but
    open to how the core model evolves, and willing to
    contribute to the discussion on units
Intro to draft-ahc-green-smartpdu-yang
- 10 min
- Presenter: Luis Contreras (onsite)
- Carlos: Out of the blue... Are you familiar with the "home
  assitant" smart home control? It is kind of similar, as it
  offers a centralized common layer interfacing with different
  smart home PDUs, allowing to monitor powe consumption, turn
  on/off and automate actions based on this.. There is a huge
  open-source community behind it so you could potentially reuse
  some of the code.
  - Luis: Happy to review it if you share a reference to this
- Nigel: Some of the design principles are related to inventory
  issues
  - Luis: Indeed. We are trying to consider this
- Altanai: Are you considering the collaboration with energy
  providers?
  - Luis: Not initially
- Qin: This looks more like an information model than a data
  model. Besides, PDU is a network device or more an inventory
  device?
  - Luis: Any help with the modeling will be welcome. Inventory
    issues should be considered
- Benoit: There is an issue with the identification of the
  individual elements, whether attached to a PDU or part of a
  device
  - Rob: This kind of devices must certainly be properly
    categorized
  - Benoit: Not sure if we need a dedicated YANG model for smart
    PDUs. The work is interesting in terms of the issues it
    addresses

New Proposals

Application-Layer Efficiency for Web Crawling
[draft-jurkovikj-collab-tunnel]
(https://datatracker.ietf.org/doc/draft-jurkovikj-collab-tunnel/)
- 10 min
- Presenter: Antun Jurkovic (remote)
- Rob: This work looks interesting, and is linked to GREEN through
  the definition of metrics. You could be interested in reviewing
  metrics proposals and comment on them. But the full-stack goals
  seem far from the current scope and the WG expertise. We plan to
  reach the ADs to see where this work would belong.

Progress on the GREEN Base Model (80 mins)

DTL update
Presenter: Jan Linblad
- Med: To me, the key goals are in work items 3 and 4: data on
  energy consumption
- Diego: In my view, W-I 1 and 2 are about data collection. W-I 3
  and 4 are about control
- Jan: W-I 1 tries to go beyond home-cooked solutions for
  reporting network-wide metrics. W-I 2 needs to consider W-I 1 to
  make any sense of individual measurements
- Benoit: There are 3 YANG modules to deal with: (1) Power and
  Energy monitoring model (Device); (2) Power and Energy control
  model (Device); (3) Collection and Aggregation model
  (Controller)
- Jan: I think there is also a fourth area, Power and Energy
  control on the Controller level. But that's for later for sure.
Green Device Model: Benoit Claise
- Jan: I believe this EMAN work goes in the right direction, but I
  think we should address any kind of devices, whatever the
  interfaces they use
- Med: This goes in the right direction, but I am not sure EMAN
  covers all the cases we need to consider, especially in what
  relates to leveraging the inventory part
- Benoit: I do not want to bring all EMAN, but I believe the EMAN
  proposal will be useful for identifying relevant issues to be
  addressed
- Marisol: Many of the things presented here rely on hardware, and
  IVY is considering these aspects
- Gen: It is not only about EMAN, but considering other results
  (ITU-T, ETSI...)
Discussion on results and further work
- Rob: There is a clear demand and a clear path, but we must
  consider (more or less) legacy devices, and that requires the
  W-I 1 mentioned by Jan. I believe work should advance in both
  directions
- Benoit: Working in both directions would be possible
- Thomas: This would be strategic for network operations. I would
  really support this work
- Marisol: To me PDUs can be considered part of a network device,
  not sure if defined as a component, something to discuss
  probably, but for sure to be considered as part of a network
  device.
- Benoît: Exactly Marisol!
- Thomas: +1 on the relationship between power measurement and
  having a relationship. Smart PDU is a good example.
- Alexandr: Just a thought: you can probably make it a bit more
  efficient by using CBOR instead of JSON. It would be a
  low-hanging fruit.
- Benoît: CBOR. sure. Let's focus first schema
- Camilo: +1 in start work right now on practical power
  monitoring, not sure if making the energy yang the point to
  match entity and ivy/hardware though, that sounds better for an
  augmentation (from our 3 vendors we collect inventory of, only
  Cisco uses entity well, the other two j and n, have their own
  mivs)
- Marisol Palmero: I just want to make my point clear, I am in
  favour to go with the proposal presented by Benoit and Gen, but
  also considering the practice that we put together in POWEFF,
  what it was strong dependency on the HW / SW inventory (work
  linked to IVY).
- Rob: I would like to ask Jan, Gen, Marisol, Jan and you to
  coordinating your work on both W-I 1 and 2.
- SoH: "Should we be focussing the work on the concrete use cases
  (rather than abstract models)?"
  - Y:19
  - N:1
  - NOP:3
- SoH: "Should we be focus the initial work only on a device
  model?"
  - Y:5
  - N:10
  - NOP:5
- SoH: "Should we focus on using the EMAN MIB (converted to YANG)
  as a starting point for the device model?"
  - Y:18
  - N:0
  - NOP:3

The chairs expect the authors of both models (device and controllers)to
start working on them. We will organize DTL bi-weekly calls to discuss
the results and make the consensus on the models progress.