GREEN WG --- 4th November 2024 9:30 (GMT)
Chairs: Diego Lopez, Robert Wilton
Notetaker(s): Joe Clarke

Intro by chairs (10 mins)

1. Welcome and Note well
   • Intro to the new WG by Mahesh Jethanandani, designated AD
   • Chair reminder on goals and priorities

Mahesh:

• Thank you to all involved in charter discussion and formation of the
  WG.

• Goals:

  • Speaking on some of the goals.

  • Requirements document

    • Should be a living document, documenting requirements, and
      specifically the use cases.
    • Will evolve as the WG evolves with its charter
    • Hopefully requirements never becomes an RFC, but will leave
      it up to the WG
    • Look at use cases first before stating requirements;
      validate each use case for "today's" environment

  • Charter is not cast in stone

  • There could be a charter update for future work after completing
    the work currently in the charter.

• Sustainability is an important topic, but not currently in scope of
  the WG

• Sustainability is now part of the IRTF in SUSTAIN RG, charter
  discussion happening now.

Chairs (Rob):

• More focus on energy efficiency than sustainability
• We need help writing liaison statements to other groups; please
  reach out to chairs if you want to help
• We want to work towards adopting documents per the charter

2. Requirements(10mins)

draft-stephan-green-ucs-and-reqs (10 min)

Speaker: Luis Contreras

Q&A:

Carlos Bernardos: Focus on use cases before
Document has useful content, but kind of hard to follow because it
doesn't lead with use cases
Focus only on use cases at this point, then requirements, then
framework

Ali Rezaki: Looks for use cases then drive the technology

Luis: We need to rationalize the approach between use case focus and the
technology requirements

3. Terminology(25mins)

draft-bclp-green-terminology (10 min)

Speaker: Mohamed Boucadair

Q&A:

Rob: The set of terms should be very narrow; reuse makes a lot of sense;
we want to consolidate the terms documents into one doc

Med: Okay

draft-pignataro-green-enviro-sust-terminology (10 min)

Speaker: Ali Rezaki

Q&A:

Rob: This document's scope is much broader than the scope of GREEN; some
useful content to feed into terminology
But we need to find a place for the wider focus; maybe SUSTAIN or
eIMPACT
This is too broad (at the moment) to adopt in this WG
Should be narrowed down

Ali: We can look to see what terms we might want to extract from this,
but the scope is definitely broader
We can look at SUSTAIN RG for this, too

General comments/observations on terminology drafts (5 min)

Rob: We should work to adopt a single document for terminology. We can
hold a virtual interim perhaps; comments?

Carlos Jesus Bernardos: An interim might be useful; we need to sync on
terminology

Diego: We have tried to identify the foundations of what we should be
doing; terminology and metrics are connected and we need to align these
foundations

Rob: Discuss on list and focus on one particular terminology draft as a
starting point and whether we do a VI to settle on this

(Order of procedure: Datatracker is having issues so we are stalling on
a consensus call on terminology)

Diego: We have seen two term drafts; show of hands on this first
terminology draft for a WG starting point:
total participants: 125
1st doc yes: 37
1st doc no: 4
1st doc no opinion: 26
[Show of hands in room via Rob]: Strong indication that this is a good
starting point

Diego: Can conclude that the first draft is suitable for a starting
point on terminology; majority of opinions in favor

total participants: 137
2nd doc yes: 9
2nd doc no: 27
2nd doc no opinion: 30

Rob: (to confirm on list) Seems like first doc is a good starting point,
but we can pull in content from the second doc

4. Metrics(25mins)

draft-bogdanovic-green-energy-metrics (10 min)

Speaker: Dean Bogdanovic

Q&A:

Jing Wang: How can metrics... (moved to chat due to audio issue)

Ali Rezaki: Have you looked at what other SDOs have done with metrics
defs?

Dean: Looked at ETSI and NIST; they are mostly in line with what we're
suggesting

Ali: 3GPP and mobility providers are coming out with their own thoughts
on energy

Dean: Looking at 3GPP to provide radio specs and should not define
end-to-end solutions

Phillip Hallam-Baker: We're looking at power consumption and efficiency;
acceptable with coal and gas; many current sources are now
non-dispatchable; we need to focus on when power is being used vs. the
power being used

Dean: Can take power into account for traffic engineering

Phillip: Power dispatchability will be an important metric

Dean: We need to get to a common set of data/metrics so we can make
clear decisions

draft-cx-green-green-metrics (10 min)

Speaker: Alex Clemm

Q&A:

Gen Chen: Please give a clarification on the difference between flow and
path metrics

Alex: Flow is about the energy consumed by a specific service or flow
instance (e.g., in the flow record); the share or energy/carbon that can
be attributed to the flow
The path helps you make better decisions about with paths/routes to
select to achieve a more energy-optimal path through the network (e.g.,
running over greener eqduipment)

General comments/observations on metrics drafts (5 min)

Rob: Running short on time; move on to models; we might have time to
circle back

5. Models(Reportingandoptimizations, differentarchitectures)(50mins)

draft-li-green-power (10 min)

Speaker: Tony Li

Q&A:

Benoit Claise: Relationship, we need other types of relationships (e.g.,
for PoE); we should focus on use cases so that we can, for example,
determine if PoE is in

Tony: Agreed PoE should be in — we are after a total power use across
the network, looking to include which AP is on which port

Arashmid Akhavain: Are we going to get some requirements on bootstrap
time for the components (how long they take to transition)

Tony: Keep it simple for now; though we could add a leaf for this

Eve Schooler: You mention the simple power usage metric; if you can't
measure that, you can insert from the datasheet; will you allow one to
reflect whether this is measured or simply reflecting the datasheet?

Tony: Not now; if the WG feels we need a new attribute, we can add it

Rob: We're too early to do adoption polls now; like the simple model;
good to do use cases but don't want to spend lots of time doing use
cases, terms, etc. in serial; would like to see work happen in parallel

Benoit: A simple use case like whether we do PoE or not will help us
determine if we need to model this; thought of some of these use cases
in EMAN; the relationship bit has a lot of architectural ramifications

draft-cwbgp-green-energy-saving-management (10 min)

Speaker: Qin Wu

Q&A:

Dean Bogdanovic: Go to slide 7...Who is the customer in your case here?

Qin: Could be the network operator who builds this system or the user

Dean: Operator or user? Becuase the operator sees the network as a whole
domain and can make the right energy decisions based on the whole
entity; else you might break energy contracts other users are not aware
of; be very specific as to who is controlling the nergy usage within the
network
This is trying to do too much to start with; happier to see a basic
approach rather than do a top-down approach; do simple things; the IETF
doesn't have people that understand energy distribution; if we learn the
basics it will be easier to build simpler systems from the bottom up

Qin: The network operator is trying to improve energy use and they have
more control; this system could be broken down to single domain from
multi-domain; we didn't want to expand this...this is a generic
framework

draft-xiong-green-host-power-monitoring-yang (10 min)

Speaker: Yanxia Tan

Q&A:

Dean Bogdanovic: Want a better expectation as to what a layer is
How is this YANG model different from Tony's? This one seems a bit more
narrow.

Yanxia: This model is concerned more with the measure of energy usage in
this hierarchical model whereas Tony is focused on control of energy
usage

Mahesh Jethandani: To the WG: what are we collecting? Are we collecting
host power consumption, facility power consumption?
What is the focus of collection metrics in GREEN?

Rob: We need to refine these different models and how they fit together.

Mahesh: We seem to be broadening scope

Diego: What do we mean when we talk about network; during charter time
it was clearer, but now things have popped up

Dean: Several Tier 1 operators have facilities that use more energy than
the network equipment inside

Mahesh: When we talk about GREEN in the context of networking what
equipment are we including (HVAC?)

Diego: We sould qualify this, but we need to move on

draft-petra-green-api (10 min)

Speaker: Alberto Rodriguez-Natal

Q&A:

Dean Bogdanovic: Fully support providing APIs that can help traffic
engineering; please keep it simple; don't talk about carbon
optimization; we have no way to know the carbon intensity of the power
at the moment; we change carbon footprints of power on the fly; some of
this happens out of band in analog ways (like phone calls); it will not
be so precise to determine the green-ness of energy

Srihari Sangli: Paths change dynamically; is it even feasible to compute
power consumption by path and keep it accurate?

Alberto: We hope so, but this is to be explored

On the POWEFF and philatelist proposals (5 min)

Speaker: Jan Lindblad

Q&A:

Rob: This is interesting work; POWEFF is definitely in scope, but
philatelist might be on the periphary of what is in scope

General comments/observations on models (5 min)

Rob: Quick show of hands on virtual interims (who would be interested in
participating) to refine some of these areas

total participants: 131
yes interested: 35
no not interested: 4
no opinion: 7

Rob: Very hopeful we can work toward consensus and get stuff done
EOM