GREEN

Summary

This session of the GREEN Working Group covered administrative items, updates on the core foundational drafts, and fostered discussion on the path forward. Updates were provided on the Terminology, Use Cases, and Framework drafts, with an emphasis on aligning definitions and concepts across these documents and with future data models. A strong call for active participation in ongoing adoption polls was reiterated. The session culminated in a decision to launch an adoption poll for the Framework document, acknowledging the current impasse between framework and data model development and the need for concrete starting points.

Key Discussion Points

Administrative Items and Working Group Status
- The chairs (Rob and Diego) opened the session, highlighting the IETF Notewell, conduct policies, and the working group's professional expectations.
- The work is structured into short-term deliverables (Use Cases, Terminology, Framework) and long-term data modeling.
- Liaison statements from previous IETF meetings still require attention; volunteers are sought to help respond.
- Adoption Calls: Reminder that 3-week adoption polls are currently active for the Terminology and Use Cases drafts, extended due to the holiday period. Members were strongly urged to provide feedback (even simple "I support adoption" or "I do not support adoption" with reasons) on the mailing list to gauge interest and progress.
- Design-team-light meetings for modeling have started, with three held in October, and will continue post-meeting to progress common understanding.
Terminology for Energy Efficiency Network Management (draft-ietf-green-energy-terminology)
- Presented by Jentchen.
- The latest version is 05, with Sheddish added as a contributor.
- New definitions include "service level energy efficiency" and "energy-aware intent." Updated definitions clarify "device level" and "component level" energy efficiency.
- Discussion:
  - Rob: Questioned if "energy-aware intent" and "service level energy efficiency" might be too early for inclusion, suggesting they should align with the framework and data model development. He proposed keeping the document as a living WG document, not rushing to RFC, to allow for evolution and alignment.
  - Mahesh: Agreed with Rob regarding the potential for these terms to be premature, noting charter alignment and the need for data model discussions to clarify scope.
  - Speaker (unidentified): Suggested referencing RFC 936 for intent definitions and clarifying the scope of intent (network-wide vs. specific parts).
  - Benoit Awas: Reiterated the importance of terms being used in other stable documents (framework, data models) before inclusion. Suggested a two-phase approach: include terms, check usage, then prune or refine. He proposed an appendix for potential future terms or storing them as GitHub issues.
GREEN Use Cases (draft-ietf-green-use-cases)
- Presented by Emil (noted a mistake in the slide title, it was for Use Cases, not Framework).
- The draft started in July 2024 and currently contains 15+1 use cases, including a recent addition by Xinu on "cross-level energy efficiency."
- A table mapping use cases to framework interfaces has been pushed to the framework draft, showing which interfaces each use case utilizes.
- Next Steps for Draft: Cross-validation to reduce overlaps (possibly using MPR 16 rules), continued documentation of the interface table, and checking relations with data modeling and terminology.
- Action: Address Mahesh's directions to finalize the Isaac use case and add Xinu to the author list.
- Discussion:
  - Mahesh: Asked if there's an effort to identify which use cases the WG specifically wants to work on.
  - Rob: Clarified that the plan is to adopt the document, then do a consensus call on the included use cases to establish the WG's focus. The document is intended to be a living WG record, not necessarily an RFC, to allow for new use cases as long as they align with the charter and bring new requirements.
  - Diego: Clarified that the Isaac use case in this document will converge with the YANG model for Isaac (draft-ietf-green-yang-isaac).
  - Benoit Awas: Emphasized that the main point is the ability to fulfill all use cases with the defined framework interfaces, supporting the idea that many use cases might use the same underlying data model.
A Framework for Energy Management in Networks (draft-ietf-green-framework)
- Presented by Benoit Awas.
- The document has seen three new revisions. A key conclusion is that most of the identified use cases can be solved with the building blocks in the framework.
- Key "green" specific concepts include energy objects, power interfaces, and metering relationships.
- A new G-interface (API service interface) has been added to support applications like PETR (Power Energy Traffic Ratio API).
- Proposal for Data Modeling: Benoit suggested using the existing Eman Working Group's MIP for power information (e.g., nameplate, unit multiplier, accuracy, current type) as a starting point for Yang modeling, rather than starting from scratch.
- Discussion on Source ID and Yang Provenance:
  - Need for unique IDs (like EUID from RFC 4133 Entity MIB v4) for reporting.
  - Discussion on using draft-ietf-netmod-yang-provenance for ensuring data origin (e.g., from routers vs. data sheets). Diego elaborated on the four options for adding signatures, with the simplest for new modules being adding a leaf element. The challenge of signing data sheet information was raised for further study.
- Open Issues: Discovering diverse capabilities (what "sleep" or "hibernate" means in different contexts), mapping intent to settings, transition handling, and ensuring safety.
- Benoit's Concern: Highlighted that the framework development and modeling efforts are in a "waiting mode," hindering progress. He emphasized that both are required and asked for adoption to break this impasse.
- Rob's Response: Acknowledged the concern and took some responsibility for slow organization. He emphasized the need for active participant engagement in the design-team-light meetings, as chairs are only facilitators. He suggested a GitHub-based approach for the model to encourage concrete contributions.
- Diego's Response: Agreed on the need for community feedback to delimit the core model and framework, especially regarding potential overlaps with other framework models (e.g., Jan's model). He urged comments on the adoption call.
- Adoption Poll: A show of hands indicated strong support for adopting the framework document.

Decisions and Action Items

Decision: The GREEN Working Group chairs will launch an adoption poll for the draft-ietf-green-framework document on the mailing list. (Confirmed by strong in-room support).
Action Item: All working group members are strongly encouraged to review draft-ietf-green-energy-terminology and draft-ietf-green-use-cases and provide comments (even simple support/non-support) on the mailing list before the adoption polls close.
Action Item: Authors of draft-ietf-green-energy-terminology to continue aligning definitions with the framework and data model, and consider how to manage non-normative or future-use terms (e.g., an appendix or GitHub issues).
Action Item: Authors of draft-ietf-green-use-cases to finalize the Isaac use case and formally add Xinu to the author list.
Action Item: Chairs to facilitate and encourage broader participation in the ongoing "design-team-light" meetings focused on data modeling to unblock progress.
Action Item: Authors of draft-ietf-green-framework to continue exploring the integration of Eman MIB concepts for power information and address Yang provenance issues, particularly for data sheets.

Next Steps

Friday Session: The working group will hold a second, two-hour session on Friday dedicated to detailed discussions on the data modeling work.
Online Meetings: Continued "design-team-light" online meetings will be organized to progress the modeling work towards concrete document contributions.
Document Reviews: Working group members are urged to actively participate in the adoption polls and provide constructive feedback on all working group drafts.

Session Date/Time: 07 Nov 2025 19:30

GREEN

Summary

The GREEN working group held its second session at IETF 124, focusing on the progress of various drafts and the critical next steps for establishing foundational YANG models for energy measurement and control. Presentations included an update on the Petra Green API, a proposed YANG model for Smart Power Distribution Units (PDUs), and a new proposal on application layer efficiency for web crawling. A significant portion of the session was dedicated to discussing the ongoing data modeling efforts, including a design team update and a proposal to leverage existing E-MAN MIB concepts for a device energy model. Through polls, the working group indicated a strong preference for focusing on concrete use cases and using E-MAN as a starting point for device modeling, while also proceeding with controller and aggregation models in parallel.

Key Discussion Points

Chairs' Update: The chairs noted that two adoption polls are currently open for existing drafts and encouraged reviews. The goal is to establish base data model drafts for the working group. Design team meetings have been set up but have struggled to gain significant traction.
Petra Green API Update:
- Louise presented an update on the Path Energy Traffic Radio API draft.
- Motivation: Provide visibility into energy consumption across network paths, retrieving metrics like watts per gigabit.
- Changes: Refined wording, expanded scope to endpoints beyond IP addresses (e.g., optical networks, network slices), added security considerations, and an annex on SLA negotiation for energy-efficient services.
- Positioning: Petra is envisioned as a potential API service interface within the Green Framework, operating at customer, service, and network levels, aligning with RFC 8309.
- Feedback: The group felt it was too early for adoption, as the draft depends on the stability of core base models. The unit watts per gigabit and its rationale (allocation proportional to throughput) are open for further discussion.
Smart PDU YANG Model Proposal:
- Louise also presented a draft for a YANG model for Smart PDUs.
- Motivation: Standardize a vendor-neutral approach for monitoring and controlling Smart PDUs, which currently rely on proprietary solutions.
- Objectives: Real-time energy telemetry, control at outlet/PDU level, support for automation (remote control, power optimization).
- Discussion: Carlos suggested reviewing Home Assistant's approach for generic wrappers. Nigel noted potential overlap with IETF IV (inventory model) for equipment. Concern was raised by Benoit about standardizing a specific "Smart PDU" module, suggesting it should be an extension or use generic building blocks from a broader GREEN framework.
- Conclusion: The work is considered in scope, and the base data models should eventually cover this use case.
Application Layer Efficiency for Web Crawling (TCT):
- Anton Yukovych presented a draft proposing a "Collaboration Tunnel" (TCT) to reduce energy waste from web crawling.
- Problem: Wasted bytes from repeatedly fetching unchanged content consume energy across network, server, and client layers.
- Solution: Using efficient JSON instead of HTML, server-generated SHA256 fingerprints as ETags, and JSON sitemaps for zero-fetch optimization (client-side hash comparison to skip unchanged URLs). Demonstrated significant bandwidth reduction (83%) and skip rates (70-90%).
- Request for Guidance: The presenter sought guidance from the GREEN WG on energy measurement methodology, whether to include kWh conversions, and how to quantify efficiency.
- Chair's Feedback: While metric terminology definitions might align with GREEN's scope, the protocol aspects and full-stack implications of this work seem to extend beyond the current GREEN charter. The chairs will consult with Area Directors and other relevant IETF working groups (e.g., HTTPBIS) to find the most appropriate home for this work.
Data Model Design Team Update:
- Jan provided an update on the design team's work, outlining a conceptual pipeline for energy data (Collect -> Aggregate -> Store -> Deliver -> Analyze -> Present -> Act).
- Requirements: Emphasized the need for a plugin architecture, traceability, rich metadata, and flexibility to align with evolving political and accounting standards.
- Work Items: Four main areas were identified: (1) Controller-side collection/aggregation, (2) Device-level interfaces for reporting, (3) Policy-based device power control, and (4) Network-level orchestrator control.
- Discussion: Med advocated for focusing on Work Items 3 and 4 (control) as core to GREEN, suggesting Work Items 1 and 2 are more generic. Jan countered that Work Item 1 (collection/aggregation) addresses a significant industry gap, while Work Item 2 (new device interfaces) would be a long-term endeavor given existing legacy.
E-MAN Based Device Model Proposal:
- Benoit presented a proposal for a "bottom-up" approach to data modeling, starting with core device power/energy reporting.
- Rationale: Operators need concrete ways to report power and energy, which is foundational for all use cases. He highlighted existing concepts from the E-MAN MIB (e.g., power tables, entity physical index, accuracy, caliber, energy tables).
- Argument: While E-MAN is MIB-based, its underlying schema and concepts address critical issues like identifying objects (device, interface), defining power values (SI units, direction), and handling measurement accuracy and calibration (actual, estimated, static). He argued that current implementations are a "mess" and a standardized YANG approach based on these proven concepts is needed.

Decisions and Action Items

Petra Green API: The working group will continue to develop the Petra Green API as an individual draft. It is not ready for a working group adoption poll at this time.
Application Layer Efficiency for Web Crawling (TCT): The working group chairs will consult with relevant Area Directors and other IETF working groups to determine the most appropriate venue for further discussion and standardization of the TCT draft, particularly its protocol and full-stack aspects.
Data Model Development:
- The working group will focus its efforts on developing concrete use cases for data models, rather than purely abstract models. (Sense of the room: strong support)
- Work on device models will proceed. (Clear consensus)
- Work on controller and aggregation models will also proceed, in parallel with device models. (Sense of the room: split, but leaning towards parallel work)
- The existing E-MAN MIB (or an RFC converted to E-MAN) will be used as a starting point for the device model, providing a concrete basis for discussion and development. (Sense of the room: clear support)
Adoption Polls: The working group will continue to review and provide feedback on the two existing drafts currently open for adoption polls (open for another week). Following this, an adoption poll for the green-framework and green-control-plane-model documents will be initiated if authors agree.

Next Steps

Bi-Weekly Design Team Meetings: The bi-weekly design team meetings will continue, providing a forum for detailed technical discussions on the data modeling efforts.
Collaboration on Data Models: Benoit, Jan, Marisol, and other interested individuals are encouraged to collaborate closely to drive the development of the foundational device, controller, and aggregation YANG models, utilizing E-MAN concepts as a starting point for the device model. The chairs will facilitate these efforts.
Document Review: Working group participants are encouraged to review the two drafts currently in adoption polls and provide feedback before the polls close.