Semantic conventions for Software Carbon Intensity (SCI / ISO 21031:2022)

Open
#3,539 5 comments 11 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
5/5
Estimated time
Over a week
Newbie friendliness
35/100
Issue type
Feature
Clarity
Needs clarification
Activity status
Quiet

Research direction

Start with the SCI specification and the previous sustainability semantic conventions proposal in issue #2020, then review this issue's proposed components: E, I, M, R, and the composite SCI score. Done means producing a draft set of semantic conventions for review by the General Semantic Conventions SIG, potentially using the federated semconv model.

Written by the indexing model from the issue text.

Description

Background

The Software Carbon Intensity (SCI) specification (ISO/IEC 21031:2022) defines a standardised method for measuring the carbon intensity of a software system. It is maintained by the Green Software Foundation (GSF), a nonprofit consortium of 60+ member organisations operating under the Joint Development Foundation.

The formula is:

SCI = ((E × I) + M) per R

  • E — energy consumed by the software system
  • I — carbon intensity of the electricity consumed
  • M — embodied carbon of the hardware the software runs on
  • R — functional unit (normalises the score per unit of work)

The problem

SCI components map directly onto concepts that observability practitioners already instrument: energy consumption, hardware utilisation, and a functional unit. However, there are currently no shared semantic conventions for representing SCI-related measurements in OpenTelemetry. Without them, teams measuring software carbon intensity use inconsistent attribute names and metric definitions, and existing tools such as Kepler cannot be composed into a coherent pipeline without bespoke integration work.

A previous attempt to address this (#2020) stalled in 2023 when the primary driver changed roles.

What we're proposing

The GSF, in collaboration with Sarah Hsu (@greenhsu123), is convening a new project, the SCI for OpenTelemetry, to develop a proposed baseline set of semantic conventions for SCI components, covering:

  • Operational energy (E)
  • Carbon intensity of electricity (I)
  • Embodied carbon (M)
  • Functional unit (R)
  • The composite SCI score

The deliverable will be a draft set of semantic conventions for SCI, which we will bring to the General Semantic Conventions SIG for review and feedback. We plan to follow the federated semconv model, hosting the registry under the GSF's own organisation.

Why now

Standardising SCI within OpenTelemetry's semantic conventions would make carbon a first-class signal alongside performance, reliability, and cost, using the same registry that already unifies other telemetry signals. Once the vocabulary is agreed, it creates a foundation for downstream tooling: collector components, SDKs, and backends that can emit and consume carbon data consistently across the ecosystem.

Regulatory requirements (CSRD, EU Taxonomy, emerging US disclosure frameworks) are also creating demand from engineering organisations that need to report on software emissions, many of whom already rely on OTel for observability.

Who we're looking for

We are actively recruiting participants for the project. We are particularly looking for people with:

  • OpenTelemetry semantic conventions experience, to ensure the output maps cleanly into OTel from the start
  • Green software / carbon measurement expertise: SCI, Kepler, carbon intensity data sources
  • Instrumentation experience, to inform how conventions would be implemented in practice
  • Interest from observability backend vendors, to ensure standardised carbon signals are consumable across the ecosystem

If you have relevant expertise, or know someone who does, we would welcome your involvement.

Applications are open to all now at: https://greensoftware.foundation/assemblies/sci-for-open-telemetry/

Contact

If you'd like to get involved or have questions, feel free to reach out to any of us directly:

Related

  • #2020 — previous sustainability semantic conventions proposal (2023)
  • SCI Specification — the ISO standard this work is based on
Dominant language
Python
Stars
1.2k
Forks
326
Avg merge
1d 12h
Merged PRs (30d)
13

Getting set up

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from open-telemetry/community

All issues in open-telemetry/community

Similar issues

More Python issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.