Currently, more than 15 working groups are actively implementing the SDIA Roadmap for Sustainable Digital Infrastructure by 2030. This technical update is from the working group "Carbon Footprint Framework for Server-Side Applications" and addresses the digital carbon footprint. Focus of the group: better data availability and reporting on CO₂ emissions in the digital value chain as a basis for better decision-making. The update introduces the simplified equation to measure the overall environmental footprint of server-side software.

The Challenge: Infrastructure Alone Doesn't Change

Until now, IDED has addressed sustainability in the digital sector primarily from an infrastructure perspective — servers, data centers, etc. However, it has become evident in recent years: Infrastructure changes slowly unless driven specifically by customer demands. Software, on the other hand — especially system administrators and software developers building applications on infrastructure — is agile and adaptable.

When it comes to determining the environmental footprint of IT, we should therefore consider the perspective of the software developer. The requirements can be reduced to two fundamental elements:

• The user should be able to see how much environmental impact a single request or operation causes — through an API or an interface.

• The entire environmental footprint of a company's IT software portfolio should be measurable and assessable.

Version 0.1 of an Environmental Footprint Metric

Perfection hinders progress.

One of the working groups on the digital carbon footprint developed a framework for measuring the environmental footprint of server-side IT applications. The initial approach from the infrastructure side resulted in over 30 metrics and data points — too complex. The simplified version looks from above at the environmental footprint of a server-side application: The largest "application" item is the server itself. A comprehensive metric that includes server energy consumption and embedded emissions from manufacturing is a sensible first step.

Measurement in Practice

Most software applications today do not run on a single physical server but in a virtualized environment. This does not fundamentally change things — it merely means distributing the environmental footprint of the physical server proportionally to the virtual instance, based on assigned resources (compute, memory, storage).

The second challenge lies in the differing requirements: Should the entire environmental footprint of a server-side software be measured over its runtime — or the footprint of a single request? Both requirements build on each other. To solve the first, the power consumption and embedded CO₂ emissions of the server must be continuously recorded, resulting in a time series.

The question "What was the environmental footprint of this request?" can then be answered with two parameters: When did the request start, when did it end? With this information, the recorded time series can be queried, a snapshot created, and the numbers aggregated.

Identifying the most compact information for a simplified metric creates data transparency to optimize overall power consumption and thus reduce energy use.

Thanks to John Laban and Leonieke Mevius (SURF) for focus and support on this journey.