PhD Position F/M Defining and implementing metrics for Cloud sufficiency

Inria, the French national research institute for the digital sciences; France; Hybride; EUR 35 000 - 45 000; Plein temps

Résumé du poste

A national research institute in France is seeking a full-time researcher for a temporary position focused on studying metrics related to digital technology and data centers. The role involves analyzing existing metrics, designing new scores, and developing software solutions for improved measurement and user engagement in sustainable digital practices. Candidates will enjoy a flexible working environment with extensive leave benefits and opportunities for personal development.

Prestations

Partial reimbursement of public transport costs

7 weeks of annual leave + 10 extra days off

Possibility of teleworking

Access to vocational training

Social security coverage

Responsabilités

• Conduct an analysis of existing metrics for data centers and identify reward mechanisms for behavior change.
• Design scores based on existing or new metrics to inform users about energy consumption impacts.
• Develop software for metrics and scores with a focus on user-friendly presentation.

Description du poste

Inria, the French national research institute for the digital sciences

Organisation/Company Inria, the French national research institute for the digital sciences Research Field Computer science Researcher Profile First Stage Researcher (R1) Country France Application Deadline 7 Jan 2026 - 00:00 (UTC) Type of Contract Temporary Job Status Full-time Hours Per Week 38.5 Offer Starting Date 1 Feb 2026 Is the job funded through the EU Research Framework Programme? Not funded by a EU programme Reference Number 2025-09647 Is the Job related to staff position within a Research Infrastructure? No

Offer Description

Since the beginning of the 21st century, the importance of digital technology in general and data centers (DC) in particular has only grown with exponential expansion in recent years. Initially invisible, DCs represent a significant share of global electricity consumption, which could rise from 2% today to 10% in 2030. Even if the efficiency of each DC increases, this does not compensate for their proliferation which reached 67 million servers in 2022 according to ADEME. The historical models proposed for digital consumption by Anders et al. in 2015 propose several scenarios. As early as 2018, the worst-case scenario was already observed in the field. Thus, DC consumption in Ireland exceeded that of urban households, representing 20% of the country's production, which jeopardises decarbonisation projects in other sectors of the economy. This issue is entering societal debates and is increasingly being covered by the media in articles or reports.

The first objective of this thesis is to study the technical and human conditions for implementing sobriety in the cloud (i.e., DCs), to identify the levers and to put their consequences into perspective.

More precisely, this thesis focuses on the definition and implementation of metrics: The objective is to define metrics (DC-score, euro-score, energy-score, carbon-score, CSR-score) that are meaningful at the level of users and the nature of requests to inform the decision to submit, postpone or abandon requests for cloud access and thus fully realise the benefits of virtuous behaviour. The perspective of providers must also be studied when they face a demand for load shedding and must both manage energy storage and reduce service offerings to users. The goal is to take rebound effects into account upstream of the design and use of data centres.

Description of Tasks

Study and Analysis of Existing Metrics and Scores

This task aims to conduct an analysis of existing metrics, on the aggregation of metrics to constitute scores (weight of each metric, interpretability, explainability), on the methodologies for creating/developing indicators and scores, on their scope with respect to users (different metrics or scores depending on the targets). An analysis of reward mechanisms and incentives for behaviour change towards frugal use of digital technology will also be included, projecting it onto the case of a data centre, by asking questions such as: What do we gain from following the indicators? How do we measure the reward?

Design of Scores Useful for a Frugal DC, Adapted to Users

This task constitutes the core of scientific advances on metrics, indicators, scores and rewards. Proposals will be made to constitute scores based on existing metrics or on metrics to be identified. The scores will serve to inform and encourage users towards virtuous behaviours beyond the charter. More than just the carbon cost, a score should reflect the impacts on a subset of the 9 pillars of planetary boundaries (such as water) and can be adapted to a user profile (or a community) allowing better efficiency towards sobriety. For example, criteria such as the social impacts of consumption and its reduction, the (perceived) usefulness of uses and services, the carbon content of energy, its cost, the strain on supply, etc., can be more or less emphasised.

The question of a score's sensitivity to change will be studied, e.g., by ensuring that a minimal improvement of a decision at the cost of too high a price is properly reflected in the evolution of the score in a significant way. The question of predicting scores will also be raised. Particular attention will be paid to the design of understandable indicators, facilitating user adoption in a logic of conviviality.

Several scores will be compared for qualitative comparison purposes.

Software Development of Scores and Metrics

This task tackles the development of new metrics and selected scores. The ergonomics of score presentation will be particularly taken into account to facilitate their adoption. The development itself will be open (free and open source) and frugal (in memory footprint, in computational impact). We will use the Alumet software developed at LIG which performs frugal, non-intrusive measurement, which is modular and which will allow, for example, the integration of modules to retrieve metrics (such as Mojit/OS developed at IRIT which will serve as proof of concept) or calculate scores.

• Partial reimbursement of public transport costs
• Leave: 7 weeks of annual leave + 10 extra days off due to RTT (statutory reduction in working hours) + possibility of exceptional leave (sick children, moving home, etc.)
• Possibility of teleworking (90 days / year) and flexible organization of working hours
• Social, cultural and sports events and activities
• Access to vocational training
• Social security coverage under conditions