We’ve funded a range of projects to guide our roadmap for the future

FAIR-Compute: A Roadmap for Fair and Efficient Allocation of Federated Digital Research Infrastructure

As the demand for High-Performance Computing (HPC) grows, the challenge of matching diverse research needs with finite computational resources becomes increasingly complex. Current scheduling mechanisms, such as Fairshare and Backfilling, are effective at maximizing local utilization but may face limitations when scaling to a Federated Environment where resources are shared across institutions.

FAIR-Compute investigates whether principles from Economic Theory and Mechanism Design can complement these existing engineering solutions. Our interdisciplinary team — spanning Economics, Computer Science, and Game Theory— is mapping the landscape of allocation strategies, from current heuristics to theoretical market models.

We ask: Can we design “Coordination Mechanisms” that incentivize efficient sharing without requiring central control or monetary transfers? By simulating alternative approaches, such as non-monetary credits (Scrip) and Dominant Resource Fairness (DRF), we aim to identify policy innovations that ensure equitable access for the UK research community. The final output will be a roadmap for the National Federated Cloud Service, exploring how local optimization can lead to globally fair outcomes.