Delivering an HPC Service Under Real-World Constraints: Lessons from the CIUK Student Cluster Challenge

This technical white paper examines the delivery and operation of a High-Performance Computing (HPC) service under realistic constraints, drawing lessons from the 2025 CIUK Student Cluster Challenge. Designed to simulate the full lifecycle of an HPC service, the challenge required teams to deploy, operate, and evolve a multi-user cluster from bare metal while balancing performance, reliability, security, and user experience.

Participants designed and operated a four-node HPC cluster using industry-standard components, including Rocky Enterprise Linux, diskless provisioning, Slurm workload management, MPI-based applications, containerised workflows, and graphical user environments. Beyond initial deployment, assessment for teams focused on service operation, policy enforcement, application enablement, user onboarding, and documentation under time pressure and evolving requirements.

The paper outlines the architectural decisions, operational stressors, and service-oriented thinking implemented during the simulation, demonstrating how small-scale environments can expose many of the same risks and trade-offs as production HPC services. The document serves as a practical reference for educators, practitioners, and teams that deliver or operate HPC platforms in constrained or training-focused environments.

This work is made available under the Creative Commons Attribution 4.0 International License.