Building and sustaining an HPC cluster as a bioinformatics resource in resource-limited settings: principles and practice from the UVRI experience

Abstract

Building and sustaining a High-Performance Computing (HPC) cluster in resource-limited settings presents unique challenges and opportunities. This article outlines ten key rules derived from the Uganda Virus Research Institute’s (UVRI) experience in establishing an HPC cluster using a mix of previously existing equipment and servers mobilised from different sources including donations from partners. Our journey demonstrates that clusters can be built from diverse repurposed hardware resources combining donated and existing components into a cohesive and functional system. These rules focus on critical aspects to facilitate a home-grown solution such as strategic investment in people, proper planning, leveraging collaborations and adopting open-source tools to ensure sustainability. The article aims to offer practical guidance to institutions facing similar constraints on how to approach HPC infrastructure development strategically and sustainably. By following these basic principles, institutions can build robust HPC clusters and associated technical capacity while ensuring long-term operational success.