Understanding Critical Software for UKRI Digital Research Infrastructure

This scoping study explores key issues around the critical software required to deliver the UKRI digital research infrastructure (DRI), and recommends what help UKRI could provide to support their investments. It also suggests next steps for future research to support DRI.

In a complex software ecosystem, understanding what software requires support and investment from UKRI is a complicated task. The Software Sustainability Institute (SSI) examined not just a list of software used by infrastructures, but considered a range of other issues. Our research questions were: 

1. What software is used by the UKRI DRI infrastructures researched?

2. What should be considered as ‘critical software’ for UKRI Digital Research Infrastructure?

3. What are the risks to critical software for UKRI Digital Research Infrastructures?

4. How might audits support the infrastructures in preparing for and mitigating risks?

To begin to address these questions, the SSI conducted a short study over the course of six months starting in November 2024, interviewing staff from 3 Compute infrastructures (ARCHER2, JASMIN, Isambard-AI) and 3 Data infrastructures (Seshat - part of iDAH, DARE-UK, PSDI) to ascertain details of their set up, funding, software stack, and audit/risk assessment processes, as well as their definitions of ‘critical software’ and opinions on the risks of disruption to the running of their infrastructures. Then, two focus groups were conducted with staff and users of the wider Compute and Data DRI community in order to corroborate and expand on the themes that arose in the interviews. 

The report explores the software stacks of the 6 infrastructures (Appendix B), and provides illustrations of different ‘critical software’. It looks at risks to software and also addresses the role of audits in understanding and preparing for risks to infrastructures. The report indicates how different parameters – types and size of infrastructure, users, available funding, and where the infrastructures were in their life cycle – influence the choices of software in use, and what this means for the risks to the infrastructures.