• Recommendation 1: any organisation considering adopting any cloud services for mission-critical applications, or for processing personal or otherwise sensitive information, should obtain specialist legal advice regarding their contracts and SLAs.
• Recommendation 2: JISC should investigate the issues surrounding the management and preservation of research data in cloud systems, and produce guidance aimed at researchers. This should support risk assessment and management, and should not design or develop technical solutions.
• Recommendation 3: JISC should investigate mechanisms for national engagement, negotiation and procurement of cloud services, primarily with AWS, Google and MS, but allowing for the engagement of smaller, niche providers.
• Recommendation 4: The NGS, and its funders, should consider whether there is a role for that organisation in supporting the development of virtual machine images for common research codes, to allow users to deploy them easily within commercial and private clouds. This may include liaising with or funding the developers or maintainers of the codes.
• Recommendation 5: unless backed by clear evidence of demand, and a robust and revenue-neutral business case, JISC should not support the development of a production UK academic research cloud.

Taken from the executive summary of a study funded last year. I hadn't spotted this report before. Its recommendations are pertinent to the UMF 'cloud' activity (particularly the JANET Brokerage service, advice being developed by DCC and the pilot cloud infrastructure that we are involved in) and to our forthcoming symposium.