JISC

Cloud computing is increasingly seen as a useful and possibly cost-effective resource for research computing but is still at an early stage of adoption. Use of the cloud is seen to give potential benefits such as: on-demand availability of compute, bandwidth and storage capabilities;  pay-as-you-go charging, thus reducing investment in underused resources; taking advantage of the economies of scale enjoyed by the cloud service providers; and  replacement of large peaks in capital expenditure by operational expenditure.

Studies carried out during 2010 developed guidance on the technical aspects of use, adoption and migration of cloud computing for research together with governance, legal and economic issues. RCUK also held a workshop to discuss inter alia the issues around the use of cloud computing for academic research, and how they could be addressed. Key issues identified by previous work and confirmed by the RCUK workshop are that:

• it is difficult to predict the cost of using the cloud;
• in many cases there is a lack of transparency of the costs of current in-house e-infrastructure and so comparisons with cloud costs are also difficult.

These two factors together make decision-making difficult. Decisions on whether to use the cloud or not, include:

•  researchers making cost-capability trade-offs for use of the cloud or an e-infrastructure;
•  institutions (or departments or heads of research groups) making investment decisions on enhancement or replacement of its current e-infrastructure;
• research councils making grant-allocation decisions regarding funding bids that use the cloud or an e-infrastructure (eg cost of long-term data storage).

This project, jointly funded by EPSRC and JISC, aims to:

• analyse the costs of ‘cloud computing’  for the types of research task within the remit of EPSRC, specifically in comparison to the costs of other approaches such as institutional facilities or distributed infrastructures such as grids;
• provide guidance to three distinct stakeholder groups:
•  researchers considering  the appropriate computing approach  to their own research problem  within the remit of EPSRC;
•  institutions making investment decisions  for computing infrastructure  for research  in engineering and physical sciences;
• EPSRC making grant allocation decisions regarding cloud and other computing infrastructures.