- This workshop aimed to give IT and strategic decision-makers in HE a chance to find out more about the potential advantages and disadvantages of cloud and shared services, and the opportunity to reflect on the key factors that are likely to influence decisions in this area.The workshop was been jointly organised by two JISC-funded projects: Responsible Energy Costs led by the Forum for the Future, and Green IT for Science led by the University of Bradford.
The Importance of Cloud and Shared Services to Future Organisational Effectiveness
Richard Anning, Head, IT Faculty, Institute of Chartered Accountants in England and Wales
- Anning provided a brief overview of the ICAEW, which was granted a royal charter by Queen Victoria in 1880 and currently has around 4,000 members.
He specifically outlined the work of the IT faculty, which provides member services and thought leadership in the area of accountancy and IT.
Anning admitted that accountants are still in the very early adoption phase with cloud services, as they are risk averse by nature, so security, control and availability are key areas of concern. However, the IT Faculty at ICAEW has held a conference to raise awareness about cloud computing and issued a publication covering many of the issues, so the themes under discussion at the Cloud and Shared Service Solutions workshop were very pertinent to their current work.
- You can view Richard Anning's introduction in full in this video:
Key Business and Environmental Factors in Making Internal versus External Choices
Martin Bennett, RECSO Project Leader
Peter James, Green IT for Science Project Leader
James introduced the key concepts associated with cloud services, including SaaS, PaaS and IaaS, and the different cloud models available: Private, Community, Public and Hybrid. He also outlined the different types of shared service identified by HEFCE and discussed the key decision factors that need to be considered when moving towards such services, focussing particularly on the complexities of costing such solutions when many of costs involved are not immediately apparent.
Martin Bennett continued this theme, observing that cost information is not always easily available, making it difficult to identify the resources consumed by the different options. He observed that unless you're able to calculate the costs involved in the different options, including your existing solution, you cannot make a clear comparison.
Bennett discussed a breakdown of Amazon Web Services' published costs, which included 13% spent on power and 18% spent on power distribution and cooling. He used this to illustrate the depth to which institutions need to probe the various costs factors when looking at the changes in ICT provision that are on the agenda at the moment.
- James and Bennett's slides are also available here:
- You can view Peter James and Martin Bennett's presentation in full in this video:
The Cloud Adoption Toolkit: Supporting Cloud Adoption Decisions in the Enterprise
Ian Sommerville, Professor of Software Engineering, University of St Andrews
Sommerville discussed some of the issues which have arisen during the development of a cloud adoption toolkit by the University of St Andrews. Whilst this has been developed primarily for medium-sized enterprise, the toolkit can also be applied to many universities.
He observed that many of the problems they encountered when creating this toolkit were human problems, not technical problems. There are a lot of concerns from users about the notion of loss of control, and concerns from IT management about losing their organisational influence.
Sommerville emphasised that in many cases cost modelling is often too simplistic. Their toolkit helps organisations to identify their requirements in more detail, carry out price comparisons, and model their elastic demands. He emphasised the importance of modelling workload patterns to get accurate costings, which informed their development of an infrastructure modelling language to help organisations describe both their infrastructure and the applications deployed on that infrastructure. This enables users to model their workload patterns over time and generate reports demonstrating the costs that would be incurred from the different cloud providers for different loads.