Putting the Government in the Cloud

My latest foray to the British Computer Society (BCS) was to learn more about Cloud Computing.

The Cloud is one of the big topics at the moment, e.g. it is one of Logica's three strategic themes, but there are many assumptions behind the terminology and I wanted to dig in to some of these.

Part of this was to try and work out what all the fuss is about as almost everything that I do with computers is in the Cloud already. All my photos are on Facebook, my bookmarks are on del.icio.us, my contacts are on LinkedIn and Google looks after my email, calendar and blogs. There is very little of importance on my hard drive.

Tony Heritage, of IBM's Central Government Team, kept a packed room enthralled with the story of the Government's adoption of the Cloud, the G-Cloud.

He opened with a warning about Shared Services, the Government's last attempt to drive down IT costs. This produced lots of providers of shareable services but no consumers.

However, the prospects of the Cloud are much better with, for example, Google able to run 2,000 servers per employee.

The G-Cloud has a fairly common layered approach with networks, data centres and applications.

Consolidation of networks and data centres is a no-brainer. There are some budgetary and governance issues to fix but the rewards are big enough to make them worth tackling.

Thinks get a little trickier after that as the promise of shared applications has never worked before because every organisation will claim vigorously and loudly that they are unique and need the applications tailored to them. With tailoring comes cost and then trying to share applications becomes even more expensive that keeping them separate.

The many questions and comments from the floor shared the cynicism, perhaps not too surprising given that most of the people there were from my generation and had lived through bureau services, outsourcing, software components, applications frameworks, etc. etc.

I was left with the conviction that the G-Cloud is an oversimplification of what is required to consolidate and share, in particular the market dynamics required to make it work (e.g. who could possibly bid to provide this, how would the government contact with them, how would they engage with the many user organisations). It won't work.