A Strategy for Cloud Computing

So you’ve heard all the hype about Cloud computing and are now thinking about how you might enter the fray, test it out, give it a try? You even have the endorsement of Cabinet Office and the Government ICT Strategy let alone secure networks (PSN) and the G Cloud under development. So what services would it be best to consider?

My colleague Martin Howitt (Enterprise Architect from Devon CC) has very rationally suggested that it is integration and SOA that is the holy grail of cloud computing and that is where we should invest. It’s hard to disagree with Martin looking at it from a purely architectural viewpoint (http://forestandtrees.wordpress.com/author/martinhowitt/).

However, in my experience which is admittedly public sector biased, this type of decision is more political rather than taking the best technical option. In Local Government individual Directors, as head of professions, hold enormous power and virtually run their Directorates as private companies with the Local Authority acting as the conglomerate HQ. It’s hard to convince all powerful Directors that they should act for the good of the Local Authority and take all the risk on behalf of the others. The decision really requires a little more subtlety by CIOs keen to move into Cloud computing.

The obvious first step is to use cloud as a stop gap – use it for development and testing services initially to prove capability rather than going straight for full applications hosting. This takes advantage of the Cloud service providers’ capability to build platforms quickly which can speed development and reduce pressure on critical staff. There is no need to procure and pay for server hardware as you only pay for actual usage. The risks at the development stage are also a lot less than for losing a production service (and data). The cloud offers big opportunities for short term hosting such as testing upgrades which only lasts for a short period. Best practice would substantially reduce the number of environments to permanently maintain whilst still retaining flexibility to restore known end states. That should increase the Cloud’s credibility whilst saving funding on new projects.

So you’ve proved the capability of your cloud services provider on development services and are willing to take the next step of hosting applications? My approach would be to look for a utility application, core to my business, but not likely to make any major technology leaps in the near future and one that my cloud services operator has heaps of experience of. You could use a portfolio analysis to help you analyse things (in McFarlan’s IT Portfolio Analysis – Factory systems) but I think there is one application that stands out. Email - a standard application, configuration and hardware that will save bundles of cost. Not only that, but it does not have a clear service owner that could object.

I would also look to achieve one other major benefit. Every few years, even if you miss one, you have to upgrade. This is a major headache where you put in lots of effort and only perform once. How much easier to pass this over to an experienced provider that has done it before, has the ‘T’ shirt and probably all the scars. Let them take the pain and roll it out on a phased basis. Undoubtedly, this would be my first choice for hosting a service via the cloud. Next I’d be looking for other similar applications. Standard operations (in every sense) where I could cut costs – ISPs and shared workspaces spring to mind.

Having made the leap into Cloud and everything is now settled and operating well I’d be looking to see where I might gain some ‘advantage’. Obviously being Local Authorities we don’t actually complete in a commercial sense but to the uninitiated do not be fooled. It’s a bit like F1, the first person you want to beat is your team mate. (For central government think departments – DWP and HMRC for example.) I’d be looking to really exploit collaboration services. Not as many have already achieved by implementing Sharepoint internally but working externally with the Local Strategic Partnership. Incorporating other public sector bodies and significantly third sector organisations (TSOs). Major performance benefits could be achieved by sharing project based information with partners.

Note now I have moved away from the Factory based systems of McFarlan. We are in the area of Turnaround systems which are more future based, innovative and experimental. In a word ‘risky’. For this reason I would be less inclined to go with a standard solution but more favour an innovative or Open Source option. It remains to be seen if the service providers of the G Cloud will be prepared to take that risk or stick with the traditional ‘safe’ options? Certainly within SOCITM, there are individuals such as Glenn Wood (Wolverhampton) who are keen to champion the open alternatives as well traditional suppliers.

Since moving out of direct Local/Central Government I have been struck by the innovation that has taken place in the Social Media arena and how central based systems are being left behind by the extent of innovation in these spaces. Frankly it reminds me of the late 80’s when the first computer networks were challenging the heartlands of the traditional mainframe suppliers. Perhaps this really deserves its own blog but I am still struck by the number of CIOs that are not engaged (don’t even have Twitter accounts) and the fact that we have systems and procedures in place that prohibit their use at work from council/departmental systems.

So, my strategy for the cloud is to start safe using development services then move into utility and finally new innovative services. But what about legacy systems, I hear you say. Well, my view is that unless you can justify moving them onto standardised platforms then it will be difficult to develop the business case. That does not mean to say that you can’t move towards common (shared) data centres but the benefits of the cloud require dumping those traditional architectures. And so those unmentionable mainframes in central government chug on, and on, and on……….