Some say that cloud computing will be like the PC revolution all over again. The availability of commodity services will convince everyone that they’re IT experts and they don’t need a CIO to tell them how to create a new system.

The cloud will be ominous for corporate datacentre (DC) managers too. Vendors such as CA, Sun and Microsoft have made dire predictions for anyone who runs a datacentre for a company whose core business is not IT.

Fortunately at least one analyst thinks this is nonsense. According to Steve Wallage, MD of research firm BroadGroup, the datacentre is the foundation of the cloud. “The great advantage of the datacentre is it’s the one physical port in the virtual economy,” he argues. “That physicality can be tremendously reassuring.”

However, Wallage warns, just as IT can be dislocated from business strategy, datacentres are often dislocated from IT. They become a low priority and can be the cause of clashes between the heads of property, facilities management and IT, so companies tend not to concentrate on their power or their impact on the business.

In a way, the rise of cloud computing could help remind the business of the strategic importance of the datacentre. “The cloud will not kill off datacentres, but will instead make them more important,” says Wallage.

Projections see the internet expanding fifty-fold over next decade, adds the analyst. Even with virtualisation’s savings, this means there’ll be a tenfold increase in server numbers.

“The uncertainty around cloud’s security, performance and ROI have created a lot of resistance to change in the datacentre,” says Wallage. The majority of investment in datacentres goes on mechanical and engineering (M&E) elements such as chillers, power distribution and generators, all solid reliable items that are as far from cloud culture as it is possible to get.

Uncertainty is a big factor driving the demand for datacentres. Any company spending a hundred million plus on a large new DC – Google spent $900m last quarter – will want it to last at least 15 years, he argues.

“Given the uncertainty around cloud and IT development, can you have any confidence in IT needs over that period?” says Wallage. Banks like Lloyds TSB, who clearly also have business uncertainties, are delaying major datacentre decisions, he reports.

The cloud is making datacentres both smaller and more remote, Wallage adds. Ten years ago, global datacentre builder Digital Realty was building installations with an average life of 14 years and 20,000ft2 in size. Today, a typical UK deal is for a datacentre of between 4000 and 6000ft2 in size with a life of five to seven years.

Similarly, in 2001 80 per cent of outsourced datacentres were in London or within the M25. A new trend, reports Wallage, is for banks to take advantage of improvements in telecoms and cloud computing to build further out. HSBC has built in Wakefield and Morgan Stanley in Bournemouth, as banks have realised that only 15 per cent of apps really need to be run locally.

 Chop or change
The cloud will certainly affect the roles of datacentre staff. Clive Longbottom, head of research at analyst Quocirca, has predicted that IT datacentre staff have two options: to leave and find a job in a technical environment, such as a service provider’s datacentre, or to change their mindset, understand the business they are in and become the trusted conduit between business and IT.

“If they choose the latter, I believe they will be doing the business far more of a service and will find it far more interesting and fulfilling,” says Longbottom. If they take the former, then they will become drones. Tasks will increasingly be automated and they’ll be left to deal with technical exceptions to the automated tasks.

On the other hand, the cloud will be great for those CIOs who perform a more strategic role. Rather than being constrained by the hardware resources available to them, the cloud puts the CIO a great position – “like a kid in a sweetshop”, according to Longbottom.

The promised elasticity of the cloud allows the CIO to concentrate on what functionality is required. “If the business has a problem with customers, let’s get a CRM solution,” says Longbottom. “This isn’t cloud computing, it’s hosting, and just takes existing problems and puts them on someone else’s hardware.”

If the business runs on a set of processes and each process is made up of a set of tasks, then by identifying the individual tasks and the technical resources available to fulfil these, you can identify the gaps that cloud apps can fill. It’s a lot quicker to buy cloud services than to build the functionality into a datacentre, he says.

The CIO's view of cloud collaboration: choose candidates wisely

But this raises questions, says Longbottom. “How are the various functional pieces to be stuck together to work in an efficient manner? Do the various risks involved with placing the function in-house or in the cloud match with the organisation’s corporate risk strategy?” he asks.

CIOs should be prepared for more dilemmas, he says. “What is Plan B should our chosen provider go out of business, be acquired by a company we don’t like or if service levels fall below our requirements? Who owns any data created, and how can we get our hands on it?”

Comments