One traditional beneficiary of this pressure is outsourcing – offering CIOs the opportunity to tap into the collective expertise of a dedicated IT services organisation. However, if you speak with any CIO, it’s quickly obvious that outsourcing is no panacea. Popular statistics such as “50% of outsourcing arrangements end in failure” have more than a grain of truth to them. Susan Roberts of the Overtum think-tank claims this is to do with the same market unpredictability that makes outsourcing attractive in the first place: “to succeed, any arrangement requires mutual benefits - traditional outsourcing often lacks the flexibility to ensure that is possible.”

This perennial stumbling block in the outsourcing model is partly responsible for driving forward the concept of utility computing. The idea that companies will be able to purchase only the computing tasks that they need, when they need it, is phenomenally attractive to CIOs. Previously, their only option was to overprovision for the times when a business unit would require a surge in resource. Now, claim believers, the IT utopia beckons: companies drawing from a huge pool of interoperable resources delivered by the internet, leaving IT overheads as dynamic as the business requirements itself.

As with all visions, when the details are examined, the picture becomes more complex. For Utility Computing to be a sustainable business model, every computing task, transaction and processing cycle will need to be matched to an account. If computing is going to be delivered in the same way as water or electricity, it will also have to be priced in the same way. This will require an incredibly detailed understanding of who has used what and for how long. The demand will be both external – so that services providers know what to charge – and internal, in order for IT departments to accurately pass on costs.

It’s in this arena that Ejasent (www.Ejasent.com) operate. Founded in 1999 and a veteran survivor of the dot.com shake-down, they have evolved from an ASP model to offer software products aimed at enhancing data center efficiency. Their MicroMeasure product offers reporting facilities that both measures and analyses the utilisation of different computing resources.

This is vital for two reasons: Firstly, it helps with the issue of accurate bill back. Accurately charging business units back for their use is an attractive tool for CIOs, offering them a tool to encourage the view that IT is a business resource – not simply part of the furniture. Secondly, of course, detailed usage information is the entire foundation upon which the Utility Computing concept will either fail or prosper.

“Utility Computing is a great idea – no-one disagrees with that,” opines Roberts, “but one early critical success factor is obviously going to be the accuracy with which companies are able to measure usage – not so simple in practise as it is in theory.”

Ejasent claims that companies can not only benefit from the information that MicroMeasure provides but they can go onto then achieve greater efficiency by using their UpScale product to dynamically distribute demand amongst a pool of physically separate servers. As demand profiles fluctuate, processing capability can be assigned and re-assigned within a data center to ensure availability. Over time, of course, this information can be used to tailor requirements to the demand profile and reduce future infrastructure costs.

Whether or not offerings such as the Ejasent one will lead to IT departments being able to fulfil the multiple “flexible, accountable, speedier, efficient, accurate, cheaper” requirements that Murphy describes remains to be seen. However, even if that ideal does not materialise – at least the IT department will be able to present those same departments with an accurate bill for their demands.