Build an IT Utility?

When Loyalty Management Services (LMS) launched the Nectar Card scheme, IT director Fiachra Woodman found himself with an all too familiar problem. Early demand for the scheme had outstripped all expectations, and Woodman’s systems were in danger of buckling.

Most IT executives would have opted for the traditional solution – a massive purchase of extra capacity that would likely remain idle once the initial surge of demand had passed. Woodman took a different approach – outsourcing the bulk of LMS’ data centre to the utility service provider, Savvis.

Today, thanks to this unorthodox decision, LMS is confident it need never worry about being short of processing capacity again. Even better than that, says Woodman, by questioning the “economics of redundancy” and the need “always to buy extra”, LMS’ IT costs are not artificially inflated.

LMS is now in the happy position of being able to access as much IT processing capacity as it needs, when it needs it, and only ever paying for what it uses. It is a situation that, says Woodman, “has really improved my relationship with the CFO.”

Woodman is not the only IT executive ready to praise the virtues of utility IT services. In this year’s Information Age Effective IT Survey, use of on-demand utility processing services was voted one of the most effective IT strategies, and at last year’s Effective IT Summit 2006, no less a luminary than Harvard Business School Professor Nicholas Carr warned that the future adoption of a utility model is not merely desirable, but inevitable. But this high praise among utility proponents is not matched by take-up in the market.

The reality is that despite the experiences of pioneering users like Woodman, and the hard-nosed economic analysis of academics like Carr, utility service uptake is still a minority practice.

Only 8.2% of Effective IT Survey 2007 respondents have adopted it so far and, despite bullish noises from utility providers like Savvis, other measures of utility service growth are equally unimpressive.

As recently as three years ago, this lacklustre interest in utility services would have been simple to explain. Then, network bandwidth scarcity, security fears, and inadequate management tools were just some of the technology hurdles that still needed to be overcome to make on-demand utility services practical.

The future adoption of a utility model is not merely desireable but is inevitable, warns Harvard Professor Nicholas Carr.

Now though, no such barriers remain. Indeed, in the past several years, the successful and widespread deployment of technologies such as Gigabit Ethernet networking and virtual machine software, and the remarkable success of metered, software-as-a-service (SaaS) offerings like, have wiped away most of the traditional structural barriers to utility deployment and even demonstrated that a real appetite exists for pay-as-you-go enterprise computing. So why aren’t on-demand utility services growing at least as fast, or even faster than SaaS?

Certainly, one key barrier to more widespread use of on-demand services is that, like SaaS services, they are not equally well suited to all classes of application, or supported by all categories of application vendor.

At LMS, said Woodman, the company was fortunate to possess a largely home-grown set of software assets, which meant it couldn’t mount its key applications on Savvis’ dynamically scalable server farm without first having to negotiate new licenses from its software suppliers. If this had been necessary, it might have stopped LMS’ strategy in its tracks since, according to Woodman: “some vendors need to come into line [with their licence policies]. Until they do, the utility model won’t come into force.”

In fact, with even Microsoft now embracing the SaaS payment models in its Microsoft Live! strategy, widespread change in this area may already be underway.

Those vendors that try to cling to traditional per seat or processor license models may yet find themselves sidelined by the growth of alternative metered charge models

Still, even as trends like SaaS are breaking down traditional software licence barriers to utility services, and technologies like server virtualisation continue to make the utility model an ever more cost-effective strategy for service operators, one key obstacle to real growth must still be overcome: the IT manager’s fear of the unknown.

However much utility service suppliers emphasise the potential their services have to create savings that the imaginative IT executive can use to more profitable effect elsewhere, there is no hiding the fact that the utility model is an outsourcing model. As such, it represents a threat to established business IT hierarchies.

That, of course, is part of the reason that business economists like Carr (who believes that “IT is nothing special”) are so enamoured of the utility model. However, it is also why, at events such as Effective IT 2006, populated by leading IT practitioners, utility arguments are typically met with stony scepticism – and rightly so.

Ultimately, the argument for utility IT services made by Carr and others is sound. Conventional corporate IT delivery models, founded on the creation and maintenance of a private IT infrastructure are wasteful.

IT utilities, which constantly juggle service loads to derive optimal value from the underlying infrastructure, should always be more effective and more economical.

But must IT utility services also always be owned and operated by a third party service provider? Conventional IT utility wisdom, which has always treated the energy utilities as its model, says yes.

Perhaps, as more corporate IT owners recognise the true potential of their own freshly virtualised and service oriented architectures, they may come to see that there are really no barriers to building IT utilities of their own.

Lee Biggins

Lee Biggins developed as an entrepreneur selling cold cans of fizzy drinks to fishermen on hot days, to running his own car washing empire as a teenager. This spirit carried through to the developmental...

Related Topics