Some see utility computing as a technology while others describe it as outsourcing with a variable bill. Another school of thought likens it to the IT equivalent of running water or electric power. More cynical observers see this as just another opportunity for a new generation of jargon, technobabble and self-interested vendor spin.
Is it any wonder that a large number of IT managers are still waiting for someone who could tell them what utility computing is and how it's supposed to make their lives better? At this point, let's step back and examine the question that utility computing is trying to answer.
The main issue is the big disconnect between making up-front investments in IT and the long-term process of collecting financial returns on that investment. Once an organisation buys a computer, there is no guarantee it will recoup its money's worth over the long run. And even though it is up to the owners of a computer to use expensive IT resources effectively, it is very hard to see how specific IT expenses connect to specific revenue or productivity benefits.
This disconnect between IT costs and business benefits leads to all kinds of IT efficiency and management problems.
The resulting symptoms get expressed in system underutilisation and inflexible installations that are hard to change in the face of changing business conditions. At the same time, there's little visibility into which computerised business operations contribute the most economic value.
On top of everything, there is a sneaking suspicion among frontline businesspeople that because no one can really measure IT's benefits, maybe IT doesn't matter as a strategic business advantage.
Utility computing addresses these issues in different ways.
- It provides technical ability to scale computing resources dynamically up and down to cope with fluctuating workloads.
- It changes IT pricing from up-front investment to real-time, pay-as-you-go.
- The fine-grained monitoring required to generate pay-for-use billing reports provides much greater visibility into IT operations, their costs, and their relationship to the business activities they support.
One myth about utility computing is that it always involves outsourcing. Outsourcing is only one way to implement utility computing. Other options include owner-operated internal utilities, managed service "multisourced" utilities, and specialised fee-for-service providers (ISPs/ASPs/hosting companies). Each model has its virtues depending on specific customer needs.
No matter which utility model an organisation chooses, customers need to be sure that utility computing is the right solution for their needs and doesn't add cost and complexity without definable, measurable benefits. "Gotchas" here include long-term contracts that lock in customers to a company's technology base or distance and alienate customers from decision making.
One thing that utility computing doesn't do is commoditise computing. Digital ones and zeros are more valuable than ever before, and their value varies according to the tasks they perform. Inflexible contracts that treat computing as a commodity, cut customers off from visibility into IT costs and benefits, and reduce customer power to make decisions is big step back for IT managers who want to extract the maximum value out of every dollar, euro, yen or peso of IT spending.
But by directly connecting information technology spending to real-world business activities, utility computing gives managers new ability to maximise the strategic business value of information technology. This can only be a good thing for businesses and the IT industry as a whole.
biography
Patricia C. Sueltz is the executive vice president of Sun Services at Sun Microsystems.
