In the early days of atomic energy, when safe, abundant nuclear power seemed just around the corner, the chairman of the Atomic Energy Commission bravely predicted that electricity would soon be "too cheap to meter."
That famous forecast, made in 1954 by Rear Admiral Lewis Strauss, never came to pass. Indeed, Admiral Strauss would probably be dismayed to learn that nearly 50 years later we can't even build a reliable grid to distribute electricity, let alone generate it so cheaply.
But you can't judge the comment too harshly. It's typical of the early phase of any new technology, when the promise seems limitless.
That's how it used to be for Web services. Pioneers envisioned a world in which enterprise-class applications would be built from myriads of self-describing, openly connected, Web-based components. Business people would build their own new apps by shuffling pieces without troubling IT.
In other words, computing power would be too cheap to meter.
The reality is less appealing. Although most enterprises have experimented with Web services, managing them has turned out to be harder than many had hoped. And legions of highly paid consultants have sprung up to help sort it all out.
In that context, it's refreshing to consider some of the findings of a recent report by our sister company, IDC. The study, published in May by analyst Sandra Rogers, reported on a survey of IT execs involved in Web service projects. Among the results:
Most companies want to improve existing systems, not to boldly go where no firm has gone. Rogers found 38 percent of projects were aimed at replacing existing functions; another 32 percent at integration. Only 28 percent focused on new functions. And 60 percent of respondents expected the projects to pay for themselves within a year.
Companies hope to save money, not get rich quick. That's no reflection on Web services; presumably you could turn a better profit by improving your efficiency, integration, and ease of upgrade. But nearly half (48 percent) of today's projects have expense reduction as their goal. Only 25 percent aim to provide a competitive advantage, and a mere 13 percent are to generate revenue.
Most companies still build their own. Some 70 percent of projects today are developed internally, versus about 10 percent by systems integrators or consultants.
Most companies want a partner that adheres to standards. That's not surprising, given that Web services have always been standards-focused. But at the other end of the spectrum, companies give the lowest consideration to whether a potential partner can provide a complete solution. Buyers understand that "complete" usually means "proprietary," and they'd rather find best-of-breed building blocks and assemble them on their own.
That's smart. It may not yield computing power too cheap to meter, but it will definitely provide a solid return on investment.