In one of Mission Critical’s very first issues, Christian Belady wrote that increasing energy efficiency in the data center could have the adverse affect of increasing energy use by causing more applications to become cost effective, thereby raising demand. He called the phenomenon Jevons Paradox and predicted it would a major part of a troubling cycle.

Since then, we have witnessed great changes in IT that affect data centers. The first examples are the increased penetration of blade servers and the use of containers in a number of applications, including at very large data centers owned by companies such as Google and Microsoft. I think announcements of new data centers by companies such as Facebook (Oregon), Yahoo (New York) and Apple (North Carolina) may be just as significant since they since to just barely precede new service announcements by these companies. Finally, we seem to be witnessing increasing growth of companies like Coresite, Digital Realty, PAETEC and many others that provide colo, SaaS, or IaaS. I recently read a Tweet about a company that had reduced its IT workforce to four: “The tail that wags nothing.” I expect that this company had outsourced its IT efforts at significant cost savings, judging by the staff reductions noted in the Tweet.

Traditional IT departments may resist these changes, particularly when jobs are on the line. After all, IT is more than physical reliability; it is valid when IT expresses misgivings about who can access proprietary data or equipment when some or all services are outsourced to a service provider.

Still, these changes and others result from efforts to control the overall cost of IT operations, including energy, while maintaining reliability. The environment is very competitive, and it is a maxim of economics that increased competition will result in reduced prices and increased services over time. Just today, for example, I learned that PAETEC is guaranteeing a 100 percent SLA to any customer that configures its equipment to take advantage of the reliability features in its facilities. Intense competition in the market will cause others to match, or even exceed, this offer. Other companies will compete on price or develop other services that reduce customer costs.

My personal experience does not provide much insight into how companies are responding to reduced IT prices, outside a few web-based services that BNP Media now employs to handle expense reports and health-care claims. I expect, however, that we can use consumer companies as a proxy for all companies. In that way, we can see that Belady had a point, as consumer companies find that offering new services is preferable to reducing costs, that demand for these new services is seemingly unquenchable, and that constant innovation is essential for remaining competitive.

One need not be a social media guru to understand that these services use a lot of energy and should be expected to use more. The Facebook, Apple, and Yahoo facilities now under construction say volumes about the business plans of these companies: Whether the product is locational services, streaming music, or greater entertainment access, these behemoths will be using more energy even while reducing the costs of offering these services and increasing profitability.

Efforts to control overall energy use in data centers may be doomed to failure, if our conservation plans and metrics do not consider Jevons’ Paradox.