Lessons for Utility Computing
Most utility computing pioneers are in dire straits. For instance, StorageNetworks, one of the highfliers of the Internet era, enjoyed a peak market value of $14 billion but today is being liquidated. Only utility computing providers such as IBM and Hewlett-Packard are hanging onand then only because of generous infusions of capital. The good news for utility computing is that a model exists: the power utility business. Consider these lessons:
Delivering the power is not enough; ease of use matters, too. The main selling point of electrical power is so understood that its rarely articulated: Electrical power has many uses and is user-friendly. Computing power is similarly dynamic, but its not as easy to access. Customization and implementation are so expensiveand such a drag on upgrades and maintenancethat the ease-of-use appeal so central to selling any utility becomes moot.
Users want the power, not the power plant. The pricing model for utility computing is based on selling the power and the system itself. If computing power is to be sold and consumed like electrical power, providers must remember that users want only the number-crunching power, not the resource-hungry hardware that generates it. Pricing must shift toward a traditional utility price plan, where billing is a function of how much power is used.
Electrical utilities were built on monopolistic advantages. The heavy financial lifting behind a utility startup has most often relied on the built-in protections of monopolies. While power utilities begin with little to worry about except serving the market, utility computing companies must consider competition and develop business plans that account for other providers.
The law of supply and demand never takes a vacation. In the 90s, the U.S. electricity market was glutted, so no major power plants were built. If theres no demand for utility computingor if its uncertain the market will be more than smallits a gamble to invest in utility computing infrastructures. The salvation of the business will rest on giving customers flexible configurations and realistic pricing, including lowering the cost of implementation, providing true pay-as-you-go pricing and breaking the vertically integrated model to focus on value-adding activities. Success can be found by studying power utilities history, but most utility computing enterprises are still learning the hard way.
George Chen and Lloyd Switzer are consultants at the strategic consulting firm Strategos, based in Chicago. Free Spectrum is a forum for the IT community and welcomes submissions at firstname.lastname@example.org.