Mike Evans, vice president of corporate development at Red Hat, says that this is also a way for companies to experiment with the technology: "They're not just going to turn over their whole IT department to an external cloud in two months. We're seeing them say, 'Let me figure it out through an internal cloud.'" Fellows adds, "That's the next rash of product offerings that we are going to see--anyone who can bottle some of the magic that Amazon and Google have got and make it available internally for enterprises."

In the longer term, "the real value in cloud computing is not going to be in the under­lying hardware and basic services," says Sloan's Brynjolfsson. "They will be relatively close to commodity provisions. It'll be in the value-adding services that go on top of that." He predicts a churn of companies offering services. Even if a company comes to control a market, its dominance will be tenuous, because "if somebody else can do it just a little bit better, they can take over the market quite quickly," he says. "Cloud computing makes it much easier for somebody to do that."

Fellows agrees that public cloud computing will be dominated by a few big players, but because so many businesses will be running private clouds, he thinks there is the possibility of coöperation between them. He believes that these owners could even compete with the big public players by opening up their private clouds to public use for specialized services--essentially, cloud bursting in reverse. Companies will have opportunities to act as brokers between clouds--providing easy ways to switch loads between providers, for example. There are even companies "setting up now to provide exchanges where cloud capacity can be traded like futures," he says. But in that case, "these folks are ahead of their time."