The recent problems experienced by the Windows Azure cloud-computing preview might have seemed like a case of Microsoft joining the party. Outages and other glitches have been relatively common in the industry as Amazon, Microsoft and others have opened up their data centers for smaller companies and developers to use for storage, computing and delivering online applications.

But those kinds of problems shouldn't be an inevitable part of cloud computing, says Doug Hauger, Windows Azure general manager. In a phone interview last week, Hauger reiterated Microsoft's explanation of the problem, which was triggered by an operating-system upgrade that caused "networking issues" and server failures. However, he said, cloud-computing customers should expect more.

"This is a community technology preview, so in that sense, I think we're learning," he said. "But it is my expectation that, if you're an enterprise business and you're betting on a cloud-computing platform, that bet you make should not include the need to expect an outage."

In technical terms, Hauger said he doesn't consider what happened to Azure to be an "outage," in that the data center didn't go completely dark. Storage remained available, and the problem didn't affect most applications that used Microsoft's recommended architecture, which essentially provided a natural backup. Microsoft called it a "malfunction."

But whatever you call it, "clearly, it was not good," Hauger said. "I'm not saying that it didn't happen, nor am I saying that it wasn't a bad thing. It was not good. And we learned a lot from it."