The prolonged outage of Amazon.com's EC2 cloud service late last month could set back adoption of hosted services by giving some companies -- especially those that are on the fence about migrating to the cloud -- a strong argument for taking it slowly.
For other organizations, Amazon's problems reinforced an already-held belief that cloud services can't match an enterprise IT operation when it comes to meeting the technology needs of business or government entities.
More than a week after the days-long partial outage started on April 21, Amazon released a detailed 5,700-word postmortem and mea culpa that identified the culprit as a configuration error that occurred during a network upgrade.
"The recent outage confirmed, for us, that cloud services are not yet ready for prime time," said Paul Haugan, CTO for the city of Lynnwood, Wash. The city had been looking into Amazon's cloud offerings, he added.
"Cloud services [as a whole] need some more maturing and a much more hardened infrastructure and security model prior to our adoption," Haugan said.
Jay Leader, a senior vice president and CIO at iRobot, whose products include the Roomba vacuum cleaner, said the Amazon outage illustrates well the limitation of cloud computing.
"We don't use Amazon or any other public cloud services, and we won't, perhaps ever, or at least until there is much more transparency about where the data lives, who controls where it lives and when and where it moves, and lots of other things," Leader said.
Analysts noted that the Amazon incident further increased IT concerns about application availability and the security of cloud services.
"Obviously, these issues are very heightened right now and will continue to be so for quite a while in light of the outage," said Gartner analyst Kyle Hilgendorf. "Amazon portrays an aura of invincibility, whether intentional or not, and this outage is going to remind enterprise customers that nobody is perfect and increased due diligence is required."
Hilgendorf said IT managers remain most concerned about cloud security -- including whether user authentication and access control measures are adequate, how much access a service provider has to a customer's systems and data, and the potential for an accidental release of protected data.
The Amazon outage also reinforces the fact that there's no way to directly migrate customer data to another provider, said Arun Taneja, an analyst at Taneja Group. If a service goes down, the host company must return the data to its customer, which then must find another provider or revert back to storing it locally, he said.
Coupled with recent decisions by EMC, Iron Mountain, Cirtas Systems and others to shut down or scale back hosted storage offerings, the Amazon incident could spur a backlash against cloud storage services, even though they can offer good value, Taneja said.
Lucas Mearian contributed to this story.