On February 15 this year, Amazon S3, the "cloud" storage service that's part of the Amazon Web Services suite of infrastructure applications, failed. Web 2.0 entrepreneurs who had been attracted to AWS based on its promised reliability and low cost had their confidence shaken. Several lost revenue when the service seized up.
Last week at the Under the Radar conference, Amazon CTO Werner Vogels sat down to an interview with Robert Scoble. The discussion of course came around to the S3 outage, and Vogels explained what happened. It was, he says, a "provisioning" and "logical" problem. Translated: They didn't program S3 to handle the load they got. It has since been fixed. Amazon also recently upgraded its hosted computing service, EC2.
But while Vogels expressed unhappiness at the outage, he also believes that Amazon's cloud services are still more reliable than any collection of servers a budding Web start-up could marshal. While that may be true, that's not what companies who signed up for AWS signed up to hear. We think a simple mea culpa would have gone over better.