Amazon Apologizes for Server Outage, Offers Credit
Amazon Apologizes for Server Outage, Offers Credit

Amazon's Elastic Cloud Computing web service was still experiencing problems four days after the initial outage that brought down many web sites. On Monday morning, the EC2 dashboard noted instance connectivity, latency and error rates at the North Virginia facility at the heart of the outage.

On Friday, Indaba Music co-founder Jesse Chan-Norris told Billboard.biz the outage is unlikely to dissuade people from using cloud computing services. All services suffer interruptions at some point, he said. "I don't think it will be a doomsday thing. People will look at the cloud just like any other service." Indaba was down much of Thursday because of the Amazon service outage.

But as Amazon's problems continued over the weekend, experts debated the outage's impact and discussed how companies should plan for similar events in the future.

Many experts used the outage as a teachable moment. Forbes.com's CIO Central advised companies to include compensation for downtown in their cloud computing contracts. Research analyst Krishnan Subramanian encouraged companies to plan for failure. "If we want to take advantage of commodity servers based public clouds, there is no option but to code for failure."

The outage also brought light to the differences in resources companies can apply to cloud computing. Amazon customer Netflix, for example, was not affected by the outage because it "has taken full advantage of Amazon Web Services' redundant cloud architecture," a company spokesperson told the New York Times. A smaller company likely knows of the risks of not having such redundancies but may not have enough money and staff to do so. That's why George Reese wrote at O'Reilly that "if your systems failed in the Amazon cloud this week, it wasn't Amazon's fault."

In fact, as one expert explained, Amazon did not even violate its terms of service. "Because Amazon's problems were confined to [Elastic Block Storage] and [Relational Database Service] in the US East region, Amazon's SLA for customers affected by the outage was not violated," wrote Arnal Dayaratna at the blog Cloud Computing Today.