Call it the summer of outages. Following downtime from Microsoft's Outlook and Google's overall darkness for about five minutes, Amazon saw some trouble of its own on Monday. Some are estimating the 25-minute outage could cost the e-commerce giant millions of dollars in lost sales.
But it could also hurt the company's reputation, which could be more costly in the long term. This isn't the first time the e-tailer has seen outage issues with its Amazon Web Services. In January, a massive outage lasted 49 minutes.
Although some are speculating that the National Security Agency had something to do with it, most industry analysts aren't buying that argument. Still, there are so far no answers to the question of what caused the outage that resulted in visitors receiving error messages instead of product displays.
The Specter of Risk
We asked Rob Enderle, principal analyst at The Enderle Group, for his take on the Amazon (and Microsoft and Google) outage. He told us he's concerned about the e-commerce company's reputation for reliability with services.
"As Amazon continues to try to run up this cloud services business, it's critical that people be able to depend upon them," he said. "If they can't, it doesn't matter what the price is because the cost of being out of business is simply too high."
As Enderle sees it, Amazon shouldn't be having outages. That, he said, is because there should be enough redundancy in the system to allow the system to failover to a backup system and continue operating.
"The fact that Amazon continues to have outages suggests that the system isn't fully cooked yet and it does raise a specter of risk," Enderle said.
Room to Grow
Newsfactor reported on Monday that 40 percent of Internet disappeared when Google went down for four minutes. The outages demonstrated how much people rely on the search engine. Google didn't offer a clear explanation about what caused the outage.
Meanwhile, Microsoft fixed a problem that kept some folks from accessing Outlook.com. Microsoft blamed an unexplained "incident" on the outage.
Are the three unexplained incidents related? Enderle doesn't think so.
"These companies are trying to learn what kind of loading they can put on servers and what kind of stress the systems can take. This is indicative of the fact that they are still pretty young and until they can have a utility-like uptime people shouldn't expect them to have utility-like reliability," he said. "But eventually they will get there. The goal is to make these services utility-like so that they are on and you can depend upon them being on, short of a natural disaster."