On Monday, Amazon Web Services’ North Virginia data center experienced yet another outage, knocking many popular websites offline. By Tuesday morning, most of the company’s cloud computing services had been restored.
“Amazon Web Services experienced an outage on Monday that resulted in downtime for many popular web destinations and services. Issues began at a Northern Virginia data center that caused a chain reaction across the Internet spanning several hours,” reported Shawn Knight for Techspot. “It started out innocently enough as a few major websites and services like Github, Minecraft and Reddit were among the first affected. As the day progressed, more sites and services were added to the list. Before it was all said and done, Airbnb, FastCompany, Fliboard, FourSquare and Pinterest were all experiencing downtime.”
According to Jack McCarthy from CRN, “Amazon’s service dashboard reported the first service interruptions at just before noon PST with its Elastic Beanstock services, followed by announcements of service interruptions with its Management Console for Elastic Beanstock Services, Relational Database Service, ElasticCache, Elastic Compute Cloud and CloudSearch. At 3:15 p.m. PST, Amazon said CloudWatch service was restored, with service brought back to Amazon Management Console at 3:45 p.m. PST, while the other service remained out late in the afternoon.”
Ars Technica’s Lee Hutchinson said that the outage related to problems in Amazon’s Elastic Block Store (EBS) service, a critical piece of the company’s Elastic Compute Cloud (EC2). “Unfortunately, this is not the first time Amazon’s distributed cloud has run into Internet-crippling issues. There was a much more serious EC2 outage in April 2011, which again stemmed from EBS-related issues. These kinds of outages are a jarring reminder of the true nature of ‘the cloud’—it’s still just servers in data centers. The basic concept of cloud computing is to abstract the annoying physicality of things away from the user, but it’s not turtles all the way down. The service is only as available as the underlying technologies make it.” He concluded, “The takeaway for anyone considering cloud hosting for your website or application is that there is no magic pill, and just because something is ‘in the cloud’ doesn’t mean it can’t come crashing down to earth when there are problems.”
VentureBeat’s Sean Ludwig noted, “A member of hacker collective Anonymous claimed responsibility for yesterday’s outage, but Amazon flatly denied that an attack occurred. We believe Amazon, as the evidence suggests this person is routinely lying about attacking websites and services including GoDaddy and Facebook.”