Amazon Web Services Hit By Power Outage

Outage affected companies relying on Amazon Web Services' U.S. East region.
Amazon's 7 Cloud Advantages: Hype Vs. Reality
Amazon's 7 Cloud Advantages: Hype Vs. Reality
(click image for larger view and for slideshow)
Amazon Web Services suffered a power outage in its Northern Virginia data center on Thursday evening, an event that disrupted service at several start-ups and Internet companies.

The Amazon Web Services Service Health Dashboard indicates that problems were first detected just before 9 p.m. PT on June 14. Affected services included Amazon Elastic Compute Cloud, Amazon Relational Database Service, and AWS Elastic Beanstalk served out of Amazon's U.S. East region.

"We can confirm a portion of a single Availability Zone in the US-EAST-1 Region lost power," Amazon said in a dashboard status report. "We are actively restoring power to the effected EC2 instances and EBS volumes. We are continuing to see increased API errors. Customers might see increased errors trying to launch new instances in the Region."

[ Read Apple's War Against Google: Time For New Tactics. ]

Amazon did not immediately respond to a request to provide information about the cause of the power outage.

Last August, a lightning strike in Dublin, Ireland, disrupted service at AWS's European zone data center. The lightning caused a transformer explosion at Amazon's electricity supplier and also knocked out Amazon's backup generators.

In April last year, an outage in Amazon's U.S. East region triggered what Amazon described as "a re-mirroring storm," in which the sudden loss of access to data prompted automated systems to try to duplicate data in a way that jammed network data traffic.

Heroku, Parse, Pinterest, and Quora were among the companies affected by the Thursday evening power outage. Heroku's incident report shows the disruption lasting eight hours, though most service appears to have been restored in two hours.

Amazon's Service Level agreement offers affected customers a service credit equal to 10% of their bill if the Annual Uptime Percentage for the customer drops below 99.95% over the service year. However, the SLA excludes events "caused by factors outside of our reasonable control," among other things. So a power failure unrelated to AWS equipment would not be covered.

On Twitter, Amazon's competitors like Rackspace and Joyent took the opportunity to remind tweet-stream readers of cloud hosting alternatives. And developer Drew Tempelmeyer, making reference to the AWS price reduction earlier this week, quipped about AWS, "We've lowered prices and availability."

Even if service problems invite sarcasm and customer solicitation, AWS isn't hurting for business. Last week, AWS evangelist Jeff Barr noted that Amazon S3 is now storing over one trillion objects.

On Thursday, AWS announced a new free support plan, a service optimization assistant called Trusted Advisor, and chat support, among other things.

Private clouds are more than a trendy buzzword--they represent Virtualization 2.0. For IT organizations willing to dispense with traditional application hosting models, a plethora of pure cloud software options beckons. Our Understanding Private Cloud Stacks report explains what's available. (Free registration required.)