Amazon Outage Highlights Data Recovery Issues with Cloud Servers

The recent server outage suffered by sites using storage at Amazon has prompted further debate on the issue of data recovery from cloud-based server systems.

On Thursday 21 April, sites such as Nasdaq, Quora.com, Reddit.com, GroupMe.com and Scvngr.com all found their sites inoperable – a situation which they were stuck with for at least the following 24 hours. All the sites had their data and applications stored on Amazon's Elastic Cloud (EC2) systems, which experienced significant problems and caused the shutdown of some sites hosted there.

Amazon said the problems were caused by a networking issue which saw their systems creating too many backups and flooding their storage volumes. In the aftermath of Amazon's high-profile failure, many have expressed renewed concern at the vulnerability of cloud computing in terms of data recovery, as well as the security of data and a lack of clear service standards and service level agreements.

For a website using a more conventional hard disk RAID array system, most problems with storage drives can be swiftly addressed by simply swapping a new drive for any which are faulty. Even a non-RAID system could be swiftly brought back into service by engineers should any problems arise, making the option of a cloud-based server arrangement a risk in terms of the time needed to address faults.

Blog archive >>