News

posted on 14 April 2008 16:57

Amazon is introducing persistent local storage to its EC2 cloud computing offering.

EC2 (Elastic Computing Cloud) is an Amazon facility wherebye users can requisition a virtual server in minutes and use it to run their own application code, uploaded to Amazon's S3 (Simple Storage Service). The server is a virtual server, running on Amazon's own IT infrastructure, and its capacity and number of instances can readily be scaled up and down, hence the 'elastic' title.

Now an EC2 application can be provisioned with raw block storage which is defined and mounted into an EC2 application instance as a storage volume. It is logically equivalent to an unformatted hard drive. The EC2 user can format it with whatever file system is desired or use it directly from within a database engine. You can even set up striping.

There can be any number of volumes mounted by an EC2 instance and they can range in size from 1GB up to 1TB. Volumes can be migrated from one EC2 instance to another but only one EC2 instance can mount a volume at a time.

Users can request that EC2 snapshots the volumes and stores the snapshot in S3. It is stored redundantly and available in multiple availability zones. It can be used as a backup or as a source for a new volume to be created. This is equivalent to restoring a snapshot and is the means for building fault-tolerant applications that can run inside availability zones and which use elastic IP addresses.

Amazon stresses that all this can be used with any kind of infrastructure, middlware or application. It is coupling virtual servers with virtual storage in the cloud.

It has been likened to creating a storage area network (SAN) in the cloud with one API call and then being able to clone it.

More information is available here.

EC2 persistent storage will become generally available later this year.

^{[Paul Roberts, news editor.]}