Amazon.com announced today a new online database service, the latest addition to its hugely successful Amazon Web Services suite of online computing resources.
Called DynamoDB, it draws on the work Amazon has done for years to develop databases for its internal businesses.
“It’s not database software, it’s a database service,” Chief Technology Officer Werner Vogels said during a webcast announcing the service.
Offering a NoSQL database service was prompted in part by a suggestion two years ago from SmugMug, an online photo sharing service that was already using AWS for storage and computing. Co-founder Don MacAskill, who appeared on the webcast, explained that his company had to invest heavily to build, scale and monitor its databases and wanted a service that would remove this administrative overhead.
“We really just want to give you data and at some point in time ask for it back,” MacAskill said.
SmugMug helped Amazon test DynamoDB last year in a private beta test. Amazon also tested the service itself, moving its consumer Cloud Drive service onto DynamoDB, where it was used to handle a traffic spike over the holidays. Cloud Drive functions as the online hard drive for devices such as the Kindle Fire, so users of the company’s new tablet who stored files online helped test the new web service.
Amazon is running DynamoDB’s NoSQL database on solid-state disk drives and said the service should provide latency in the single-digit millisecond range for a 1 kilobyte object.
The company charges $1 per gigabyte stored per month. Data request pricing varies on the capacity requested; it’s charging a penny per hour for 10 units of “write capacity” or 50 units of “read capacity.” A unit equals one read or write per second for items up to 1 kilobyte.
Amazon is also offering a free introductory tier of service.