After nine months of testing, Amazon Web Services is now offering its Aurora database as a full commercial service, positioning it as a high-performance alternative to the widely used open source MySQL database.
“For the majority of workloads, we’re recommending Amazon Aurora for database use,” said Matt Wood, Amazon Web Services general manager for product strategy. “We’re expecting customers will use Aurora as their default choice.”
The company claims Aurora can run five times as quickly as a standard MySQL deployment, while offering the reliability of a high-end commercial database.
Amazon unveiled Aurora last November as its Reinvent conference. It adapted the interface of MySQL but used its own back-end cloud technologies to give the system a powerful performance boost.
The software is easily scalable, according to Amazon, an advantage over in-house databases. An organization can start off with a small test deployment and ramp it up later to full production use. The database can then be replicated for further use, such as for analytics or mobile applications.
“You don’t have to worry about implementation or all the management or overhead that goes with scaling,” Wood said.
Amazon is promising greater than 99.99 percent availability. The service can automatically detect and recover from most failures, without the need to rebuild the database cache. The service replicates data across multiple nodes on the Amazon Simple Storage Service.
In an internal test, Amazon ran 100,000 updates a second and 500,000 selects per second on both Aurora and MySQL, and found Aurora ran five times as quickly, Wood said.
Aurora is one of the options offered through the Amazon Relational Database Service (DRS), and will now be Amazon’s recommendation for general purpose, full-scale database workloads.
AWS offers a hosted version of MySQL on RDS but will encourage new users to opt for Aurora instead. Users can easily deploy their MySQL scripts and queries on Aurora instances, according to Amazon. Customers using the RDS MySQL instance can convert their databases to Aurora using a tool added to the AWS Management Console.
Over 1,000 users have deployed the service in beta phase, including the Nasdaq financial exchange, Pacific Gas And Electric, and the Alfresco content management service provider.
With Aurora, Alfresco was able to run a database holding over 1 billion documents that was updated, on average, 3 million times an hour, according to Wood.