Amazon has announced a rebuilt version of MySQL created for the cloud, and engineered to be extremely low maintenance and easy to protect and affordable.

The announcement was made at Amazon’s re:Invent developer conference in Las Vegas. The new database, Aurora, has been created ‘from the ground up’, according to Andy Jassy, head of Amazon’s cloud services platform. Taking the stage at re:Invent Aurora's General Manager Anurag Gupta said the project has been underway for a number of years in secrecy and it is a complete redesign of the idea of a database that takes advantage of the scalability of AWS and its services, uses a service oriented, decoupled architecture and is fully managed.

Amazon RDS for Aurora is fully compatible with existing MySQL 5.6 applications and tools, but five times faster on the same hardware. Amazon is still offering MySQL in its RDS lineup, adding Aurora as an alternative that’s self-maintaining and fault-tolerant.

Discussing the new database in a press conference at Re:Invent, Jassy said;

“Lots of enterprises are using MySQL, and they would like to put more workloads on MySQL, but it is very challenging to get the kind of performance they want from these sorts of open source database engines.”

On the Amazon Web Services blog, Jeff Barr says that all the databases Amazon has until now offered on its RDS relational database service (MySQL, Oracle, SQL Server, and PostgreSQL), were:

“designed to function in a constrained and somewhat simplistic hardware environment -- a constrained network, a handful of processors, a spinning disk or two, and limited opportunities for parallel processing or a large number of concurrent I/O operations”.

Barr says that the RDS team decided to take a fresh look at the problem and to create a relational database designed for the cloud, with an integrated design encompassing the storage, network, compute, system software, and database software. This should eliminate bottlenecks caused by I/O waits and by lock contention between database processes.

Aurora is built on an SSD-based virtualized storage layer that is purpose-built for database workloads. Barr says that if you need to add storage, there’s no need to take your application off line.

“Instead, Amazon Aurora will add storage in 10 GB increments on as as-needed basis, all the way up to 64 TB. Baseline storage performance is rapid, reliable and predictable—it scales linearly as you store more data, and allows you to burst to higher rates on occasion.”

Aurora is designed to be completely compatible with MySQL in areas such as table structures and SQL calls. However, alongside the faster performance, Aurora is designed to be easier to administer. There’s no need to manage data replication as six copies of the data are automatically replicated across three Availability Zones. Hardware failure is also less of a problem as the data is backed up to the Amazon Simple Storage Service (S3) rather than needing to be backed up using snapshots. Aurora detects any database failures and transparently recovers with instance restarts typically requiring less than a minute. If the failure is permanent, will automatically failover to a replica without losing any data.

Among Aurora's benefits, outlined in the announcement at re:Invent, is that it provides similar performance and availability as high-end commercial database offerings, but at one tenth of the cost with no minimum commitment or up-front fees. Pricing is based on what you actually use, starting at 29 cents per hour for 15.25GB of memory and two virtual CPUs. You then pay separately for storage at 10 cents per gigabyte per month, and I/O at 20 cents per million requests. 

Aurora is currently available as a free-to-use technology preview. Sign up to join the waiting list here.

Comments

Make a Comment or View Existing Comments Using Disqus

or email your comment to: comments@i-programmer.info