Amazon develops MySQL-compatible database engine Aurora

Amazon has released a relational database engine on its AWS cloud platform that would make MySQL databases perform better and scale more easily. The engine, called Aurora, is not open source, but it is compatible with MySQL 5.6.

According to Amazon, Aurora can perform up to five times better than standard MySQL implementations running on its AWS cloud platform, while the database engine remains fully compatible with Oracle’s open source database software. The performance gains are due to virtualized storage systems based on SSD storage, along with software optimizations in the engine.

Amazon promises that Aurora will also provide highly reliable database services thanks to fault-tolerance mechanisms and “self-healing” storage systems. This would keep databases available if disk errors occur. Amazon also claims that a database crash is automatically detected and that Aurora is available again within 60 seconds without having to perform a recovery procedure or rebuild the database cache. Furthermore, it is possible to automatically switch to a backup server if an entire Aurora instance crashes. Database data is replicated up to six times within three AWS regions.

An AWS user can activate and manage Aurora through their AWS management console. According to Amazon, a system administrator needs to perform very few tasks himself. For example, a database of a few gigabytes could grow to terabytes without the database having to be temporarily taken off the air to make more storage space available.

Aurora is not open source and is currently in preview for AWS users only, alongside existing platforms MySQL, PostgreSQL, Oracle, and Microsoft SQL Server. Like other AWS cloud services, customers pay based on the actual use of the database servers.