Send in the (MySQL) clones
Amazon takes relational to a higher plane, and cuts prices
Amazon is making it easier, and cheaper, to roll out copies of MySQL for very large websites across its cloud.
On Tuesday, the company announced an update to its Amazon Relational Database Service that will let users create and delete multiple Read Replicas of MySQL instances in minutes, via a point-and-click interface.
MySQL has been the top relational pick on the web for those building websites and hyperscale services - Twitter and Facebook are MySQL users, for example.
Read Replicas are spun up to help hyperscale customers maintain query times as demand for a particular service increases.
Once a Read Replica is created from a specified source Database Instance, subsequent updates to the source code will be automatically replicated to the Read Replica, Amazon said.
The ability to spin up Read Replicas is important on Amazon because MySQL can be deployed across the tapestry of zones that comprise Amazon's service. Amazon splits its compute and storage service across four geographic zones, or AZs.
Automatic replication of changes across what Amazon calls Multi-AZ should help ensure there's consistency across the different portions of the Amazon cloud.
Vice President of database services at Amazon, WS Raju Gulabani said: "Amazon RDS now offers advantages in scalability, multi-data centre availability, elasticity and ease of administration at a fraction of the cost of operating MySQL servers on-premise."
Amazon also said that it is cutting the price of big databases. The organisation said this was in order to help organisations operating on tight margins squeak by - and to stop them from hopping to rival clouds.
Quadruple Extra Large Database Instances - 68GB of memory - and Double Extra Large Database Instances - 34Gb of memory - have been cut by 15 per cent.
Amazon also offers IBM's DB2, Informix Dynamic Server, Microsoft's SQL Server Standard Edition, and Oracle 11g.
Managing MySQL in large, distributed server settings is difficult, and there's no easy off-the-shelf approach to achieving scale or performance.
Sharding is one answer, where a very large database is partitioned as a shared-nothing system across a number of servers to achieve the high-levels of performance. Many of the social networks using MySQL have had to build their own sophisticated approaches.
A start-up called Xeround offers what it claims is an approach that delivers MySQL performance and scale without the sharding. Xeround uses in-memory, instead.
The company last month announced beta availability of its MySQL in the Cloud service, built on the MySQL Storage Architecture. The architecture acts as a pluggable storage engine that lets you upload your data and replace your existing, on-site MySQL database.
Xeround's beta is available on EC2 and through Xeround.com, or as a virtual appliance. ®