Overview
The ScaleBase Database Load Balancer™ performs load balancing between a database application and its backend database instances in the same way as any standard web-based load balancer.
But database load balancing is more than simple HTTP load balancing! ScaleBase partitions the database data into multiple instances (a technique known as sharding) and then directs clients’ requests to the relevant instances, achieving a shared-nothing database architecture. In this Cloud Computing architecture, each database and instance is independent, and there is no single point of failure across the system. Request results are aggregated and the client is presented with a single unified view of the requested data, thus ensuring that the number of configured database instances and their statuses are transparent to the application. This sharding technique has proven itself in some of the world’s most popular web sites. However, until now, sharding was typically considered too complex to implement for regular database applications.
The ScaleBase solution serves as both a database front end and a database manager, as depicted below:
On the front-end side, the ScaleBase Load Balancer serves as a database proxy that mimics the behavior of a regular database by supporting the same network protocol as your existing server (MySQL, Oracle Database, Microsoft SQL Server or IBM DB2).
On the back-end side, the ScaleBase Load Balancer™ acts as a database manager to distribute data between multiple database servers and to ensure the most effective usage of existing hardware. Since each database contains only a portion of the data, the ScaleBase Engine is responsible for directing SQL commands to the correct database server, aggregating results from the database servers (if necessary) and returning a single result set to the calling application.
Database Scaling
The ScaleBase architecture is based on share nothing servers, which enables its database cluster to achieve linear scalability. This means that it is always possible to add a new machine and that adding a new machine always improves the database performance.
Most applications can benefit from this handling of BigData, data scaling and data high availability. You can see how well ScaleBase performs when increasing the number of database servers:
High Availability
To achieve high availability, ScaleBase Load Balancer™ ensures that every piece of data is stored on at least two servers: one acting as the primary server and the other as a backup. When an instance fails, its backup instance springs to life and handles the data requests of the crashed database.
This active/passive solution provides the extremely MySQL high availability required by scalable applications.
What’s Next?
