Research And Application Of High-availability Database Cluster System

Database is the foundation of the electronic commerce, the E-government, MIS, ERP, as well as financial system, usually saving the important information. However, with the data quantity and the business volume fast growth, and the high concurrent visit's appearance, cause traditional the stand-alone database unbearable, so the system's performance and availability urgently need to be solved. The cluster technology is the first choice, the deep research database cluster technology, will be helpful in enhancing its performance and the availability, the high availability as an important measure of database system, this paper will focus on.The cluster has been applied into the multi-layer software architecture successfully, for instance the high concurrent visit causes the Web server bottleneck, already might be solved through it. But lacks this kind of tool in the database system, when the database presents the bottleneck, for most users, it is not a good idea to use expensive commercial product like Oracle RAC, they even more favorite in using free software. This paper analyzed the cluster technology based on the middleware and the factor of affecting the high availability, finally constructed a high performance and high availability database cluster system using the source software Sequoia, and has applied it into practice.High availability database cluster must have a good scalability, data consistency, load balance and the ability to eliminate single point failure. Open source middleware Sequoia through JDBC interface to introduce the concept of a virtual database, the applications using JDBC can be run in the cluster environment without any changes. It is a flexible architecture to support large and complex architecture, it can provide high-performance, strong fault-tolerant ability, easy-to-use platform. In this paper, analyzed the principle of Sequoia deeply, detailed description of the installation and configuration, and build a database cluster using Sequoia, Oracle and PostgreSQL, then following the consistency, failure switching, load balancing and other testing, the experimental results show that the cluster using Sequoia has a good scalability, data consistency, fault-tolerant as well as the ability to load balance, improved query speed, enhanced query performance, improved database system availability and could be applied to the actual field.Finally, this paper constructed a company sponsor business management system. According to project request, in order to reduce the cost, overall system's hardware is PC machine, through the high speed Ethernet connection, the development kit and database uses the source free software completely. The project uses Struts+Spring+Hibernate+Sequoia+PostgreSQL, contains the Web server cluster and the database server cluster, database cluster's architecture based on the above Sequoia middleware technology. This management system is developed under the Linux platform, and has been successfully applied in practice.