Research Of Linux-based Clustered Database System

Cluster system is composed by a set of computers which are inter-connected to provide a single whole computing resource. The goal of clustering technology is to implement large-scale, high performance application system. We discussed how to establish a database system based on Linux Virtual Server, which is a famous cluster project on Linux operating system. We first analyze he key technologies of Linux cluster system, espicially we discuses the Linux Virtual Server in details. Then we review the current popular database clusters such as Oracle9i RAC and IBM UDB. Basing on our research, we propose a database system prototype, which is established on Linux cluster and use the DBMS such as DM4 and Mysql as real database servers. DMC comprises three layered modules: load banlancer,data servers and data synchronization. Focusing on the key technologies employed when deploying a cluster system such as SSI, data synchronization and load balancing,this paper give a detailed disscusion and give its solution perspectivly.According the DMC model and basing on Linux virtual Server, the prototype system is implemented with C programming languages on Linux operating system. This system can implement SSI in three ways and can provide load balancing strategies in eight different means. It is a cost-effective system and caters for the requests of current application such as WebAPP,DSS(Decision Supporting systems) and OLTP(Online Transaction Processing).