High available database system: With TGPMS as an example

Dealing with the service aspect of a system as an unbroken whole, the availability of database management system (DBMS) is more and more critical as the base of information system. In this paper, we present an overview of the distributed DBMS and parallel DBMS technologies, highlight the characteristics related to system availability of each and try to give out the theoretical fundamental of high available DBMS. By introducing the features of the clustering technology, and discussing the real implementation of TGPMS DBMS, it is instructive for us to optimize system design to get high availability.