Distributed And Parallel Database System Dpsql Intelligent Remodeling, And Implementation

Along with the great development of the Internet, E-business and enterprise informitition, the requirement of reliability and availability of the database system is increasing. Among the database systems, DPDBS (Distributed Parallel Database System) is the ideal choice due to its distributed data and parallel processes. In the research of DPDBS, computer designers constantly study the arithmetic and technologies that can enhance the reliability and availability, and the research of system reconfiguration is progressed accordingly. In DPDBS, on the one hand, each node in the system joins and exits random because of its autonomy, on the other hand, the shortcoming of software and hardware of node itself, Internet and natural factors may result in the fault forecast not available. Therefore, to reach the high reliability and availability of DPDBS, the problem of system reconfiguration should be settled commendably. How its realization, good or bad, is the determined factor to the availability of data resources of system. To reach the high performance of system, especially the availability, the system reconfiguration of DPDBS should be designed not only ensure the system run normally after some nodes fail or join, but also balance the load of each node in the system. A great many of technologies and theories have been produced in the research course of system reconfiguration, but all of them are limited to the failure of single node. In system DPDBS, system reconfiguration can't be limited as that because the state changes of multiple nodes are inevitable. So, DPSR(Distributed Parallel System Reconfiguration), one intelligent system reconfiguration mechanism based on dynamic data redundancy, is presented. DPSR is applied in DPSQL, the first practicality of DPDBS developed by our lab. In DPSQL, DPSR finishes the automatic reconfiguration of data resources through distributed collection of databases information and by the way of "rebirth" and "transfer" of data to reach high availability, and reconfigures all existing nodes to an organism to implement high reliability while the node number is changing. The application indicated that, DPSR could solve the state changes of multiple nodes in DPSQL and ensure the system reliability and data resources availability. Accordingly, we can say that it is the key foundation of technology to reach high reliability even never breakdown system.