Research Of Reflective Middleware For Data Replication Based On Dynamic Priority

Computational grid has been used to solve a wide area network distributed resource sharing, networking and interoperability issues, with the rapid development of data-intensive applications and the explosive growth of data, Data Grid emerged and become another focus of grid.It is a great challenge to the grid designers to ensure an effective access to such huge and widely distribute data set because of the hindering of high latencies of the Internet. Data replication technology offers high data availability, low bandwidth consumption and robustness by copying a large amount of data storage on different nodes. Most of the current research is base upon static or user-driven data replication techniques. But the Gird environment however, is highly dynamic. The resource s availability and network performance change constantly and data access requests also vary with each application and each user.Traditional "black-box principle" of Middleware has been unable to meet the current developments in the field of new technologies and reflective Middleware is gradually becoming the mainstream of the next generation middleware technology. Reflective middleware provides re-allocation of the system to through self-inspection and self- adaptation to achieve run-time flexibility and adaptability.In view of the data grid characteristics and the performance of Reflective Middleware, we introduce the concept of computational reflection to data replication technology and use self-inspection and self-adaptation of reflective middleware to help data replication system cope with the dynamic changes in the environment and solve the key issues facing the data replication. DPRM(Research of reflective middleware for data replication based on dynamic priority),which is a prototype system for data replication according to this theory,can be dynamically adjusted the action of the system according to data replication needs.