Research On Metadata Management In Grid-DBMS

Nowadays many data grid applications need to manage and process a huge amount of data distributed access multiple grid nodes and stored into heterogeneous databases. Because of the data's distribution and heterogeneous, there is much difficulty for data grid applications to deal with the data. So, in order to manage and process a huge amount of data in the grid, we must introduce special software that can lighten the burden of grid applications to manage those grid data. That is what we called Grid-DBMS.Grid-DBMS is expanse of the DBMS technology. The aim of the Grid-DBMS is to hide many implementation details strongly connected with network, data replication and data fragmentation and gives the grid applications a global logical view. So, a"Grid-DBMS"is a system, which automatically, transparently, and dynamically reconfigures at runtime components such as Data Resources, according to the Grid state, in order to maintain a desired performance level. It must offer an efficient, robust, intelligent, transparent, uniform access to Grid-Databases.Metadata management, which is crucial for Grid-DBMS to manage grid data, has an important effect in the Grid-DBMS. There must be the support of Metadata management, if the Grid-DBMS wants to give grid applications a global logical view. The Metadata management utilizes the layered distribution structure, and provides both the metadata management and Data source discovery. In order to improve the response ability of the Grid-DBMS, we introduce Data source replication management. However, the Metadata management must maintain the consistency of Data source replication. As the grid is dynamic, the Metadata management must provide the way for Data source to join or leave the Grid. The Data source discovery is based on the layered distribution structure model. According to the model, we put forward log-based Data source discovery algorithm, which also takes the user management into account.