Research And Implementation Of A Data Management System In A Service Grid

With the development of grid technology, it appears that science data increase explosively. At the same time, the users distributed across the globe want to access and analyze these vast data. Because of the complication of calculation, the analysis of huge volumes of data needs billions or more times computing power, but current architectures and solutions of data management can't satisfy the requirements of users in high performance, distributed storage and distributed process. Therefore, people are looking for an effective technique in vast data storage and management.In this thesis, we introduced how to construct a data management system in the Service Grid. Our research work is a part of the study on enterprise flexible composition in service grid; one of the nation 863 projects, and the data management system is an important component of Enterprise Flexible Composition Services (EFCS). The functions of our data management system that is similar to the data grid include managing multiple distributed and heterogeneous storage resources, providing a transparent metadata access for users.According to the requirement of EFCS system, we design a prototype system named Data Management System (EFCS-DMS) by researching existing data management systems applied in Grid. The DMS provide a platform with which users can transparently access the metadata. There are four layers in the EFCS-DMS, including data layer, interface layer, service layer and application layer. The service layer, as the kernel of DMS, is component with four services that include security service, data service, metadata service and global namespace service.