Research On Key Issues Of Data Management In Grid Environment

Naming's transparency, position's transparency and protocol's transparency are the goal of data management in the grid environment , data management is to establish an environment which integrates data access, storage, transmission and maintenance. Facing the bottom it screens the various heterogeneous storage devices and data resources, for the upper it provides the general and reliable data accessing service, providing users with an integrated data management service and a high-performance data processing service.With the enhancement of education informationization, many scientific research and the teaching activities need to use each school's education resource. Presently, each university has massive, complex education resource, but most of these resources are heterogeneous, distributed, idle, isolated, therefore it hindered people's exchange and sharing seriously. The plan of education resource grid enables the national various universities' distributed, heterogeneous, autonomous resources to be well shared and integrated. How to easily visit the national various universities' mass education and teaching management resources, is the main study of the data management in education resource grid environment.This thesis has analyzed the current grid data management's advantages and disadvantages, compared several kinds middleware technologies used in data management, designed a grid data management middleware model combining the education resource grid's characteristic, and made the related research on the data management key issues. The experiment indicates that using the designed data management middleware model in the education resource grid's data management can provide a good solution to massive data transmission's efficiency issue and avoid the accessing bottleneck, increase reliability and consistency and so on . Specifically, the major works and innovative points of this thesis are as follows:1.According to the education resource grid's characteristics, a three step-P2P topology which suits the education resource grid is proposed, the structure not only has the high performance and the high reliability features, but also has the strong P2P scalability, self-organization and fault tolerance features. This enables the hierarchical distribution management of education resource.2.Given the grid data management middleware model ER-D3M, based on the ER-D3M model, this thesis discussed each service and strategy's application. This model provides users with an integrated data management service and a high-performance data processing service.3.Proposed the education resource integration service, it combines the education resource and the teaching management well, provides the unified data view for users, it is also the explore of the further education resource grid supported cooperative work. Proposed the aggregation strategy based on users'access similarity, this strategy can explore the inner connection between the users who share the resources, it also can reduce the transmission flow of education resource grid.4.Proposed the file remote browsing control service and the grid data transmission proxy strategy based on file monomer measurement, the former combines Web Service technology, has realized users remotely transparent browsing access, it has good compatibility with the existing OGSA system. The latter choices different transport protocols to transfer data through the transmission proxy model in view of the different file monomer quantity size, it has achieved the purpose of optimization transmission speed.5.Given the replica consistency algorithm based on replica directory index, this algorithm combine the trace label based consistency maintenance algorithm and the three-step-P2P education resource grid's structure, it has the advantages of less redundant information dissemination, less update messages, low-cost index storage and high updated coverage.6.The algorithm and strategy in grid environment need simulating in order to validate the accuracy and high performance. Firstly, it gives the implementation of the education resource integration service and the file remote browsing control service. Secondly, it carries the data transfer experiment using the grid data transmission proxy strategy based on the file monomer measurement, and obtains the good experimental result. Finally, according to some rules, the network architecture topology of education resource grid is generated by using the tools of BRITE, this thesis also makes some comparative simulation experiments between the consistency maintenance algorithm based on replica directory index and the other consistency maintenance algorithms, it verifies the superiority of the algorithm.