The Study On The Model Of Grid Information Service Based On Tapestry

With the rapid development of Grid technologies, the study on Grid information service(GIS) is an emphasis and difficulty. The goal of the GIS is to discover, character and monitor resources of Grid. In the study of GIS, there are three models: Monitor and Discover Service(MDS), the Relational Grid Monitoring Architecture(R-GMA) and RGIS. The three models can do well in the tasks of GIS, but they meet the same question: how to organize the index system so that it can possess self-organization , fault-tolerance , scalability attributes, and have the abilities to efficiently handle multi-attribute based queries and the updating of dynamic attributes. With the development of Grid , it's scale is larger, the number and the type of resource, the communities and people involved in the Grid are more, the speed of change becomes more rapid. So those attributes are necessary.Grid is a method of distributed computing . And another method is Peer-to-peer technology(P2P).There are some commonnessesand differences between them. In brief, we argue that (1) both are concerned with the same general problem, namely resource sharing within virtual communities; (2)both deal with the problem by logical network; (3)each has made genuine technique advances, but each also has in current instantiations-crucial limitations, which we characterize as "Grid computing addresses infrastructure but not yet failure, whereas P2P addresses failure but not yet infrastructure". The complementary nature of the two approaches suggests that we can use P2P technology to supply the difficiencies of Grid.Based on the above discussion, the paper builds a GIS model based on a P2P model-Tapestry It can do well in calability , self-organization , fault-tolerance, and improves the surrogate routing algorithm of Tapestry . Besides, the paper suggests two approaches to efficiently deal with the updating of dynamic attributes, and gives out the solution for multi-attribute based queries.We hope that our work can benefit the development of Grid information service.