Research On Layered And Semantic-based Model For Grid Resource Discovery

Grid, an integrated computing and resource environment, makes it possible for geographically distributed heterogeneous resources to link up through networks and then provide users with all kinds of information. The resource discovery mechanism is one of the most important components in grids. Although there has been many discovery mechanisms presented so far, the majority of them adopt an index-like structure which is centralized and/or text-based. Therefore, not only are these traditional mechanisms inaccurate in searching, but also they have been the performance bottleneck in the process of resource discovery.In this paper, we presented a layered and semantic-based model for grid resource discovery to solve these problems mentioned above. This discovery mechanism is based on semantics in that it introduces RDF (Resource Description Framework) to describe grid resources. We also build upon this a set of query routing and node clustering mechanisms, thus enabling the resource discovery process to be smarter. Moreover, this discovery model is a kind of layered model. We divided grid nodes into different types according to the different roles they play in grids. Then the model is logically formed by two layers. They are resource aggregation layer and resource distribution layer. The resource aggregation layer is located in the virtual organizations, using centralized discovery mechanism. The resource distribution layer is built upon virtual organizations where the nodes are fully decentralized just like P2P systems. In the last part of this paper, we simulated a tourism-oriented grid system and performed some experiments on it.The model and solutions this paper presented will not only be valuable in practical use, but also provides a broader space for improvement in grid resource discovery technologies.