Grid Information Service Based On Chord And Bloom Filter

Grid technology, known as the next generation application of the Internet, is used to make the nodes in the network interconnected. Its final purpose is to eliminate resource or information islands so that the information and resources can be fully shared in the network. Grid information service (GIS) is the core and one of the basic modules of grid system. It is mainly responsible for the resource registry, discovery and other functions in grid environment. Almost all the activities rely on the result of queried by GIS so the efficiency of GIS affects the performance of whole grid system to a great extent. The existing grid information service models have the problems such as the scalability of the system, the overload of the information server, the low efficiency of inquires. So it is necessary to build a new grid information service model to solve the problems mentioned above.On the basis of researching the related theories such as grid information service, P2P and Bloom Filter and so on; the paper analyzes the shortcomings of the existing GIS models in detail. A high efficient GIS model is proposed by integrating the DHT and Bloom Filter algorithms.The paper uses the following solutions to solve the shortcomings of the existing grid information service models.①Given the scalability and load balance of the system, Chord protocol is chosen in the proposed model to organize the information server nodes in grid. There is no need to adjust any arguments to achieve the scalability of the system. All the information server nodes share the load with same probability to realize the load balance of the system.②Each information server in the system uses the data structure -Bloom Filter to manage the registration information of resources. It can store a large amount of registry information with a small amount of space. The characteristic element matching pattern of Bloom Filter improves the efficiency of inquiries.③Each information server will store the Bloom Filter replicas of its successors in order to reduce the transmission time of the request in grid. The compressed Bloom Filter algorithm is used to compress the Bloom Filter on an information server so that the size of update message and the network flow to update the replicas would be minimized. The simulation results show that the proposed model is better than the existing GIS model in resource matching rate and consuming price rate. The model improves the efficiency of information server on matching resources and giving the feedback of no-result resources to users. The model reduces the transmission time of the requests and improves the efficiency of the whole grid information service.