Grid computing environments include heterogeneous resources shared by a large number of computers to handle the data and process intensive applications. The required resources must be accessible for Grid applications on demand, which makes the resource discovery a criti More
Grid computing environments include heterogeneous resources shared by a large number of computers to handle the data and process intensive applications. The required resources must be accessible for Grid applications on demand, which makes the resource discovery a critical service in Grid environments. In recent years, diverse techniques are provided to index and discover the Grid resources. The response time and message load during the search process highly affect the efficiency of resource discovery. This paper proposes a new technique to forward the queries based on the resource types which are accessible through each branch in hierarchical Grid resource discovery approaches. The proposed technique is simulated in GridSim and the experimental results indicated that it is able to reduce the response time and message load during the search process especially when the Grid environment contains a large number of nodes.
Manuscript profile
Grid is a new generation of distributed networks and allows users to share files like the Internet. With regard to the specific features of Grid environments, such as high dynamicity and resource/members heterogeneity, there are some challenges dealing to it. One of the More
Grid is a new generation of distributed networks and allows users to share files like the Internet. With regard to the specific features of Grid environments, such as high dynamicity and resource/members heterogeneity, there are some challenges dealing to it. One of the most important services in grid environments is the resource discovery. The purpose of resource discovery is to identify a list of available resources for assigning to tasks. In this paper, using the assignment of prime numbers as the weight for tree nodes, and combining the hierarchical and super-peer structure, a new algorithm is presented with multiple trees. The results of the experiments and comparison with the previous methods indicate the improvement of the proposed method in terms of the number of visited nodes during the search process and the reduction of processing load caused by the smaller number of weights in the indexing tree.
Manuscript profile
