Distributed discovery of semantic grid service using semantic clustering of service ontologies

Abstract

This paper presents a novel approach of distributed discovery of semantic grid services using semantic clustering of service ontologies. In a distributed environment like grid there is a necessity for decentralized distributed registries containing the advertisement profile of the large number of grid services. The grid services can be of any diverse field. Hence distributed discovery becomes a major issue in grid computing. To solve the problem we cluster the semantic grid services and derive on a local cluster index for every distributed registry peer and finally a global cluster index. The global cluster index will give the location of the registry which can contain the requested service. Then in the local registry of the peer the local cluster index is used for discovery process. Thus the discovery time is reduced. For finding semantic similarly between the services, sense refinement algorithm [2] is used. To cluster, a hierarchical clustering algorithm, CHAMELEON [24] is used. The performance analysis of the system is made to prove that this technique reduces the latency of retrieval of the service and the precision of retrieving the most appropriate service is also high.