Ontology based Resource Discovery Approach in Grid Environment

Abstract

Grid is an emerging technology for enabling resource sharing and coordinated problem solving in dynamic multi-institutional Virtual Organizations. In the Grid environment, the resources may belong to different institutions, have different usage policies and pose different requirements on acceptable requests. One of the fundamental operations needed to support location-independent computing is Resource Discovery; it is process of locating relevant resources based on application requirements of a user. The description of a resource is essential for automated Resource Discovery and search, selection, matching, composition and interoperation, invocation and execution monitoring; different Middleware specify different rules for describing a resource. Hence, the information gathered from these diverse sources tends to be semantically heterogeneous and needs to be correlated. Efficient Resource Discovery needs uniform unambiguous Resource Description. To date there is no universal resource description language common to all state of the art Grid middleware. Different Grid middleware systems have different methods of resource description and it is not yet known how well these can interoperate. Hence, there is a need to utilize semantic matching of these resource descriptions. An Ontology can provide unambiguous, asymmetric resource description that can be easily shared across various Virtual Organizations. The expressiveness of Ontology makes matchmaking process more efficient by providing list of resources based on semantics as the result of Resource Discovery process. Ontology based Resource Discovery also helps to improve Quality of Service (QoS) of Grid System and provide transparent access of Grid resources which is major goal of any Grid system In this thesis work, a Resource Discovery Approach based on Ontology has been proposed and implemented to discover the resources similar to the ones requested by the user. Using this approach matchmaking process achieves better results than exact keyword matching prevalent in most of the Resource Discovery approaches used today.