An architecture based SKOS ontology for scalable, efficient and fault tolerant grid resource discovery

Abstract

Grid computing allows coordinated and collaborative resource sharing across several institutions, to solve large scientific problems that could not be easily solved within the boundaries of a single institution. To take advantages of all these resources, which have a heterogeneous and dynamic nature, an efficient resource discovery mechanism is required. As the number of nodes is constantly increasing, scalability and efficiency are highly desirable. Hybrid techniques are considered among the solutions that guarantee the efficiency and the scalability at the same time. This paper introduces a semantic clustering of nodes based on their domains of interest to form groups called federations. It presents a new process for constructing the federations and the three layers overlay network, based on Semantic Web technologies that use a SKOS lightweight ontology for describing domains of applications in the grid, and routing queries to target federations in an efficient way. We present a scalable and fault-tolerant process to leaders' federations, which minimizes the overhead between leaders during the process of updating information about themselves.