A Semantic Search Engine for the Storage Resource Broker

Information discovery is looming as a major challenge with the growth of tera-byte size datagrids. In order to manage their distributed data collections, many scientific organizations are adopting San Diego SuperComputer's Storage Resource Broker (SRB). Indexing and retrieval of data stored in SRB is via SRB's Metadata Catalogue (MCAT). MCAT focuses primarily on system or administrative metadata but supports domain-specific metadata through user-defined extensions. Although this approach provides maximum flexibility, it will lead to interoperability problems when searching across distributed collections described using different user-defined metadata schemas. The aim of the work described in this paper is to semantically augment SRB through an ontology and Resource Description Framework (RDF) descriptions in order to support arbitrary metadata schemata and to enhance the system's search capabilities. In particular we describe a semantic search engine and interface built on top of an OWL ontology, RDF instance data and a Jena reasoning engine that enables easier and more sophisticated searching of heterogeneous data stored using SRB.