Archived: 30 January 2006
University of Adelaide Masters by coursework thesis, November 2004.
Supervisors: Paul Coddington and Andrew Wendelborn
The demand for sharing resources scattered on machines of dissimilar platforms gives rise to research on grid technologies. In this project, we investigate the design issues and requirements for developing data grids applications that enable access to large-scale datasets. The goal of this project is to construct a data grid in Adelaide as the Australian node in the International Lattice Data Grid (ILDG), which will allow high-energy physicists to share terabytes of particle physics simulation data. We are working with the ILDG Middleware Working Group (MWG) on designing a standard Web services interface to allow interoperation of the data grids among ILDG members. In this project, we focused on the development of a prototype for the Metadata Catalog (MDC), which is a Web service component proposed by the ILDG MWG. We first developed a program to produce metadata in Extensible Markup Language (XML) for the particle physics simulation data files generated in Australia, because those data files originally did not have structured metadata. The metadata XML files our application generated conform to the latest metadata schema, i.e. the QCDml1.1, specified by the ILDG Metadata Working Group. Then we constructed an XML database, i.e. Xindice, which supports adding, deleting, updating and querying the metadata XML documents. We investigated different approaches for providing a standard interface for querying the metadata, including Open Grid Service Architecture Data Access and Integration (OGSA-DAI) and Metadata Catalog Service (MCS). The OGSADAI is a software development kit (SDK), which aims to provide a standard interface for metadata, data and database on the grid. The recently developed MCS provides a set of Application Programming Interfaces (APIs), which extend the OGSA-DAI to specially facilitate querying on metadata. We finally used the OGSA-DAI to design the architecture of the ILDG node featuring the Metadata Catalog and the Replica Catalog, which are two Web service components proposed by the ILDG MWG.