Geospatial Data Grid with Computational Capability

Grid-computing technology is used to solve large or complex computational problems on distributed computational resources belonging to multiple organizations regardless of their geographical locations. In this paper, we introduce an integration of the SAFORAH geospatial data grid, provided by the Canadian Forest Service, and the computational grids, offered by the University of Victoria and National Research Council Canada, using Globus Toolkit 4 (GT-4) as a common grid middleware. New GT-4 services, such as an Integration Service (IS), a Metascheduler Service (MS) and a Registry Service (RS) were designed and implemented. MS is a broker used to schedule jobs on many distinct computational clusters at the project partner sites. The RS is used to track the availability of resources at these sites. IS allows the data grid and computational grids to work together, so that users are able not only to access geospatial data directly from the SAFORAH data grid but also to create virtual image products on the fly using the computational grids. The new system is a prime example of sharing data and computational resources and collaborating between different organization and research communities. We envisage that the results of our system would be valuable to other projects that wish to combine large-scale data management systems with computational resources.