A high performance system for processing queries on distributed geospatial data sets

The size of many geospatial databases has grown exponentially in recent years. This increase in size brings with it an increased requirement for additional CPU and I/O resources to handle the querying and retrieval of this data. A number of proprietary systems could be ideally suited for such tasks, but are impractical in many situations because of their high cost. On the other hand, Beowulf clusters have gained popularity for providing such resources in a cost-effective manner. In this paper, we present a system that uses the compute nodes of a Beowulf cluster to store fragments of a large geospatial database and allows for the seamless viewing, querying, and retrieval of desired geospatial data in a parallel fashion i.e. utilizing the compute and I/O resources of multiple nodes in the cluster. Experimental results are provided to quantify the performance of the system and ascertain its feasibility versus traditional GIS architectures.