ABSTRACT

Spatial Database Systems are very large databases requiring storage in disk devices. The join process between two relations that is based on spatial nature of the attributes needs repeated and redundant disk accesses. This necessitates the implementation of join index which accelerates the join processing between two relations and it incurs high over head. This work studies the heuristics that are being in use for join processing from the context of optimal page access of disk I/O and that of spatial join processing with join index. The study presents a summary of heuristics that are implemented for joinprocessing of spatial relations. Spatial-Queries modeled with spatial statistical measures provide an indication of spatial clusters. This model with an appropriate clustering algorithm identifies clusters that would require page re-accesses. An Optimal Page Access Sequence can reduce the page re-accesses and thus bring down the I/O cost of the join process. The work evaluates the performance of the statistical measure Moran’s I to model spatial queries and identified to have a better performance than the existing measure Ripley’s K function.