Clustering Spatial Data for Join Operations Using Match-based Partition

The spatial join is an operation that combines two sets of spatial data by their spatial relationships. The cost of spatial join could be very high due to the large sizes of spatial objects and the computation-intensive spatial operations. In spatial join processing, a common method to minimize the I/O cost is to partition the spatial objects into clusters and then schedule the processing of the clusters such that the number of times the same objects to be fetched into memory can be minimized. In this paper, we propose a match-based approach to partition a large spatial data set into clusters, which is computed based on the maximal match on the spatial join graph. Simulations have been conducted and the results have shown that, when comparing to existing approaches, our new method can significantly reduce the number of clusters produced in spatial join processing