Optimizing Join Index Based Spatial-Join Processing: A Graph Partitioning Approach

A Join Index is a data structure that optimizes the join query processing in spatial databases. Join indices use pre-computation techniques to speed up online query processing and are useful for applications which require low update rates. The cost of spatial join computation using a join-index with limited buuer space depends primarily on the page access sequence used to fetch the pages of the base relations. Given the join-index, we introduce a suite of methods based on spatial-clustering to compute the spatial-join. The spatial clustering we employ is based on graph partitioning techniques. For all the methods we derive upper-bounds on the lengths of the page-access sequence. Experimental results with Sequoia 2000 data sets, on a sequential system, show that spatial clustering method outperforms the existing methods based on sorting and online clustering heuristics.