Optimizing storage utilization in R-tree dynamic index structure for spatial databases

Spatial databases have been increasingly and widely used in recent years. The R-tree proposed by Guttman is probably the most popular dynamic index structure for eciently retrieving objects from a spatial database according to their spatial locations. However, experiments show that only about 70% storage utilization can be achieved in Guttman's R-tree and its variants. In this paper, we propose a compact R-tree structure which can achieve almost 100% storage utilization. Our experiments also show that the search performance of compact R-trees is very competitive as compared to Guttman's R-trees. In addition, the overhead cost of building a compact R-tree is much lower than that of a Guttman's R-tree because the frequency of node splitting is reduced signi®cantly. Ó 2001 Elsevier Science Inc. All rights reserved.