The RD-tree Allowing Data in Interior Nodes of the R-tree

The RD-tree is a dynamic spatial access method which is a variant of the R-tree. The major difference between the R-tree and the RD-tree is the use of the RD-tree's interior nodes, which contains not only index entries but also data entries whenever possible. When an appropriate data object is stored in an interior node, instead of a leaf node in the subtree, the expansion of the minimum bounding rectangles (MBRs) in the subtree can be avoided. Thus, the overlapping area between MBRs can be reduced. As a result, the number of disk accesses in the RD-tree can be improved, compared to the R-tree. From the experiment, it is shown that the RD-tree clearly outperforms the R-tree when minimum bounding rectangles of the data are markedly oblong