Modern database applications like geographic information systems, multimedia databases, and digital libraries dealing with huge volumes of high dimensional data, make use of multidimensional index structures. Among them,SR-tree has been shown to outperform the R-Tree and its variants and the SS-Tree. For static datasets, packed index structures provide better retrieval performance. This paper presents schemes for SR-Tree packing based on different pre-processing techniques. The results show that these schemes consistently outperform packed R-Tree and conventional SR-Tree structures in terms of storage space and query performance.
Jayendra Venkateswaran, S. R. Subramanya