Benchmarking Coding Algorithms for the R-tree Compression

Multi-dimensional data structures have been widely applied in many data management fields. Spatial data indexing is their natural application, however there are many applications in different domain fields. When a compression of these data structures is considered, we follow two objectives. The first objective is a smaller index file, the second one is a reduction of the query processing time. In this paper, we apply a compression scheme to fit these objectives. This compression scheme handles compressed nodes in a secondary storage. If a page must be retrieved then this page is decompressed into the tree cache. Since this compression scheme is transparent from the tree operations point of view, we can apply various compression algorithms to pages of a tree. Obviously, there are compression algorithms suitable for various data collections, therefore, this issue is very important. In our paper, we compare the performance of Golomb, Elias-delta and Elias-gamma coding with the previously introduced Fast Fibonacci algorithm.