A Bitmap Index for Multidimensional Data Cubes

Abstract

Multidimensional arrays can be used to store multidimensional data cubes. One way of storing such array onto disk is to partition the array into a set of small chunks that can fit in a disk block. Chunks are usually of the same side length and are compressed before being stored, in order to achieve high query processing speed and space utilization. This scheme is called the chunk-based cube storage scheme. Although the scheme is considered to be a position based storage scheme, the chunks cannot be retrieved fast without using indexes since they are compressed. In this paper, we propose a bitmap index for such a storage scheme. The index can be constructed with the data cube in parallel. The relative positions of the chunk retrieval can be done in constant time. We placed in an index block as many chunks as possible so that the number of index searches is minimized for OLAP operations such as slice, dice, and range queries. We evaluated the proposed index by comparing it with existing multidimensional indexes such as UB-tree and grid file in terms of time and space.