X-square: A hybrid three-dimensional bitmap indexing for XML document retrieval

XML is studied and used as a key technology in many research and applications today. Accordingly there is an increasing need for the efficient XML document retrieval. Bitmap indexing is an efficient technique for determining true and false fast and it has been used mainly for reducing search extent rather than retrieving data. BitCube, an existing three-dimensional bitmap indexing for XML document retrieval, constructs bitmap from the entire index., In case high volumes of documents are loaded in a single cluster, however, this causes significant performance degradation in memory usage and operation speed. xPlaneb which is another bitmap indexing technique solves this problem by reconstructing three-dimensional bitmap index of BitCube using linked list. xPlaneb, however, has a high memory usage problem compared with BitCube when low volumes of documents are loaded because BitCube index consists of small field of one bits. This paper proposes X-Square which is a hybrid of BitCube and xPlaneb. X-Square takes both advantages of BitCube and xPlaneb. Experimental results show that X-Square has better performance in memory usage than xPlaneb and BitCube although the operation speed is a bit worse than xPlaneb.