Home // ICSEA 2011, The Sixth International Conference on Software Engineering Advances // View article

A maintenance Approach of a BJI Index Configuration

Authors:
Saïd Taktak
Jamel Feki

Keywords: data warehouse; bitmap join indexes; tuning.

Abstract:
In data warehousing domain, OLAP (On Line Analytical Processing) queries are complex since they use several tables with huge cardinalities. Several optimization techniques have been studied in the literature as materialized views and bitmap join indexes (BJI). BIJ indexes are useful to pre-calculate star joins in order to reduce the execution cost. Current approaches for the selection of BJI define a configuration that optimizes a beforehand definite workload of queries. However, this workload can evolve in time and is likely to make obsolete the configuration of index created. In order to take into account the evolution of a workload of queries, we propose, in this article, a maintenance approach for the recommendation of a new configuration of indexes. Our approach starts with an evaluation of the current configuration of indexes and then adapts it to the new workload of queries with an aim of guaranteeing the stability of performances. Queries of the new workload are directly extracted from log files. Furthermore, to validate our approach, we carried out a series of experimentations on a data warehouse created with the DWEB benchmark.

Pages: 221 to 226

Copyright: Copyright (c) IARIA, 2011

Publication date: October 23, 2011

Published in: conference

ISSN: 2308-4235

ISBN: 978-1-61208-165-6

Location: Barcelona, Spain

Dates: from October 23, 2011 to October 29, 2011