OPTASSIST: A RELATIONAL DATA WAREHOUSE OPTIMIZATION ADVISOR

Abstract

Data warehouses store large amounts of data usually accessed by complex decision making queries with many selection, join and aggregation operations. To optimize the performance of the data warehouse, the administrator has to make a physical design. During physical design phase, the Data Warehouse Administrator has to select some optimization techniques to speed up queries. He must make many choices as optimization techniques to perform, their selection algorithms, parameters of these algorithms and the attributes and tables used by some of these techniques. We describe in this paper the nature of the difficulties encountered by the administrator during physical design. We subsequently present a tool which helps the administrator to make the right choices for optimization. We demonstrate the interactive use of this tool using a relational data warehouse created and populated from the APB-1 Benchmark.