On the Optimal Selection of Multilist Database Structures

Abstract

The optimal selection of secondary indexes asks for the quantitative evaluation of the performance of a number of candidate secondary indexes in order to determine the particular combination of indexes which satisfies the anticipated user transactions at a minimal cost. Previous studies determine the optimal selection by assuming that the cost of satisfying a query using a secondary index is not affected by the existence of other indexes in the database. This assumption is realistic when the inverted file organization is used to organize secondary indexes. The main reason is that inverted files do not alter the size of the file. However, the assumption is not valid when the next most popular method for structuring secondary indexes is used, namely, the multilist database structures. This is so, because each multilist increases the size of the file. This paper studies the secondary index selection problem by making the assumption that the multilist organization is utilized to structure secondary indexes and develops a dynamic programming algorithm to solve it. The practical significance of the study lies in the fact that multilists can be easily implemented on network databases.