A dynamic model for the optimal selection of secondary indexes

摘要

A model is presented which determines the optimal degree of secondary indexing for data processing requirements which follow variations over different time periods in a manner known in advance. The paper proves a number of properties which characterize the model and develops an algorithm which greatly reduces the solution space to be searched. The algorithm uses dynamic programming techniques and it may be used by Database Administrators to determine the time at which the entering and dropping of secondary indexes to and from the database should take place.