Database Management Systems (DBMS) are complex software systems which are used in an increasing number of applications. To obtain the best performance for business relevant applications, e.g., Decision Support Systems (DSS). DBMS have to be tuned and monitored. In order to reduce the costs researchers and DBMS vendors have focused on self-management techniques for recent years. In particular, materialized views were used to pre-compute results in Online Analytical Processing (OLAP) systems to improve the performance, e.g., the query processing time. In this work, we propose an automatic selection of materialized views based on statistics about a workload that is represented in the Query Graph Model (QGM). Our approach allows the mapping of workload queries into a map graph and materializes views by using access statistics and cost functions. We evaluate our approach with the TPC-H benchmark and compare it to a commercial DBMS.
展开▼
