Heuristics-Based Workload Analysis for Relational DBMSs

Database systems are widely used in heterogeneous applications. However, it is difficult to decide which database management system meets requirements of a certain application best. This observation is even more true for scientific and statistical data management, because new application and research fields are often first observed in this domain. New requirements are often implied to data management while discovering unknown research and applications areas. That is, heuristics and tools do not exist to select an optimal database management system. We develop a decision support framework to support applicationperformance analyses on database management systems. We use mappings and merge workload information to patterns. We present heuristics for performance estimation to select the optimal database management system for a given workload. We show, these heuristics improve our decision framework by complexity reduction without loss of accuracy. Finally, we evaluate our heuristics considering standard database benchmarks.