Self-tuning performance of database systems based on fuzzy rules

Self-tuning performance of the database management system is a challenging and time-consuming work since multiple performance parameters need to be automatically adjusted. In this paper, we present a new self-tuning method based on fuzzy rules. We first automatically extract Automatic Workload Repository (AWR) report from a database system, and then construct fuzzy rules by identifying the relationship among the performance parameters and performance indicators. Finally, we develop an adaption algorithm to reconfigure the performance parameters with the fuzzy rules. Experiments on TPC-C workload show that the proposed method is feasible and effective.