Performance Optimization of Analysis Rules in Real-Time Active Data Warehouses

Analysis rule is an important component of a real-time active data warehouse. Performance optimization of analysis rules may greatly improve the system response time when a new event occurs. In this paper, we carry out the optimization work through the following three ways: (1)initiating non-real-time analysis rules as less as possible during rush hour of real-time analysis rules; (2) executing non-real-time analysis rules using the same cube at the same time interval; and (3) preparing frequent cubes for the use of real-time analysis rules ahead of time. We design the LADE system to get all the reference information required by optimization work. A new algorithm, called ARPO, is proposed to carry out the optimization work. Empirical studies show that our methods can effectively improve the performance of analysis rules.