Enhancing Query Performance by Avoiding Negative Interactions

While optimizing database performance, parallel query technology can significantly improve the throughput and resource utilization of the database server. Interactions among different queries which run concurrently can have a great impact on system performance. However, conventional optimizing strategies of parallel query often only focus on individual queries rather than their interactions. In order to capture the impact of parallel query interactions, we propose an interaction-aware scheduling algorithm of parallel queries. Compared with other traditional algorithms, the experimental results demonstrate that our optimizing approach can effectively improve the database performance by avoiding the negative interactions. Finally, some future trends in this area are prospected and discussed.