BestConfig: Tapping the Performance Potential of Systems via Automatic Configuration Tuning
This addresses the challenge for system users who lack expertise in tuning numerous configuration parameters to optimize performance under specific workloads.
The paper tackles the problem of untapped performance potential in systems due to suboptimal configuration settings across varying workloads, presenting BestConfig, an automatic configuration tuning system that improves throughput by up to 430% for MySQL and reduces running time by up to 80% for Spark jobs.
An ever increasing number of configuration parameters are provided to system users. But many users have used one configuration setting across different workloads, leaving untapped the performance potential of systems. A good configuration setting can greatly improve the performance of a deployed system under certain workloads. But with tens or hundreds of parameters, it becomes a highly costly task to decide which configuration setting leads to the best performance. While such task requires the strong expertise in both the system and the application, users commonly lack such expertise. To help users tap the performance potential of systems, we present BestConfig, a system for automatically finding a best configuration setting within a resource limit for a deployed system under a given application workload. BestConfig is designed with an extensible architecture to automate the configuration tuning for general systems. To tune system configurations within a resource limit, we propose the divide-and-diverge sampling method and the recursive bound-and-search algorithm. BestConfig can improve the throughput of Tomcat by 75%, that of Cassandra by 63%, that of MySQL by 430%, and reduce the running time of Hive join job by about 50% and that of Spark join job by about 80%, solely by configuration adjustment.