Performance Improvement of Distributed Systems by Autotuning of the Configuration Parameters

• Published in: Tsinghua science and technology Vol. 16; no. 4; pp. 440 - 448
• Main Author: 张帆 曹军威 刘连臣 吴澄
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.08.2011; Tsinghua National Laboratory for Information Science and Technology, Beijing 100084, China%National CIMS Engineering and Research Center, Tsinghua University, Beijing 100084, China; National CIMS Engineering and Research Center, Tsinghua University, Beijing 100084, China%Research Institute of Information Technology, Tsinghua University, Beijing 100084, China; Tsinghua National Laboratory for Information Science and Technology, Beijing 100084, China
• Subjects: autotune configuration parameters; covariance matrix algorithm; distributed systems; ordinal optimization; performance evaluation; 分布式系统; 分布式计算系统; 性能改进; 测试时间; 网络配置参数; 自动调整; 进化策略; 配置文件; performance evaluation; autotune configuration parameters; covariance matrix algorithm; distributed systems; ordinal optimization
• ISSN: 1007-0214; 1878-7606; 1007-0214
• DOI: 10.1016/S1007-0214(11)70063-3

The performance of distributed computing systems is partially dependent on configuration parameters recorded in configuration files. Evolutionary strategies, with their ability to have a global view of the structural information, have been shown to effectively improve performance. However, most of these methods consume too much measurement time. This paper introduces an ordinal optimization based strategy combined with a back propagation neural network for autotuning of the configuration parameters. The strat- egy was first proposed in the automation community for complex manufacturing system optimization and is customized here for improving distributed system performance. The method is compared with the covariance matrix algorithm. Tests using a real distributed system with three-tier servers show that the strategy reduces the testing time by 40% on average at a reasonable performance cost.