Performance Improvement of Distributed Systems by Autotuning of the Configuration Parameters  

作　　者：张帆 曹军威 刘连臣 吴澄 

机构地区：[1]National CIMS Engineering and Research Center, Tsinghua University [2]Research Institute of Information Technology, Tsinghua University [3]Tsinghua National Laboratory for Information Science and Technology

出　　处：《Tsinghua Science and Technology》2011年第4期440-448,共9页清华大学学报（自然科学版（英文版）

基　　金：Supported by the National Natural Science Foundation of China(No. 60803017);the National Key Basic Research and Development (973) Program of China (Nos. 2011CB302505 and 2011CB302805);supported by 2010-2011 and 2011-2012 IBM Ph.D. Fellowships

摘　　要：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.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.

关 键 词：distributed systems performance evaluation autotune configuration parameters ordinal optimization covariance matrix algorithm 

分 类 号：TP338.6[自动化与计算机技术—计算机系统结构]