基于机器学习的APM性能分析子系统
首发时间:2015-12-24
摘要:性能分析对构建高效率的计算系统意义重大,但随着云计算的流行,计算系统日益复杂,性能分析工作变得十分困难,下一代性能分析工具的设计开发迫在眉睫。本文基于机器学习算法设计了一种深入挖掘应用各项性能指标内在联系,统筹多指标联合定位应用问题的性能分析系统,解决了现有分析工具定位分析问题能力不足,适用场景有限的问题。
For information in English, please click here
Performance analysis subsystem in APM based on machine learning mechanisms
Abstract:Performance analysis is necessary to construct a computing system with high efficiency. However, as the complexity of computing system increases in the age of the cloud, it gets harder and harder to do accurate performance analysis. Therefore, it's urgent to design and implement an analysis tool of the next generatiron. This paper proposes a system based on machine learning mechanisms, which analyze the internal relationship among performance parameters. During the analysis of system exception, it will take all parameters into considration to localize the source of the exception. This system will solve the problems exists in current tools, such as lack of analysis abilitiy and not easy to be applicable.
Keywords: Performance Analysis Expert System Machine Learning Inference Knowledge
基金:
论文图表:
引用
No.4672004112568914****
同行评议
勘误表
基于机器学习的APM性能分析子系统
评论
全部评论0/1000