论文部分内容阅读
本文在比较现有异常检测技术的基础上,提出了一种虚拟机异常检测框架,并对框架内关键模块的技术实现进行了研究,包括虚拟机的性能指标收集方法和传输方式、基于K-means聚类算法对虚拟机划分检测域模块、基于PCA算法实现对虚拟机性能指标数据的降维的数据处理模块、基于LOF的异常检测机制实现虚拟机的异常检测发现模块。最后,在上述算法研究的基础上进行了有针对性的实验和分析。
Based on the comparison of existing anomaly detection techniques, this paper proposes a virtual machine anomaly detection framework, and studies the key modules in the framework of the technology, including the performance of virtual machine collection methods and transmission methods based on K- means clustering algorithm to divide the detection domain module into virtual machines, implement the data processing module to reduce the dimension of the virtual machine performance index data based on the PCA algorithm, and realize the anomaly detection detection module of the virtual machine based on the LOF anomaly detection mechanism. Finally, based on the above algorithms, we conducted targeted experiments and analyzes.