摘要
传统异常检测技术受到冗余数据影响,检测准确率较低,存在一定片面性,由此,本文提出云环境下异常波动状态协作检测方法。分析云计算大规模、虚拟化、高可靠性等特征,据此建立具有部署简单、方便维护、延时较小等优势的云平台集中式监控结构;其次,将准确率、灵敏度与误报率确定为异常波动状态检测指标,利用小数定标方法对数据进行标准化处理,统一数据属性,提取融合后数据的异常特征参数,为数据融合奠定基础;在信息熵分析与非参数CUSUM算法协作基础上判断异常状态,确定异常波动状态,完成检测。仿真结果表明,所提方法检测准确度高,虚警率低,在理论创新的同时具有较强实用价值。
Traditionally,the abnormal detection technology was affected by redundant data,so the detection accu⁃racy was always low.Therefore,this article proposes a method of collaborative detection for abnormal fluctuation state in cloud environment.Based on the analysis of large-scale characteristic,virtualization,high reliability of cloud com⁃puting,a centralized monitoring structure of cloud platform with simple deployment,convenient maintenance and small delay was established.After that,the accuracy rate,sensitivity and false alarm rate were determined as the de⁃tection indexes of abnormal fluctuation state.Moreover,the decimal scaling method was adopted to standardize the da⁃ta and thus to unify the data attributes.Then,the abnormal characteristic parameter of the fused data was extracted to lay the foundation for data fusion.On the basis of cooperation between information entropy analysis and non-paramet⁃ric CUSUM algorithm,the abnormal state was judge to determine the abnormal fluctuation state.Thus,the detection was completed.Simulation results show that the proposed method has high detection accuracy and low false alarm rate.Meanwhile,this method has practical value and theoretical innovation.
作者
施永军
高祥斌
SHI Yong-jun;GAO Xiang-bin(Guangdong University of Petrochemical Technology,Guangdong Maoming 525000,China;Linyi University,Linyi Shandong 273400,China)
出处
《计算机仿真》
北大核心
2020年第9期390-394,共5页
Computer Simulation
基金
云计算环境中面向状态的云智能运维协作支撑平台研究(MM2017000011)。
关键词
云计算环境
异常波动状态
协作检测
数据融合
特征提取
Cloud computing environment
Abnormal fluctuation state
Collaborative detection
Data fusion
Fea⁃ture extraction
作者简介
施永军(1974-),男(汉族),江苏江都人,硕士,实验师,研究方向:网络安全/信息化;高祥斌(1980-),男(汉族),山东临沂人,讲师,研究方向:计算机科学与技术。
