基于小波分析的云计算在线业务异常负载检测方法
发布时间：2018-07-17 16:31
【摘要】：随着越来越多的在线业务被迁移到基于云的平台上,如何检测云平台上在线业务的异常运行状态成为了一个重要的问题。现有方法通过分析在线业务的实时负载数据来判断业务是否存在异常,在应对由程序异常或突发用户访问引起的异常负载时存在准确率低、误报率高的问题。该文提出并实现了一种面向云计算在线业务的异常负载检测方法。该方法利用小波分析技术,将原始负载数据分解成频率不同的多条曲线,并利用统计分析技术,通过检测各个频率上的异常增长或降低来判断负载是否存在异常。实验结果表明:同现有方法相比,该方法更准确,同时可以大大降低误报率。
[Abstract]:With more and more online services being migrated to cloud-based platforms, how to detect the abnormal running state of online services on cloud platforms has become an important issue. The existing methods analyze the real-time load data of online services to determine whether there is an anomaly in the service. There are some problems such as low accuracy and high false alarm rate when dealing with abnormal load caused by program exception or unexpected user access. In this paper, an anomaly load detection method for cloud computing online services is proposed and implemented. The method uses wavelet analysis technology to decompose the original load data into multiple curves with different frequencies. By using statistical analysis technology, the abnormal growth or decrease of each frequency is detected to determine whether the load is abnormal or not. The experimental results show that the proposed method is more accurate and can greatly reduce the false alarm rate than the existing methods.
【作者单位】： 清华大学计算机科学与技术系;
【基金】：国家国际科技合作专项项目(2013DFB10070)
【分类号】：TP274;TP3
本文编号：2130246
[Abstract]:With more and more online services being migrated to cloud-based platforms, how to detect the abnormal running state of online services on cloud platforms has become an important issue. The existing methods analyze the real-time load data of online services to determine whether there is an anomaly in the service. There are some problems such as low accuracy and high false alarm rate when dealing with abnormal load caused by program exception or unexpected user access. In this paper, an anomaly load detection method for cloud computing online services is proposed and implemented. The method uses wavelet analysis technology to decompose the original load data into multiple curves with different frequencies. By using statistical analysis technology, the abnormal growth or decrease of each frequency is detected to determine whether the load is abnormal or not. The experimental results show that the proposed method is more accurate and can greatly reduce the false alarm rate than the existing methods.
【作者单位】： 清华大学计算机科学与技术系;
【基金】：国家国际科技合作专项项目(2013DFB10070)
【分类号】：TP274;TP3
【相似文献】
相关期刊论文 前1条
1 彭敏;在线业务:反病毒行业先行一步[J];软件世界;2005年11期
，本文编号：2130246
本文链接：https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/2130246.html
