| 本文已被：浏览 4715次 下载 8058次 |
码上扫一扫！ |
|
|
| 基于负载模式识别的Web应用在线异常检测方法 |
|
王焘1,2,3, 魏峻1,2, 张文博1, 钟华1
|
|
1.中国科学院软件研究所软件工程技术研究开发中心, 北京 100190;2.计算机科学国家重点实验室(中国科学院软件研究所), 北京 100190;3.中国科学院研究生院, 北京 100049
|
|
| 摘要: |
| 负载模式的动态变化会影响系统度量,使得异常难以准确检测.针对此问题,提出一种基于负载模式识别、在线检测Web应用异常的方法.该方法基于在线增量式聚类算法,运行时识别动态变化的负载模式,根据特定负载模式对应的度量空间,利用局部异常因数检测异常状态,并量化异常程度,并通过学生 t 测试方法计算度量异常值,以定位异常原因.实验结果表明,所提方法能够准确识别负载模式变化,有效检测出 Web 应用典型错误所引起的异常状态,并定位异常原因. |
| 关键词: Web应用 异常检测 动态负载 增量式聚类 局部异常因数 |
| DOI：10.3724/SP.J.1001.2012.04197 |
| 分类号: |
| 基金项目:国家自然科学基金(61173004); 国家重点基础研究发展计划(973)(2009CB320704); “核高基”国家科技重大专项(2011ZX03002-002-01) |
|
| Online Anomaly Detection Approach for Web Applications with Workload Pattern Recognition |
|
WANG Tao1,2,3, WEI Jun1,2, ZHANG Wen-Bo1, ZHONG Hua1
|
|
1.Technology Center of Software Engineering, Institute of Software, The Chinese Academy of Sciences, Beijing 100190, China;2.State Key Laboratory of Computer Science (Institute of Software, The Chinese Academy of Sciences), Beijing 100190, China;3.Graduate University, The Chinese Academy of Sciences, Beijing 100049, China
|
| Abstract: |
| The dynamic fluctuation of workload influences system metrics, affects the precision of anomalydetection. This paper proposes an online anomaly detection approach for Web applications, which handles workloadfluctuation in both request pattern and volume. The study proposes an incremental clustering algorithm to recognizeonline workload patterns automatically. For a specific workload pattern, the study adopts local outlier factor todetect anomaly and qualify the anomaly degree, and then locate the abnormal metrics with a student’s t-test method.The experimental results show that the clustering algorithm can accurately capture workload fluctuations in atypical Web application, and demonstrate that the approach is capable of not only detecting the typical faults in Webapplications, but also locating the abnormal metrics. |
| Key words: Web application anomaly detection dynamic workload incremental clustering LOF (local outlierfactor) |
