A Two Dimensional Approach for Detecting Input Validation Attacks Based on HMM

摘要：

Web applications are routinely used in securitycritical environments, like financial and medical systems. A wide range of attacks exploit vulnerabilities, typically derived from input validation flaws. This paper proposes a two dimensional approach for detecting input validation attacks in web applications by a two-phase training. During the first phase, normal values of the targeted web application attributes are learned by our proposed IDS, while in the second phase of training, the sequence of normal probabilities of these attributes are modeled as the second dimension. Hidden Markov Models are used both in the training and detection phase. The experimental results demonstrate that the proposed two-dimensional approach results in at least 2％ increase in the detection rate and around 0.3％ decrease in the false alarm rate in lieu of a reasonable delay in the detection process compare to the same system without the second dimension.

关键词： Anomaly Intrusion Detection Two-dimensional detection Hidden Markov Models

作者: Mona Hosseinkhani Ebrahim Tarameshloo Babak Sadeghiyan

作者单位: Dept. of CE and IT Amirkabir University of Tech, Tehran, Iran Dept. of CEand IT Amirkabir University of Tech, Tehran, Iran

会议类型: 国际会议

会议名称: 2011 International Conference on Database and Data Mining(ICDDM 2011)(2011年数据库和数据挖掘国际会议)

会议地点: 三亚

会议语种:英文

页码: 195-199

在线出版日期: 2011-03-25（万方平台首次上网日期，不代表论文的发表时间）