【摘要】 基于校园网实际域名系统(domain name system, DNS)服务日志,研究DNS异常行为的检测方法,提出针对DNS源IP异常检测的基于多维时序特征的局部异常因子检测(local outlier factor, LOF)算法,并在此基础上,提出基于多特征的域名异常分析方法,以实现更为精准的DNS异常识别,保障校园网的稳定和安全.更多还原

【Abstract】 In this paper, we propose a local outlier factor(LOF) algorithm based on multi-dimensional timing characteristics for detecting abnormal source IPs of DNS. The algorithm is used to identify abnormal source IPs of the DNS traffic of a campus network. Based on the algorithm, we further introduce a multi-feature-based abnormal domain name detection method and efficiently improve the detection of DNS anomalies of the campus network.更多还原