作者: Douglas S. Reeves , Xiaoyong Wu , Vinay A. Mahadik
DOI:
关键词:
摘要: In this paper, we describe a method of detecting denial Quality Service attacks on DiffServ networks. Our approach focusses real time and quick detection, scalability to large networks, negligible false alarm generation rate. Sensors sample QoS parameters like bit rate, packet dropping jitter specific Virtual Leased Line (VLL) flows at predefined strategic points in their paths. We detect anomalies sampled network flow statistics using the EWMA Control Chart test for highly stationary measures rest adapt SRI’s χ statistic based NIDES approach. implementation shows that has 100% detection rate above its threshold level those produce statistically significant degradation. The is low less than about 15 minutes. maximum inherent both tests any monitored combined order 1 1000 valid status alerts either normal or under attack. believe given results our system, strong candidate intrusion low-cost commercial deployment. ∗Vinay A. Mahadik pursuing Master Science Computer Networking NC State University, Raleigh. Email : vamahadi@unity.ncsu.edu †Xiaoyong Wu with Advanced Research Group, MCNC, Triangle Park. xwu@anr.mcnc.org ‡Douglas S. Reeves Department Science, reeves@unity.ncsu.edu