Kittikhun Thongkanchorn.. Evaluation of three intrusion detection systems under various attacks. Master's Degree(Computer Science). Mahidol University. Mahidol University Library and Knowledge Center. : Mahidol University, 2012.
Evaluation of three intrusion detection systems under various attacks
Abstract:
Intrusion detection system (IDS) tools for detecting malicious traffic have been widely used in many organizations and they use a variety of technologies. Each IDS tool deploys different approaches and has been developed under different purposes. Intrusion detection systems include a set of IDS rules which can be defined differently. Therefore, choosing an IDS tool to work efficiently and appropriately in a specific environment would not be easy. The goal of this research was to evaluate three popular open-source IDS tools in terms of performance and accuracy. The selected IDS tools were Snort, Bro and Suricata. In addition, their system architecture and the main components were compared and analyzed. The experiments conducted used various attack types including DoS attack, DNS attack, FTP attack, Scan port attack, and SNMP attack. Each experiment was run under different traffic rates and only a specific set of rules was active. Moreover, the performance metrics used to measure was the number of packets lost, the number of alerts, the CPU utilization and the memory usage. The results showed that each attack type had significant effects on the performance of each IDS tool in different ways. Specifically, Bro IDS showed better performance than other IDS tools when evaluated under different attack types and using a specific set of rules. The results also indicated that the accuracy dropped when the three IDS tools activate the full rule sets instead of a specific set of rules.