Investigating Cellular Automata Based Network Intrusion Detection System For Fixed Networks (NIDWCA)

摘要

Network Intrusion Detection Systems (NIDS) are computer systems which monitora network with the aim of discerning malicious from benign activity on thatnetwork. With the recent growth of the Internet such security limitations arebecoming more and more pressing. Most of the current network intrusiondetection systems relay on labeled training data. An Unsupervised CA basedanomaly detection technique that was trained with unlabelled data is capable ofdetecting previously unseen attacks. This new approach, based on the CellularAutomata classifier (CAC) with Genetic Algorithms (GA), is used to classifyprogram behavior as normal or intrusive. Parameters and evolution process forCAC with GA are discussed in detail. This implementation considers bothtemporal and spatial information of network connections in encoding the networkconnection information into rules in NIDS. Preliminary experiments with KDD Cupdata set show that the CAC classifier with Genetic Algorithms can effectivelydetect intrusive attacks and achieve a low false positive rate. Training aNIDWCA (Network Intrusion Detection with Cellular Automata) classifier takessignificantly shorter time than any other conventional techniques.