Abstract: In this work, a Multiagent-based architecture for Intrusion Detection System (MIDS) is proposed to overcome the shortcoming of current Mobile Agent-based Intrusion Detection System. MIDS is divided into three major phases namely: Data gathering, Detection and the Response phases. The data gathering stage involves data collection based on the features in the distributed system and profiling. The data collection components are distributed on both host and network. Closed Pattern Mining (CPM) algorithm is introduced for profiling users’ activities in network database. The CPM algorithm is built on the concept of Frequent Pattern-growth algorithm by mining a prefix-tree called CPM-tree, which contains only the closed itemsets and its associated support count. According to the administrator’s specified thresholds, CPM-tree maintains only closed patterns online and incrementally outputs the current closed frequent pattern of users’ activities in real time. MIDS makes use of mobile and static agents to carry out the functions of intrusion detection. Each of these agents is built with rule-based reasoning to autonomously detect intrusions. Java 1.1.8 is chosen as the implementation language and IBM’s Java based mobile agent framework, Aglet 1.0.3 as the platform for running the mobile and static agents. In order to test the robustness of the system, a real-time simulation is carried out on University of Agriculture, Abeokuta (UNAAB) network dataset and the results showed an accuracy of 99.94%, False Positive Rate (FPR) of 0.13% and False Negative Rate (FNR) of 0.04%. This shows an improved performance of MIDS when compared with other known MA-IDSs.

Onashoga S. Adebukola, Ajayi O. Bamidele and Akinwale A. Taofik, “A Simulated Multiagent-Based Architecture for Intrusion Detection System” International Journal of Advanced Research in Artificial Intelligence(IJARAI), 2(4), 2013. http://dx.doi.org/10.14569/IJARAI.2013.020406

@article{Adebukola2013,
title = {A Simulated Multiagent-Based Architecture for Intrusion Detection System},
journal = {International Journal of Advanced Research in Artificial Intelligence},
doi = {10.14569/IJARAI.2013.020406},
url = {http://dx.doi.org/10.14569/IJARAI.2013.020406},
year = {2013},
publisher = {The Science and Information Organization},
volume = {2},
number = {4},
author = {Onashoga S. Adebukola and Ajayi O. Bamidele and Akinwale A. Taofik}
}