Monitor Node Selection Algorithm Based on Mutual Information in Wireless Mesh Networks

Abstract

Intrusion Detection System (IDS) is considered as one of the most effective security mechanisms in wireless networks. However, nodes which are used to monitor the network abnormal behaviors will carry out amount of overhead and degrade network performance. In this paper, we propose a monitor node selection algorithm based on mutual information to solve the above issue. In the algorithm, we consider the importance of nodes and the influence between nodes. Moreover, we represent the importance of nodes' topological position with bridge connection coefficient, and takes degree of interaction between nodes into account. The mutual information value of node is determined and regarded as selection condition of monitor node. In order to allocate the range of monitor node efficiently, we introduce linear threshold model in the social network and consider similarity of node is as influence edge weight. Simulation results show that the proposed algorithm can effectively reduce system cost, ensure the accuracy of intrusion detection system and improve the network performance compared with the traditional node section schemes.