A Context Adaptive Intrusion Detection System for MANET

Due to the ad hoc and mobile nature of a MANET, it is much more vulnerable to attacks than a wired network. As a result, there has been a significant research focusing on designing an Intrusion Detection System (IDS) for MANETs to detect anomalous behavior and misuse. However, each mobile node in a MANET typically has limited energy and thus it is not efficient to perform IDS functions within a node to detect every incoming packet. There is a need for an IDS to implement an intelligent control mechanism in order to monitor and recognize security breach attempts efficiently over a period of the expected network lifetime. By leveraging the Network Node Intrusion Detection (NNID) strategy, we developed a context adaptive IDS controller that advises an IDS to carry out intrusion detection while being prepared for a possible "cut through" if it is likely that the residual energy is not sufficient. By being embedded with the context adaptive IDS controller, the proposed Context Adaptive Intrusion Detection System (CAIDS) is able to adapt to the current node context (such as residual energy, security threats and traffic loading) for accommodating and inspecting new arriving packets. The performance is evaluated using a reward function that discovers an effective way to perform intrusion detection and delivers security benefits while meeting the energy budget. The numerical results show that CAIDS offers a good trade-off between lifetime performance and security. This study demonstrates empirically that the CAIDS model intelligently monitors and recognizes security breach attempts while adhering to the resource budget plan over the period of expected network lifetime. (C) 2010 Elsevier B.V. All rights reserved.