Utilizing the Uncertainty of Intrusion Detection to Strengthen Security for Ad Hoc Networks

Dennis Dreef, Sanaz Ahari, Kui Wu, Valerie King

Abstract: With Mobile ad hoc networks rapidly approaching practical use by the masses, security has become a main concern. Intrusion detection, as the second line of defense, is an indispensable tool for highly survivable networks. Nevertheless, intrusion detection systems suffer from false alarms, which are extremely hard to control in mobile ad hoc networks. As a result, local intrusion detection engines in mobile ad hoc networks may only be able to claim anomaly with low confidence. Such uncertain knowledge is not very helpful in assisting end users with the final decision for intrusion response. The uncertainty of intrusion detection results, however, can provide a rough guideline on potential hazards and can be utilized to enhance security. In this paper, we present a randomized algorithm to utilize such uncertain knowledge to help routing protocols defend against the rushing attack. By eliminating excess bandwidth usage and adopting an adaptive randomized forwarding mechanism, the presented method is more efficient than previously existing solutions.