Detecting black-hole attack in mobile ad hoc networks

The paper presents a general approach for detecting black-hole attacks in mobile ad hoc networks, which, due to their mobility and being broadcast in nature, are particularly vulnerable to attacks compared to traditional wired networks. In particular, black-hole attacks can be easily deployed by an adversary. To defend against this attack, we devise a neighborhood-based method to detect whether a black hole attack exists and a routing recovery protocol to set up a correct path to the true destination. Our methods have the remarkable advantage that the number of encryption/decryption operations for authentication is much reduced compared to those methods completely relying on cryptography-based authentication, thus saving many system resources (e.g. reducing energy consumption). Through simulation, we evaluate these methods in terms of packet throughput, routing control overhead, detection probability, false positive probability and false negative probability. Simulation results show that our methods can effectively detect black hole attacks in the sense that detection probability (the probability that one attacker might be detected) in most cases is above 93%. At the same time, it does not introduce much routing control overhead. With the routing recovery mechanism, the packet throughput can be improved by at least 15% and the false positive probability of the detection approach is very low (less than 1.7%).