Routing attacks detection and reaction scheme for mobile ad hoc networks using statistical methods

Due to the vulnerabilities of nodes and channels, absence of central management, and the resources constraints, mobile ad hoc network (MANET) is relatively vulnerable to malicious attacks. In MANET, routing attacks try to disrupt the routing protocol functions by intentionally or unintentionally dropping packets or by propagating faked routing messages. However, due to their computation requirements and inefficiency against internal attacks, the prevention only mechanisms are not powerful enough to secure MANET routing protocols. In this paper, we present a distributed and cooperative scheme for MANETs routing attacks detection. The scheme collects and analyzes data that are sensitive to routing attacks so as to detect anomalous events. In this scheme, both direct observations and indirect observations are used so as to characterize the behaviors of both neighboring and remote nodes. The scheme utilizes the Grubb´s test to develop new methods for detecting routing attacks and evaluating the exchanged information.