A robust Kolmogorov-Smirnov detector for misbehavior in IEEE 802.11 DCF

The CSMA/CA protocols are designed under the assumption that all participant nodes would abide to the protocol rules. This is of particular importance in distributed protocols such as the IEEE 802.11 distributed coordinating function (DCF), in which nodes control their own backoff parameters. A selfish node may deliberately modify its random assignment and gain unfair access to the network resources. This would result in an increased observed collision probability for the rest of the nodes, that would increase their backoff windows as a result, further increasing the benefit of the selfish nodes. In this work, we develop of a robust non parametric batch detector based on the Kolmogorov-Smirnov (K-S) statistics that does not require any modification on the existing CSMA/CA protocols, and we apply it to detect misbehaviors in an IEEE 802.11 DCF network using the ns-2 simulator. We show that our method has a performance comparable to the optimum detectors with perfect information for the majority of misbehaviors, and it is able to detect any deviation from the protocol after just a few transmissions from the offending terminal.