Detection of Network Congestion and Denial of Service (DoS) Attacks in Cognitive Radio Networks

Abstract

A Cognitive Radio Network (CRN) is based on a technology that enables secondary users (SUs) to access available licensed spectrum not occupied by primary users (PUs). However, due to its open and wireless nature, a CRN is vulnerable to fraudulent attacks, which might attempt to eavesdrop or modify the contents of packets being transmitted. Moreover, denying the opportunity to SUs to use a free band leads to underutilization of the spectrum space. In this context, it is most important as well as challenging to differentiate between networks under denial of service (DoS) attack from the ones experiencing congestion. This paper adopts the SUs performance measures of packet loss probability, mean queue length and mean throughput of the transmission node in order to devise a packet delivery ratio (PDR) for SUs aiming to determine whether or not the network is experiencing a DoS attack. PDR in this case is the ratio of the number of packets successfully forwarded from the encryption node to the SU transmitter. To this end, a generalized stochastic Petri net (GSPN) is proposed in order to investigate if the network is under a DoS attack and suggest a preventive strategy for an efficient network protection. Based on the application of the Mobius Petri Net Package, typical numerical simulation experiments are carried out and related operational interpretations are made.