Primary user emulation attack in dynamic spectrum access networks: Threats, mitigation and impact.

摘要

Cognitive radio enabled dynamic spectrum access (DSA) networks allow unlicensed "secondary users" to use the spectrum bands left vacant by licensed "primary users" provided that primary user communications are not disrupted by secondary users. While this innovative communication paradigm brings promise in improving spectrum utilization, it also introduces entirely new classes of security threats that must be taken care of to ensure successful implementation of DSA networks. This dissertation studies primary user emulation attack (PUEA), a denial-of-service (DoS) attack that is unique to DSA networks.;In the first part of this dissertation, we present an analytical model to study the feasibility of PUEA. A lower bound on the probability of successful PUEA on a secondary user is obtained. We then present a Neyman-Pearson composite hypothesis test (NPCHT) and a Wald's sequential probability ratio test (WSPRT) for each secondary user to mitigate PUEA.;We then study the effect of cooperation between secondary users on thwarting PUEA. An individual detection mechanism for each secondary user is proposed to achieve preliminary spectrum decisions in the presence of PUEA. Based on individual decisions, we develop a centralized spectrum decision protocol against PUEA with the help of a centralized controller. We also develop a distributed protocol, called NEAT: NEighbor AssisTed spectrum decision protocol, for each user to mitigate PUEA with the aid of neighboring nodes. Byzantine failure problem in the context of data fusion is also discussed and taken into account in the protocols.;In the last part of this dissertation, we analyze the impact of PUEA on the network performance for DSA networks carrying real-time and non-real-time traffic, respectively. We propose a three dimensional continuous time Markov chain (3D-CTMC) model to determine the call blocking and dropping probabilities as well as the mean delay for secondary users. We also evaluate the performance of DSA networks that deploy different PUEA-mitigation mechanisms, including the centralized and distributed protocols we proposed.