Game Theory Meets Wireless Sensor Networks Security Requirements and Threats Mitigation: A Survey

Mohamed Abdalzaher, Karim Seddik, Maha Elsabrouty, Osamu Muta, Hiroshi Furukawa, Adel Abdel-Rahman
2016 Sensors  
We present a study of using game theory for protecting wireless sensor networks (WSNs) from selfish behavior or malicious nodes. Due to scalability, low complexity and disseminated nature of WSNs, malicious attacks can be modeled effectively using game theory. In this study, we survey the different game-theoretic defense strategies for WSNs. We present a taxonomy of the game theory approaches based on the nature of the attack, whether it is caused by an external attacker or it is the result of
more » ... n internal node acting selfishly or maliciously. We also present a general trust model using game theory for decision making. We, finally, identify the significant role of evolutionary games for WSNs security against intelligent attacks; then, we list several prospect applications of game theory to enhance the data trustworthiness and node cooperation in different WSNs. In [28] , the authors discuss the promising features of game theory approaches for the wireless networks, while in [29, 30] , different trends of using game theory for WSNs have been reviewed. Recently, with the emergence of infrastructure-less and distributed systems, game theory has found its way in decentralized communication systems [31] . One of the problems in this category is related to security in WSNs. The security situation, which involves an interaction between the defender(s) and attacker(s), can be directly mapped to a game among players in which each player strives to promote its benefit. More particularly, having the action of the attacker(s) or the defender(s) depending on the counter-action of the other party places game theory as a perfect fit for this security model [22, [24] [25] [26] 32] . In this paper, we introduce a brief interpretation of the different game techniques presented in the literature to address WSN security. In addition, an overall view of the desired WSN properties in terms of security fulfillment is presented. We study the game theory based approaches for mitigating different WSN security threats from the state-of-the-art literature on the topic. We classify those approaches into two main categories, namely, cooperative games and non-cooperative games, and each summarizes the involved defense strategies based on game theory. Then, we propose a taxonomy of game theoretic defense strategies taking into consideration the attacked layer, attack features, attack consequences, convenient defense game approach, and game type. Afterward, a general trust model based on the discussed game theory approaches and scenarios is introduced to take into account the variability and features of the attack types. Consequently, we would utilize this model to any network environment (cooperative/non-cooperative game with internal/external attack). Finally, we present some applicable future trends for the interested researchers, showing the capability of facing intelligent attacks [33-36] using the evolutionary game approach [37] [38] [39] . The rest of the paper is structured as follows. Section 2 provides a brief overview about game theory. Section 3 outlines game theory classifications, addresses the different game types that are involved in WSN security, and provides the security properties needed for WSN security. Section 4 illustrates the proposed taxonomy of game theory defense strategies for WSN security showing the types of attacks and the types of suitable games to mitigate these attacks. Section 5 presents the proposed general trust model based on the discussed game theory types and attack types in WSNs. In Section 6, the applicable future trends are listed. Section 7 concludes the paper. Game Theory: A Brief Overview Game theory is an advanced branch of intelligent optimization. The model of game theory represents a game between player groups that choose to behave cooperatively or non-cooperatively and try to promote their benefits (payoffs) through the used strategy(ies) executed through the cumulative players actions. [19, 20] survey the fundamental definitions of game parameters, which can be summarized as follows: Definition 1. A game is a description of the strategic interaction between opposing, or cooperating, interests where the constraints and payoff for actions are taken into consideration. Definition 2. A player is a basic entity in a game, which is involved in the game with a finite set of players denoted by N that is responsible for taking rational actions denoted by A i , for each player i. A player can represent a person, machine, or group of people within a game. Definition 3. The Utility/Payoff is the positive or negative reward to a player for a given action within the game denoted by u i : A → R, which measures the outcome for player i determined by the actions of all players A = × i∈N A i , where the symbol × denotes Cartesian product. Definition 4. A strategy is a plan of action within the game that a given player can adopt during game play denoted by a strategic game N, (A) , (u i ) .
doi:10.3390/s16071003 pmid:27367700 pmcid:PMC4970053 fatcat:3utgdvzzdvhpjnvvsaweywmyvm