Moving Target Defense for Internet of Things Based on the Zero-Determinant Theory

Abstract

At present, the proliferation of the online connected devices conceives the Internet of Things (IoT), in which many wireless sensors, smart devices are implemented. However, the nature of openness rooted in IoT makes itself vulnerable to be attacked. One of the pioneer countermeasures is the moving target defense (MTD), which encourages an active and dynamic defense in IoT. In this article, a macroscopic research in MTD is carried out. The existing macroscopic studies take advantage of a traditional game theory. Consequently, protected IoT devices need extra operations to dominate the game. In this article, we take a dramatically different approach where a player can dominate the game without extra operation. Our approach benefits from the power of the zero-determinant (ZD) strategy, in which the player who adopts ZD can unilaterally set the expected payoff of the adversary or itself. Aware of such a powerful strategy, both players may want to employ it for dominating the confrontation. In this case, two fundamental questions need to be answered: who should take the ZD strategy? and to what extent can the ZD player dominate the game? To solve these problems, we model the interactions between the IoT devices and the malicious attackers as a Markov game. Besides, we obtain the conditions to adopt ZD, based on which we deduce the effectiveness of the ZD player. To the best of our knowledge, we are the first to employ the ZD strategy theory to enhance a better counterattack performance in IoT.