On Secure Wireless Communications for IoT Under Eavesdropper Collusion

Wireless communication is one of the key technologies that actualize the Internet of Things (IoT) concept into the real world. Understanding the security performance of wireless communications lays the foundation for the security management of IoT. Eavesdropper collusion represents a significant threat to wireless communication security, while physical-layer security serves as a promising approach to providing a strong form of security guarantee. This paper studies the important secrecy outage performance of wireless communications under eavesdropper collusion, where the physical layer security is adopted to counteract such attack. Based on the classical Probability Theory, we first conduct analysis on the secrecy outage of the simple noncolluding case in which eavesdroppers do not collude and operate independently. For the secrecy outage analysis of the more hazardous M-colluding scenario, where any eavesdroppers can combine their observations to decode the message, the techniques of Laplace transform, keyhole contour integral, and Cauchy Integral Theorem are jointly adopted to work around the highly cumbersome multifold convolution problem involved in such analysis, such that the related signal-to-interference ratio modeling for all colluding eavesdroppers can be conducted and thus the corresponding secrecy outage probability can be analytically determined. Finally, simulation and numerical results are provided to illustrate our theoretical achievements. An interesting observation suggests that the SOP increases first superlinearly and then sublinearly with M. Note to Practitioners-This paper was motivated by the problem of securing the important data collection process for the Internet of Things (IoT) applications based on the wireless communication technologies (e.g., wireless sensor networks). The traditional cryptography security method might not be well suited for the highly distributed data collection process in IoT, where edge points (e.g., sensing nodes) are extremely constrained in terms of computing capabilities. This paper explored the application of the promising physical layer (PHY) security approach in the data collection process of IoT, which simply exploits the inherent randomness of wireless channels and noise to provide a strong form of security guarantee. To model the security performance of the data collection process, we mathematically characterized the outage probability that the data is successfully intercepted by the malicious adversaries (i.e., eavesdroppers). Such outage metric serves as the guideline for system designers to initiate the tradeoff between data collection efficiency and data security guarantee, and thus to identify the suitable PHY security schemes for different IoT application scenarios. Simulation and numerical results were provided in this study, while the experiments based on the test bed or real-world implementations have not been performed, which remain as our future research work.