Game Theory and Reinforcement Learning Based Secure Edge Caching in Mobile Social Networks
Qichao Xu,Zhou Su,Rongxing Lu
DOI: https://doi.org/10.1109/tifs.2020.2980823
IF: 7.231
2020-01-01
IEEE Transactions on Information Forensics and Security
Abstract:Edge caching has become one of promising technologies in mobile social networks (MSNs) to proximally provide popular contents for mobile users. However, since caching contents inevitably consume resources (e.g., power, bandwidth, storage, etc.), edge caching devices maybe selfish to cheat the content provider for earning service fees. In addition, due to the open access of edge caching devices, the edge caching service is vulnerable to various attacks, such as man-in-the-middle attack and content tamper attack, etc., resulting in the degradation of content delivery performance. To efficiently tackle the above problems, in this paper, we propose a secure edge caching scheme for the content provider and mobile users in MSNs. Specifically, we first develop a secure edge caching framework consisting of the content provider, multiple edge caching devices, and some mobile users. To motivate the participation of edge caching devices, Stackelberg game is exploited to model the interactions between the content provider and edge caching devices. The content provider serves as the game-leader to determine the payment strategy of secure caching service and each edge caching device is the game-follower to make the strategy on the quality of secure caching service. Especially, the zero payment mechanism is adopted to suppress the selfish behaviors of edge caching devices. Apart from this, for lack of the knowledge on interactions between the content provider and edge caching devices in dynamic network scenarios, we also employ the Q-leaning to derive the optimal payment strategy of the content provider and the security strategy of edge caching device. Extensive simulations are conducted, and results demonstrate that the proposed scheme can efficiently motivate edge caching devices to provide the content provider and mobile users with high-quality secure caching services.