Blockchain and digital twin empowered edge caching for D2D wireless networks

Abstract

Edge caching is considered a promising technology to fulfill user equipment (UE) requirements for content services. In this paper, we explore the use of blockchain and digital twin technologies to support edge caching in a Device-to-Device (D2D) wireless network, where each UE may fetch content from its own caching buffer, from other UEs through D2D links, or from a content server. A digital twin monitors and predicts the operating status of UE by storing crucial data such as the location, estimated processing capability, and remaining energy of each UE. To enable secure and credible trading between UEs, the blockchain technology is used to supervise transactions and constantly update UEs’ reputation values. We formulate an optimization problem to maximize an objective function that considers the content fetching performance, network lifetime and UE’s handover costs by optimizing the content placement and fetching strategies, subject to constraints on the UE’s storage capacity, the upper limit of serving other UEs, and latency requirements. To solve this complicated problem for a dynamic network environment, we propose a proximal policy optimization-based deep reinforcement learning framework. Simulation results demonstrate that our proposed algorithm converges rapidly and can efficiently maximize the rewards, network lifetime and content fetching gain while minimizing handover costs.