Trusted Virtual Network Embedding in Blockchain-Based Smart Cyber-Physical Systems

Abstract

With the rapid development of cyber-physical systems (CPS), the rising variety of services in different application scenarios has posed a serious challenge to the traditional network resource allocation problem. In addition, the current CPS is facing many security challenges and threats. To this end, this paper proposes a trusted virtual network embedding algorithm in CPS based on blockchain combined with software defined network (SDN) technology. The algorithm is optimized by deep reinforcement learning (DRL), and the important attributes of the network are extracted to simulate the real network environment. DRL agent obtains the optimal virtual node embedding scheme by training, and then completes the virtual link embedding by breadth-first search algorithm. Finally, we design a simulation experiment for the three indicators of VNE: long-term average revenue, long-term revenue-cost ratio and VNR acceptance rate. The experimental results show that our algorithm has excellent performance while ensuring security.