SFC Orchestration and SDN-VNF Trusted Deployment Scheme Based on Deep Reinforcement Learning Algorithm

This paper introduces the background knowledge of network function virtualization, SDN, NFV, service function chain and related resource allocation problems in VNF, and explains the principles of reinforcement learning algorithms including Q-learning, A3C algorithm and DDQN algorithm in detail. Software Defined Network (SDN) separates the network data forwarding function, and the control plane is responsible for formulating and issuing the structured data packet forwarding rules according to the needs of network users. Compared with other algorithms, A3C algorithm can significantly reduce the requirement of hardware resources. This paper presents SFC orchestration and SDN-VNF trusted deployment scheme based on deep reinforcement learning algorithm. Firstly, the layout scheme of SDN-VNF based on A2C and DQN algorithm is compared. Finally, the efficiency of A3C and DDQN algorithms in SFC orchestration scheme is compared. Experimental results show that the proposed deep reinforcement learning algorithm is effective in SFC Orchestration and SDN-VNF.