Efficient SFC Orchestration Method Based On Intelligent Algorithm

Network Function Virtualization(NFV)technology implements the decoupling of network functions from proprietary hardware by deploying hardware-based network functions in traditional networks on general-purpose servers in software,improving the flexibility,elasticity and efficiency of network service.Through NFV technology,operators can efficiently and flexibly meet the service needs of users.In the network of NFV technology,a series of virtual network functions(VNF)are connected to form a service function chain(SFC)in a certain order.An SFC needs to be orchestrated and deployed in a specific physical network to provide stable end-to-end services for users.Different arrangements of SFC will have a great impact on the service quality and energy consumption of SFC,so how to ensure the efficient and reasonable arrangement of SFCs is the key issue in the current NFV researches.SFC orchestration refers to selecting appropriate physical nodes to instantiate VNF s in SFC according to the necessary constraints and specific business requirements,and connecting them according to the established order in SFC,so that the traffic can pass through each VNF instance to provide end-to-end services for users.To ensure the stability of services provided by SFC,the service quality of SFC needs to be guaranteed.Different SFC arrangement schemes will have a great impact on the service quality of SFC,so we need to take a reasonable arrangement to ensure the service quality of SFC.At the same time,since SFC deployment to the physical network will increase the energy consumption of the server,we need to take a reasonable way to reduce the energy consumption of the server brought by SFC deployment.The orchestration problem of SFC is an NP-hard problem.To achieve an efficient solution to the SFC orchestration problem,this paper studies the ant colony algorithm and reinforcement learning algorithm in the intelligent algorithm,and designs efficient SFC orchestration methods based on different optimization goals.The research contents of this article are as follows:1)An SFC orchestration method based on packet loss optimization is proposed.This topic uses the ant colony algorithm to minimize the end-to-end packet loss rate of the SFC under the premise of ensuring the experimental tolerance and bandwidth requirements of the SFC,to ensure the quality of service of the SFC.This method aims at minimizing the end-to-end packet loss rate of the SFC,transforms the SFC orchestration problem into the problem of finding the optimal solution in the ant colony algorithm,and uses the shortest path algorithm to orderly link each VNF node to achieve resource allocation and traffic steering.2)An SFC arrangement method based on energy consumption perception is proposed.In this paper,aiming at the scene of SFC dynamic arrival,the problem model in(1)is improved.On the basis of ensuring the service quality of SFC,the reason of increasing server energy consumption is analyzed,and the integer programming model of SFC orchestration problem with energy consumption awareness is established.In order to solve the problem efficiently,this paper proposes an SFC orchestration method based on Q-learning algorithm according to the specific problem scenario.Aiming at the problem caused by the excessive expansion of Q table in Q-learning algorithm,this paper also proposes the SFC orchestration method based on DQN,which realizes the efficient solution of SFC orchestration problem.