Research On Traffic-awareness Based Intelligent Orchestration Algorithm Of Virtual Network Function

In response to the rapidly growing demand for data traffic and diversification of business types,introducing the network slicing technology into next-generation wireless network architecture is an inevitable trend in the future development of 5G networks.As a key technology for network slicing,network function virtualization is based on general purpose server hardware and provides various virtual network functions in the form of software.For different service scenarios,dynamic deployment of virtual network functions and on-demand resources allocation can reduce network operation costs while meeting different service QoS requirements.However,considering the limited nature of network resources and the dynamic nature of service requests,how to effectively and rationally orchestrate virtual network functions is critical to the implementation of network slicing.This thesis focuses on the research of the virtual network function orchestration on the 5G network slicing scene.The main research contents and innovations of this paper are summarized as follows:1.In order to solve the unreasonable virtual resource allocation caused by the dynamic change of service request and delay of information feedback in virtualized network,a traffic-aware algorithm which exploits historical service function chaining(SFC)queue information to predict future load state based on neural network is proposed.With the prediction results,the virtual network function(VNF)deployment and the corresponding computing resource allocation problems are studied,and a VNFs' deployment algorithm based on delay minimization is developed.On the premise of satisfying the minimum resource demand for future queue non-overflow,the on-demand allocation method is used to maximize the computing resource utilization.Simulation results show that the prediction model based on neural network in this thesis obtains good prediction results,realizes online monitoring of the network.The VNF deployment algorithm method effectively reduces the bit loss rate and the average end-to-end delay caused by overall VNFs' scheduling at the same time.2.In order to solve the VNF migration optimization problem caused by the dynamicity of service requests on the 5G network slicing architecture,firstly,a stochastic optimization model based on constrained Markov decision process(CMDP)is established to realize the dynamic deployment of multi-type SFC.This model aims to minimize the average sum operating energy consumption of general servers,and is subject to the average delay constraint for each slicing as well as the average cache,bandwidth resource consumption in the meantime.Secondly,in order to overcome the issue of having difficulties in acquiring the accurate transition probabilities of the system states and the excessive state space in the optimization model,a VNF intelligent migration online learning algorithm based on reinforcement learning framework is proposed.The method approximates the behavior value function by convolutional neural network,so as to formulate a suitable VNF migration strategy and CPU resource allocation scheme for each network slice according to the current system state in each discrete time slot.The simulation results show that the proposed algorithm can effectively meet the QoS requirements of each slice while reducing the average energy consumption of the infrastructure.