Design And Implementation Of 5G Core Network Resource Optimization Algorithm Based On SDN And NFV Fusion

With the diversification and individuation of network requirements,the unified standard network is no longer suitable for today’s network environment.To meet the dynamic network requirements,network providers must provide users with excellent performance and customized network services to ensure the service quality of network services.NFV and SDN lead to the network technology update revolution.Network function virtualization technology removes the coupling between network hardware and software devices and realizes the deployment of network functions on virtual network servers.Software defined network technology uses SDN controller as agent to dynamically adjust and distribute network traffic.This paper mainly studies the migration algorithm of service function chain and virtual network routing algorithm.This paper mainly does the following two aspects of work:(1)For NFV networks,in order to achieve load balancing of SFC migration node resources and elastic deployment of service functions,a bandwidth resource-aware migration algorithm based on the coordination of predictive computation and real-time computation is proposed.Migration algorithm With the help of resource awareness algorithm RAIL senses the resource occupation of nodes and selects the destination node for migration according to the perception.Since there are many kinds of resource constraints triggering VNF migration,TOPSIS is introduced into the migration algorithm.In order to solve the problem of real-time computation caused by multi-objective decision algorithm,a predictive algorithm is introduced into the migration algorithm,which reduces the cost of real-time computation.Simulation results show that the algorithm can effectively solve the problem of network load imbalance under the premise of low migration cost.(2)Aiming at the problem of intelligent route planning in SDN environment,the routing selection model of traffic load balancing is designed.The multi-arm slot machine learning problem was applied to route planning and selection,and a route decision MDP based on multi-arm slot machine was designed.Then the MAB-Routing algorithm based on multi-arm slot machine is proposed.In the simulation experiment,the performance of MAB-Routing algorithm,Q-learning algorithm and Dijkstra algorithm was compared.The simulation results show that MAB-routing algorithm has certain advantages in improving network link utilization rate,balancing link bandwidth,delay and data packet loss rate.