Design And Traffic Scheduling Research Of 5G Network Slicing Simulation Platform Based On SDN And NFV

With the fast development of communication technology,the demands of users and enterprises for communication networks have grown exponentially.To this end,5G network slicing technology has emerged as the times require.By integrating software-defined network technology and network function virtualization technology,network slicing can divide and reorganize limited network resources,and centrally manage and control traffic,so as to achieve service independence and isolation,which providing users with better quality of service.However,due to the limitation of physical resources,in the face of large-scale slice subscription,it is hard to completely isolate resources for all slices.Parts of slices need to share resources.Maximizing the resource sharing situation,improving the resource utilization of network slicing and ensuring the quality of service delivery have become the key issues for operators.This thesis focuses on the design and implementation of the simulation platform for5 G network slicing,and establishes an experimental environment to verify the proposed scheme and algorithm.The main work contents are as follows:(1)This thesis integrates SDN and NFV technologies,and designs a 5G end-to-end network slicing simulation platform from the wireless access network to the transmission bearer network to the cloud core network.It realizes the mapping and isolation of network resources.On this basis,a slicing network resource simulation management platform is developed,which realizes the self-service slicing function subscription and automatic deployment capabilities.This thesis also conducts the deployment of actual slicing services based on container technology,and ensures the service quality of network slicing by allocating customized slicing resources for video services played by the end users,and then verifies the effectiveness of the slicing simulation platform.(2)Considering the problem that local link congestion tends heavily influence the quality of network delivery,a traffic scheduling management scheme for bearer network slices based on SDN software-defined network and deep reinforcement learning is proposed.This scheme takes advantage of the separation of forwarding and control of SDN and makes decisions with the support of deep reinforcement learning algorithm,so as to minimize the network load and improve the resource utilization.The proposed traffic scheduling scheme is shown superior to the traditional network topology management approaches through experimental comparison.