Microservice-based Management And Orchestration For 5G Core Network

5G empowers thousands of industries,becoming the new engine of the digital economy.Unlike traditional mobile communication systems,5G will face a wide variety of application scenarios,which will need to meet the very different performance indicators of various vertical industries.This poses a challenge to the management and orchestration(MANO)of the 5G core network.This thesis aims to design and implement microservice-based 5G core network management and orchestration architecture,and to implement auto scaling of 5G core network on the basis of it,so as to achieve on-demand deployment of 5G core network instances and improve resource utilization.The main research of this thesis are as follows:(1)This thesis introduces microservice-based management and orchestration plane into the5 G core network architecture,and realizes the on-demand deployment of 5G core network in multiple application scenarios.This thesis first designs a 5G core network system architecture that includes application plane and management and orchestration plane with reference to the Network Function Virtualization(NFV)architecture,and proposes to refactor management and orchestration plane based on the microservice architecture.Secondly,this thesis designs and implements each network function in the management and orchestration plane,and designs the calling process between each network function for multiple stages of the 5G core network life cycle,so as to realize the creation,instantiation,expansion and contraction,termination and deletion of 5G core network instance.Finally,an experimental platform was built,and the microservice-based 5G core network management and orchestration was functionally verified and tested.The experimental results show that the microservice-based management and orchestration function has better resource utilization and disaster tolerance.(2)This thesis studies the expansion and contraction problems during the 5G core network management and orchestration process.Aiming at the problem of how mobile operators dynamically adjust resource allocation when the 5G core network faces fluctuating loads,this thesis designs an 5G core network auto scaler that combines horizontal scaling and vertical scaling.The design of the auto scaler is based on automatic control theory,supports the simultaneous operation of the horizontal scaler and the vertical scaler,combines the advantages of a wide horizontal scaling range and high vertical scaling accuracy,and adds a load prediction mechanism to ensure the timeliness of expansion and contraction.This thesis firstly designs each module of the auto scaler in detail,the reasonable range of the auto scaler parameter setting is deduced,and the more reasonable parameters setting is given through simulation.Finally,the performance of the designed auto scaler is verified through comparative experiments.The simulation results show that the resource utilization of the auto scaler designed in this thesis is better than that of Kubernetes Horizontal Pod Autoscaler.