| Mobile edge computing technology extrapolates or sinks computing power to the edge of the mobile access network in order to facilitate deep integration of cellular networks and cloud computing services,which can minimize end-to-end delays in delivering services.However,the combination of cloud computing and wireless access network also meets some challenges,especially the resource limitation and user mobility in mobile edge environment.Based on the open source cloud platforms included OpenStack and Kubernetes,this paper designs and implements a Customer-Facing-Service migration system based on the specific mobile edge application scenarios.The main work contents and innovations are as follows:Firstly,this paper analyzes and compares the mainstream open-source cloud computing management platforms,and chooses OpenStack and Kubernetes as the research objects.Then this paper focuses on the system architecture,core components,workflow and some source code of OpenStack and Kubernetes.After this work,the best practices for Customer-Facing-Service migration on both platforms were explored and implemented.Secondly,combining the resource management functions provided by the virtualization platform with the terminal location technology of Cell-ID,this paper designs and implements a Customer-Facing-Service migration system in the mobile edge environment.After the test,the user's cross-region behavior can be monitored in time,and the real-time memory,service image and incremental image will be migrated in the system.Finally,this paper proposes a service scheduling model and a service prediction model for resource allocation in complex multi-user scenarios.These two models are solved by genetic algorithm,regression algorithm and BP(Back Propagation)neural network algorithm.The result is compared with what calculated by GLPK,which shows that in small-scale scenarios,the proposed model solving method can achieve an order of magnitude improvement in time at the expense of less than 30% performance. |
PDF Full Text Request