| Mobile edge computing technology sinks the processing power of the data center to the edge of the access network which is closer to the user,reducing service transmission delays and improving the user experience.As one of the important business types in the era of big data,streaming computing services still have many problems to be solved in the deployment of such services in the mobile edge computing environment.Based on the open source cloud platform Kubernetes,this thesis design and develop a MEC platform for streaming computing services.The main work contents and innovations are as follows.First,the basic technology of the MEC platform are inroduced,including the technology of cloud computing and the business characteristics of streaming computing services,and common task scheduling algorithms.Kubernetes is used as a basic platform to design and develop a MEC system for streaming computing services.After testing,it proves that the prototype system designed in this thesis can effectively manage the streaming computing service in the MEC scenario.Secondly,based on the characteristics of the streaming computing service,a task deployment model is proposed,and a scheduling algorithm based on genetic algorithm is used to solve the model.The result is compared with the professional optimization software CPLEX,which shows that in small-scale scenarios,the algorithm in this thesis can complete the scheduling calculation faster at the expense of less than 20% performance,which is suitable for the task deployment of streaming computing services in the MEC scenario.Finally,in response to the data backlog caused by the time-varying and sudden changes of streaming data,a traffic management scheme has been added,including two modules of real-time monitoring and traffic prediction.Through comparative experiments,it is found that the traffic management function can effectively reduce the data backlog at the platform entrance due to the characteristics of streaming data. |
PDF Full Text Request