Research On Unloading Strategy Of Edge Computing Tasks Based On Container Technology

The demand of massive Internet of Things device connection has given rise to the application and development of edge computing models.Since most terminal devices have limited computing and storage capacity,they are bound to rely more on edge server nodes in exchange for longer delay and lower terminal energy consumption.Edge computing task unloading has become an effective way to efficiently execute terminal computing tasks and improve the resource utilization of edge nodes,among which the edge computing task unloading strategy plays a crucial role.Due to the growing maturity of container technology,it has been widely used and deployed on the edge of the Internet of Things.Therefore,this thesis studies the task unloading optimization strategy based on container technology in edge computing environment,and further gives the task unloading strategy based on container load when considering container load.Specifically,the following work has been accomplished:1.A large number of Io T devices deployed on the edge have great differences due to different business scenarios,and the tasks carried on them have different demands on resources.In order to reduce the energy consumption of Io T devices,a container technology-based computing task unloading strategy is proposed.In this strategy,the time delay energy consumption models of terminal and edge containers are constructed respectively.Furthermore,the cost-benefit functions of the matching degree between the container and the computing task type and the distance between the container and the terminal location are established,and the target container is selected according to the costbenefit function to unload the task.Compared with random strategy and greedy strategy,the experiment results show that the proposed task offloading strategy can greatly reduce the time delay and energy consumption.2.In view of the problem that the strategy proposed in Job 1 has high load on some containers in the concurrent scenario,further consider to allocate and unload computing tasks according to the container load,and put forward the optimization strategy of unload computing tasks based on the container load.This strategy considers the matching degree between the container and the computing task type,the location distance between the container and the terminal,the current load of the container and so on to establish a multiobjective function.In the unloadable container set,the target container is selected according to the evaluation result of the container load model.Compared with the strategy proposed in Task 1,the experimental results show that,compared with the strategy proposed in Task1,the optimized task unloading strategy can ensure the execution delay requirements,and make the load of the whole container cluster more balanced,and the computing resources of the container cluster are more fully utilized.The research results show that the proposed unload strategy can reduce the delay and energy consumption of computing tasks,and can be unloaded flexibly according to the load of the edge container.It not only ensures the efficient execution of tasks,but also makes full use of the resources of edge computing nodes,so it has good practicability.