Research On Task Scheduling Mechanism Based On Edge Computing

With the continuous development of the Internet of Things,mobile Internet,5G communication,and big data,the amount of data and calculations of various mobile applications has also increased rapidly in recent years.In order to cope with the growing demand for task processing,as an innovative computing model,edge computing has gradually attracted people's attention.By expanding the computing capabilities previously deployed in the cloud data center to the edge of the network,various tasks can be processed at the nearby edge servers of users.Compared with cloud computing,edge computing can reduce the delay of task transmission and improve service quality.However,due to the limited resources of an edge server,collaboration with other edge servers or cloud servers is required to improve the overall network performance.This paper mainly studies the problem of task offloading in edge computing from two aspects.First,this article studies the task offloading decision problem in the device-to-edge collaboration mode.By designing the cost type in the network,this problem is transformed into an NP-hard cost minimization problem.The paper designs a task offloading algorithm based on collaboration of edge servers,and compares the algorithm with a typical task scheduling algorithm.It is found that the algorithm proposed in this paper has better performance under different task densities,and has a better system performance after task scheduling.Second,this article investigates the problem of offloading decisions with task deadlines.Based on the collaboration mechanism of the deviceedge-cloud,this paper introduces three kinds of costs,which converts the task offloading decision problem into a cost minimization problem with a task completion time limit.This paper designs a task offloading algorithm based on task deadlines.By introducing the concept of service level,tasks with different deadlines are optimized in order of different service levels.Compared with the strategy of local processing and complete offloading,this algorithm has can better control the total system cost while meeting the deadline constraint of the tasks,and fully optimize the overall network performance.By studying the task scheduling and offloading mechanism and related algorithms in the edge network,we can effectively improve the efficiency of the edge cloud and improve the quality of service for users.