Research On Modeling And Algorithm Of Distributed Task Scheduling For Edge Computing

With the vigorous development of the Internet of Things technology,the applications on the Internet of Things terminal become more and more abundant,and the amount of calculation required is also increasing.However,the computing power and storage resources of Io T terminal devices are very limited,and the processing of tasks on the terminal devices will cause high delays.Moreover,terminal equipment will also be limited by the power of the power supply.Long-term data processing will cause large energy consumption and insufficient battery life.The applications on the equipment often have relatively high requirements for delay and power consumption.In order to solve these problems,the traditional method is to upload some complex computing tasks to the central server for execution,but the central server is often far away from the terminal device,and data transmission will cause a high delay.It may be considered to upload these complex calculations to an edge server closer to the terminal device to perform,this can reduce the delay of data transmission.However,edge servers are also limited by computing resources and network resources.How to schedule computing tasks to improve the task time and terminal life time is particularly important.Facing the distributed task scheduling problem of edge computing,this paper mainly carried out the following research work:(1)Aiming at the problem that the terminal equipment needs to perform task scheduling due to limited computing power,this paper analyzes the influencing factors of uploading tasks to the central server in detail,and gives the delay and energy consumption of the task to be executed locally or uploaded by the server Consumable components,establishs the problem model of task scheduling,and gives a task scheduling algorithm that optimizes delay and energy consumption.(2)Using physical devices of the Internet of Things build a task scheduling experiment system.The system includes terminal equipment,edge servers,and a central server.The terminal equipment is used for data collection,and after data collection,you can choose to process the calculation task locally or upload it to the server for processing.A preliminary test was conducted on the Raspberry Pi.The data processing efficiency of the system using dynamic scheduling algorithms is higher than that of data processing locally or uploading only to the server,which initially proves the effectiveness of the scheduling algorithm.In order to simulate the verification of the task scheduling algorithm under the above conditions for large-scale terminal devices,edge servers and central servers,we conducting a simulation test.The analysis of the experimental results shows that the algorithm is an approximately optimal algorithm and can be effectively reduce the delay and energy consumption in the system.