Task Generation And Scheduling Mechanism For Computing Power Network Applications

With the in-depth development of communication technology and industrial transformation,new computing power network applications represented by intelligent collaborative driving and intelligent security are emerging.These applications require extensive data transmission and a large amount of computing resources,and the functional components that make up these applications are becoming more and more complex,and there are various dependencies among various functional components.Especially in multi-terminal motion scenarios,these applications are concurrent,which increases the challenge of application scheduling.At present,cloud computing and edge computing are separated from the network,and it is difficult to meet the application service quality requirements.The computing power network deeply integrates computing and the network,provides more intelligent services for the efficient collaboration of cloud,edge,and end computing power,promotes computing and network integration to a new stage,and provides a new idea for solving the above-mentioned new computing network application scheduling.Due to the large number of functional components and complex dependencies of new computing power network applications,direct scheduling in the cloud-edge-device environment collaboration environment will incur very high costs.At the same time,distributed deployment also needs to split the application to generate various modules with cohesive functions to form tasks that can be executed independently,and then complete the scheduling in the actual computing power network environment to ensure the effective completion of application tasks.However,the current computing power network application task generation mechanism is not perfect.It only considers the characteristics of the application itself,and cannot match the heterogeneous cloud edge node resources in the computing power network environment with the resource requirements of the computing power network application.As a result,the generated computing power network application tasks cannot meet the requirements of task scheduling.At the same time,for the computing power network environment with strong terminal mobility,the current scheduling algorithm cannot effectively perceive the terminal motion state,resulting in unstable connections between terminals and edge nodes.This makes it impossible to schedule multiple concurrent computing power network application tasks to appropriate cloud edge nodes in a timely and efficient manner.Aiming at the deficiencies in existing research,this paper proposes a task generation and scheduling mechanism for computing power network applications.Aiming at the problem of task generation that does not match between computing power network applications and heterogeneous cloud-edge-end nodes,this paper proposes a new task generation mechanism based on application component aggregation.In addition to considering the factors of each functional component of the application,including resource requirements and dependencies,this mechanism also perceives the resource differences of cloud-edge-end nodes in the actual network environment and describes task generation by establishing an optimization problem.This method takes the expected execution time of the application as the target,realizes the division of the application modules by matching the resource requirements of the functional components of the application with the resource conditions of the cloud-edge-end nodes,and generates suitable tasks that can be directly scheduled.Compared with other methods,the task generation method proposed in this paper reduces the expected execution time of the application by 77.48%on average,the communication volume between expected tasks by 86.11%on average,and the algorithm execution time by 96.66%on average.Aiming at the problem that the current task scheduling mechanism has insufficient awareness of terminal mobility and poor cloud-edge-end collaboration performences in the smooth execution of applications,this paper proposes a task scheduling mechanism for terminal mobility awareness in a computing power network environment.This method takes into account the movement of the terminal,associates the location information of the terminal with the communication connection through the prediction and analysis of the movement trajectory of the terminal,and schedules the task to the appropriate cloud-edge-end nodes according to the resource requirements of the application task.The simulation shows that,for the concurrent application scenario of multiple mobile terminals,even under the condition of fast movement of the terminals,the mechanism can still effectively realize the task scheduling and guarantee the corresponding quality of service.Compared with other task scheduling methods,the task scheduling algorithm proposed in this paper has the following improvements:1)When the number of generated tasks changes,the average execution time of the application is reduced by 6.74%,and the execution energy consumption is reduced by 4.45%;2)When the number of schedulable edge nodes changes,the execution time of the application is reduced by 17.95%on average,and the execution energy consumption is reduced by 8.05%;3)When the maximum speed of the terminal changes,the execution time of the application is reduced by 34.67%on average,and the execution energy consumption is reduced by 21.78%.To sum up,with the development of complex applications of computing power networks and the continuous improvement of the performance of heterogeneous cloud edge computing nodes,this paper discusses the task generation based on the aggregation of computing power network application components and the task scheduling scheme based on terminal mobility perception.It is of great significance to make full use of the heterogeneous cloud edge resources of the computing power network,reduce the total application execution time and total cost,and improve the application service quality.