Research On Collaborative Scheduling Strategy For Tasks In Edge Computing

With the rapid development of the Internet of Things and 5G,the era of the Internet of Everything has arrived,resulting in a sharp increase in the number of network edge devices,and the amount of data generated by such devices has reached the ZB level.The current centralized processing model with the cloud computing model as the core can no longer meet such huge data transmission,calculation,and processing needs,mainly manifested in: shortage of computing resources,and the expansion speed of cloud data center resources cannot keep up with the exponential growth of edge layer data speed.Network congestion,equipment at the edge layer needs to communicate with the cloud server across the core network,and massive data is aggregated in the core network,which greatly increases the bandwidth load,and the large transmission delay reduces the service experience of Internet users worldwide.Transmitting data across the core network increases the risk of user data leakage,and centralized storage cannot meet enterprise data privatization and data security regulations.To this end,a distributed processing model centered on the edge computing model emerged at the historic moment.By deploying an edge server at the user edge,the computing resources are close to the user,and the user is provided with the nearest computing and storage services.The computing and storage capacity of the edge server is much higher than the user's terminal equipment,which can generate lower task processing delay.Compared with the server in the cloud data center,the presence of the edge server greatly shortens the distance between the user and the computing resources,reduces the bandwidth overhead,and also reduces the task transmission delay,which can effectively improve the user's service quality.The nearby storage and distributed storage mechanism of data can further ensure the privacy and security of data.Therefore,the processing mode of edge computing can better solve a series of problems in the cloud computing architecture.The proposal of edge computing model makes the task scheduling model change from end cloud scheduling to end cloud scheduling.The change of scheduling model also promotes the improvement and adaptation of scheduling algorithms.In the previous research on edge computing task scheduling,the computing resources of terminal devices are often ignored,and the tasks are all offloaded to the edge layer and cloud layer.However,with the development of hardware devices,most terminal devices have certain computing resources.Therefore,this paper fully considers the computing resources of the terminal layer and offloads the tasks reasonably to the terminal local or edge server.For the tasks that the edge layer cannot complete,this article sets all the tasks to be offloaded to the cloud data center for processing.In Software Defined Network(SDN),the SDN server has the characteristics of a global network view.This paper proposes two strategies for collaborative scheduling of edge tasks that meet the minimum delay of tasks.The first is a single-task scheduling strategy based on greedy algorithm.The strategy is to minimize the offload overhead of a single task as an optimization goal,to meet the minimum delay of a single task as a constraint,according to the principle of first-come-first-served(FCFS)Schedule tasks.The second is a multi-task joint scheduling strategy based on improved particle swarm optimization algorithm.This strategy will cluster task offload requests that arrive within a period of time before task scheduling,and set scheduling priority for multiple task clusters generated by clustering After that,the optimization goal is to minimize the total offload cost of multiple tasks in the task cluster,and to meet the minimum required delay of multiple tasks in the task cluster as a constraint to establish a joint optimization system model.The experimental results show that the two strategies proposed in this paper can effectively reduce the system's service overhead,reduce the task offload delay,and improve the task processing success rate.For the task-removable type of application,this paper proposes a fine-grained side-end collaborative scheduling strategy that can dynamically adjust task disassembly.This strategy is based on an improved genetic algorithm,which can flexibly adjust the fine-grained task disassembly according to the current resource utilization status of the system,and minimize the processing delay of the task as the optimization goal,and make the most suitable allocation plan for the subtasks after disassembly.Simulation experiments show that the algorithm proposed in this paper can effectively reduce the task processing delay and system energy consumption,and ensure that the resources of the edge layer can be effectively used.