Task Scheduling Strategy Based on Resource Constraint in Edge Computing System

With the advent of the era of the Internet of Everything, edge computing, characterized by its high reliability and low latency, has been widely applied in various intelligent systems. This paper addresses the challenge of optimizing task execution in edge systems, which differ from traditional cloud computing with their heterogeneous nodes and varied task demands. We introduce a task-scheduling approach focused on resource constraints. Our method involves analyzing the effects of network IO and CPU resources under soft constraints, developing a task performance prediction model based on single resource pressure testing, and implementing a throughput maximization scheduling (MTSched) strategy. Our experimental analysis on a custom-built edge heterogeneous cluster demonstrates that MTSched significantly outperforms three benchmark algorithms (random, round-robin, and weighted round-robin scheduling) in system throughput and task execution delay.