Dynamic Task Scheduling Strategy for Cloud-Edge Collaborative Environment

Dynamic Task Scheduling Strategy for Cloud-Edge Collaborative Environment

연구 분야: Software Development

학회: China Conference on Wireless Sensor Networks

With the growth of IoT and edge computing, Kubernetes and KubeEdge offer cloud-edge collaboration opportunities but face challenges like limited edge node resources, geographical distribution, and unstable communication links. While previous studies mainly focused on efficient resource utilization in common scenarios, but ignored the impact of a large number of short-term tasks on resource utilization. To address this, this paper proposes a dynamic task scheduling strategy (DTSS) that aims to improve resource utilization in clusters under mixed-task scenarios, reduce cluster resource disparity (CRD), and minimize resource fragmentation. DTSS divides tasks into long-term and short-term categories. For short-term tasks, it maintains load balance while minimizing load imbalance duration, reducing cluster resource volatility. For long-term tasks, DTSS focuses on optimizing resource balance across the cluster. The DTSS algorithm is validated through comparative experiments, showing it reduces cluster resource disparity by an average of 68.06% compared to Kubernetes’ default scheduling algorithm.