Intelligent Scheduling Strategies for Computing Power Resources in Heterogeneous Edge Networks

Abstract

AbstractThe edge computing model enables real-time and low-power processing of data, while contributing to data security and privacy protection. However, the heterogeneity and diversity of edge computing devices pose a great challenge to task scheduling and migration. Most of the existing studies only consider the allocation of computational resources, but lack comprehensive consideration of data resources, storage space, etc. In this paper, we proposed intelligent scheduling strategies for computing power resources in heterogeneous edge networks. We define the relevant models and construct a comprehensive matching matrix in terms of task matching with computing resources, data resources, storage resources, load balancing of computing devices and storage space matching, and design an intelligent scheduling algorithm based on iteration and load balancing according to the matching degree of tasks and computing devices in the heterogeneous edge network environment. The iterative and load-balanced scheduling algorithm is based on the least-cost flow solution scheduling strategy, which effectively reduces the task computation response time and improves the computation and storage resource utilization of computing devices. Experimental validation of the proposed intelligent scheduling strategy is carried out based on a simulation environment. The experimental results show that the proposed intelligent scheduling strategy has obvious advantages over random scheduling methods in terms of task processing delay, computing power resource utilization and number of satisfactory tasks.KeywordsEdge computingResource heterogeneityTask schedulingSystem simulation