Computer Engineering and Applications ›› 2022, Vol. 58 ›› Issue (20): 28-42.DOI: 10.3778/j.issn.1002-8331.2202-0243

• Research Hotspots and Reviews • Previous Articles     Next Articles

Review of Research on Computing-Intensive Task Scheduling in Edge Environments

LIU Yanpei, ZHU Yunjing, BIN Yanru, CHEN Ningning, WANG Liping   

  1. College of Computer and Communication Engineering, Zhengzhou University of Light Industry, Zhengzhou 450001, China
  • Online:2022-10-15 Published:2022-10-15

边缘环境下计算密集型任务调度研究综述

刘炎培,朱运静,宾艳茹,陈宁宁,王丽萍   

  1. 郑州轻工业大学 计算机与通信工程学院,郑州 450001

Abstract: With the dramatic increase in the number of mobile devices and the widespread use of computing-intensive applications such as face recognition, internet of vehicles and virtual reality. In order to achieve the optimal matching of tasks and collaborative resources to meet user QoS requests, using a task scheduling scheme for reasonably computing-intensive applications can solve the problems of extended time, high cost, unbalanced load and low resource utilization in the edge cloud center. Firstly, the scheduling framework, execution process, application scenarios and performance indicators of computing-intensive application tasks in the edge computing environment are described. Secondly, this paper analyzes and compares three task scheduling schemes from the optimization goals of time and cost, energy consumption and resource utilization, load balancing and throughput, and summarizes the advantages, disadvantages and applicable scenarios of these schemes. Then, by analyzing the SDN-based edge computing architecture in the 5G environment, the task scheduling strategy for edge computing-intensive data packet based on SDN, task scheduling strategy for computing-intensive application based on deep reinforcement learning, multi-objective cross-layer task scheduling strategy in 5G IoV network are proposed. At last, the challenges of task scheduling in edge computing environment are summarized from the aspects of fault-tolerant scheduling, dynamic microservice scheduling, crowd aware scheduling, security and privacy.

Key words: edge computing, computing-intensive, task scheduling, software defined network(SDN), scheduling strategy

摘要: 随着移动设备数量的急剧增长及计算密集型应用如人脸识别、车联网以及虚拟现实等的广泛使用,为了实现满足用户QoS请求的任务和协同资源的最优匹配,使用合理的计算密集型应用的任务调度方案,从而解决边缘云中心时延长、成本高、负载不均衡和资源利用率低等问题。阐述了边缘计算环境下计算密集型应用的任务调度框架、执行过程、应用场景及性能指标。从时间和成本、能耗和资源利用率以及负载均衡和吞吐量为优化目标的边缘计算环境下计算密集型应用的任务调度策略进行了对比和分析,并归纳出目前这些策略的优缺点及适用场景。通过分析5G环境下基于SDN的边缘计算架构,提出了基于SDN环境下的边缘计算密集型数据包任务调度策略、基于深度强化学习的计算密集型应用的任务调度策略和5G IoV网络中多目标跨层任务调度策略。从容错调度、动态微服务调度、人群感知调度以及安全和隐私等几个方面总结和归纳了目前边缘计算环境中任务调度所面临的挑战。

关键词: 边缘计算, 计算密集型, 任务调度, 软件定义网络(SDN), 调度策略