Computer Engineering and Applications ›› 2019, Vol. 55 ›› Issue (19): 96-104.DOI: 10.3778/j.issn.1002-8331.1807-0196

Previous Articles     Next Articles

Computing Resource Allocation Scheme Based on Fog Computing

TANG Linyu, JIANG Jiafu, GU Ke   

  1. School of Computer & Communication Engineering, Changsha University of Science & Technology, Changsha 410114, China
  • Online:2019-10-01 Published:2019-09-30

基于雾计算的计算资源分配方案

汤琳煜,蒋加伏,谷科   

  1. 长沙理工大学 计算机与通信工程学院,长沙 410114

Abstract: Fog computing can provide users with data storage, computing and other services, so task scheduling and resource allocation in fog computing have become an emerging research hotspot. Considering that terminal users and fog devices are usually in a relatively open environment, the architecture of fog computing is extended and a computing resource allocation scheme based on stable matching is proposed for open fog computing environment, which utilizes the dynamic computing resource in fog network to provide computing services for users and gains computing rewards while terminal users submit task requests to fog server and need to pay a fee. Based on the stable matching strategy, the priority list of sub-tasks and the preference lists of sub-tasks and computing service devices are established to solve allocation problem between sub-tasks and computing service devices. Finally, the performance of the proposed scheme is analyzed through experiments, the results show that the time of resource allocation is relatively stable and tasks’ delay and violation rate are superior to SGA algorithm and ACOSA algorithm.

Key words: fog computing, task scheduling, computing resource allocation, matching strategy

摘要: 雾计算可以为用户提供近距离的数据存储、计算和其他服务,因此雾计算中的任务调度和资源分配已经成为一个新的研究热点。考虑终端用户和雾设备通常处于一种相对开放的状态,扩展了雾计算的体系结构,提出一种开放式雾计算环境中基于稳定匹配的计算资源分配方案,利用雾网络中动态的计算资源协同为用户提供计算服务并收取计算收益,同时终端用户向雾服务器提交任务请求并支付一定的费用。基于稳定匹配的思想,利用子任务的优先级列表、子任务和计算服务设备的偏好列表解决子任务与计算服务设备的分配问题,保证任务的完成时间和计算服务设备的收益。通过实验对方案性能进行了分析,实验结果表明该方案的资源分配时间相对稳定,且在执行雾计算任务时延以及任务违规率上都优于SGA算法和ACOSA算法。

关键词: 雾计算, 任务调度, 计算资源分配, 匹配策略