Computer Engineering and Applications ›› 2014, Vol. 50 ›› Issue (5): 51-54.

Previous Articles     Next Articles

Research of cloud resource dynamic scheduling strategy on migration technology

GONG Suwen1, AI Haojun2, YUAN Yuanming2   

  1. 1.School of Electrical Engineering, Jiujiang Vocational & Technical College, Jiujiang, Jiangxi 332007, China
    2.School of Computer, Wuhan University, Wuhan 430072, China
  • Online:2014-03-01 Published:2015-05-12

基于迁移技术的云资源动态调度策略研究

龚素文1,艾浩军2,袁远明2   

  1. 1.九江职业技术学院 电气工程学院,江西 九江 332007
    2.武汉大学 计算机学院,武汉 430072

Abstract: Currently, cloud resource management platform exists problems that instantaneous peak value of resource utilization easily leads to migration and dynamic load effect is not good. According to given cloud resources dynamic scheduling model, this paper proposes a virtual machine dynamic scheduling algorithm on migration technology. The algorithm combines with the physical node load evaluation and the virtual machine migration loss evaluation, multiple trigger control, the destination node localization, in order to achieve an efficient dynamic load balance of a cloud computing data center. The experimental results show that compared to dynamic voltage and frequency scaling strategy of the CloudSim, the proposed algorithm ensures a better application service level.

Key words: migration technology, cloud recourse, dynamic scheduling

摘要: 现有云资源管理平台存在着瞬时资源利用率峰值易引发迁移、动态负载效果不佳等问题。依据云资源动态调度模型,提出了有效的基于迁移技术的虚拟机动态调度算法。算法将物理节点负载与虚拟机迁移损耗评估、多次触发控制、目标节点定位三者有机结合,实现云计算数据中心高效的动态负载均衡。实验结果表明,该算法优于CloudSim的DVFS调度策略,在保证应用服务水平的同时能减少虚拟机迁移次数和物理机启用数量。

关键词: 迁移技术, 资源, 动态调度