云数据中心高能效的虚拟机放置算法

发布时间：2019-06-17 16:52

【摘要】：面对云数据中心高能耗的挑战,以节能为目标的虚拟机放置算法成为研究热点.现有研究大多只考虑CPU一种系统资源对能效的影响,并且多采用基于贪心算法的传统启发式算法进行虚拟机放置.已有研究开始考虑多种系统资源对能效的影响,并且提出了多资源的能效模型,但是在多资源能效模型下虚拟机放置算法的研究还未引起关注.本文根据多资源的能效模型提出了基于粒子群算法的高能效虚拟机放置算法,包括采用首次适应算法生成粒子,定义粒子个体最优解和全局最优解的信任度指导粒子进化,根据多资源能效模型定义粒子群的适应度函数并以此来评价粒子.仿真实验结果表明,与传统启发式算法相比较,该算法使虚拟机的放置结果更接近系统能效的最佳状态,同时也有效地提高了系统资源的利用率.
[Abstract]:In the face of the challenge of high energy consumption in the cloud data center, the virtual machine placement algorithm with energy saving as the target becomes the research focus. Most of the existing researches only consider the effect of CPU on the energy efficiency, and the traditional heuristic algorithm based on the greedy algorithm is used for virtual machine placement. In this paper, the effect of multiple system resources on energy efficiency has been considered, and the energy efficiency model of multi-resource is put forward, but the research of the virtual machine placement algorithm under the multi-resource energy efficiency model has not attracted the attention. Based on the energy efficiency model of multi-resource, a high-efficiency virtual machine placement algorithm based on particle swarm optimization is proposed, which includes the following steps: using the first adaptive algorithm to generate the particles, defining the optimal solution of the particle and the global optimal solution to guide the particle evolution, According to the multi-resource energy efficiency model, the fitness function of the particle swarm is defined and the particles are evaluated. The simulation results show that, compared with the traditional heuristic algorithm, the algorithm makes the placement result of the virtual machine closer to the optimal state of the energy efficiency of the system, and also effectively improves the utilization rate of the system resources.
【作者单位】：重庆邮电大学计算机科学与技术学院;
【基金】：国家自然科学基金项目(21203259)资助重庆市教委科学技术研究项目(KJ130514)资助重庆邮电大学自然科学基金项目(A2012-31)资助
【分类号】：TP308

【参考文献】