一种迁移成本可感知的虚拟机整合算法

发布时间：2018-06-23 09:32

本文选题：云计算 + 基础设施即服务　；参考：《中国科技论文》2017年14期

【摘要】：针对当前IaaS数据中心硬件资源利用率比较低的问题,提出了1种迁移成本可感知的虚拟机整合算法,可以在给定的迁移成本约束条件下完成多维资源虚拟机实例的整合。该算法基于模拟退火算法的思想,通过逐次增加虚拟机的迁移次数迭代产生新的虚拟机整合策略,能够准确地计算出虚拟机整合过程所造成的迁移成本;通过为新的虚拟机整合策略设置接受概率,保证虚拟机整合过程实现开机数减少,资源利用率提高的优化目标。模拟实验的结果表明,所提出的多维资源虚拟机整合算法可以在给定的迁移成本约束条件下将服务器的开机数平均减少4.71%,有效提高IaaS数据中心的硬件资源利用率。该算法对不同的虚拟机资源需求类型同样有效,利用该虚拟机整合算法,有助于降低IaaS服务提供商的基础设施成本和电力成本。
[Abstract]:Aiming at the problem of low utilization of hardware resources in IaaS data center, a migration cost perceptible virtual machine integration algorithm is proposed, which can integrate multi-dimensional resource virtual machine instances under given migration cost constraints. Based on the idea of simulated annealing algorithm, a new strategy of virtual machine integration is generated by increasing the number of times of virtual machine migration step by step, which can accurately calculate the migration cost caused by the process of virtual machine integration. By setting the acceptance probability for the new strategy of virtual machine integration, we can ensure that the process of virtual machine integration can achieve the optimization goal of reducing the number of boot and improving the utilization of resources. The simulation results show that the proposed multi-dimensional resource virtual machine integration algorithm can reduce the average number of servers startup by 4.71 under given migration cost constraints and effectively improve the utilization of hardware resources in IaaS data centers. The algorithm is also effective for different types of virtual machine resource requirements. The integration algorithm can help to reduce the infrastructure and power costs of IaaS service providers.
【作者单位】：西安交通大学电子与信息工程学院;
【基金】：国家自然科学基金资助项目(61272460) 高等学校博士学科点专项科研基金资助项目(20120201110010)
【分类号】：TP302

【相似文献】