Cloud systems have become an essential part of our daily lives owing to various Internet-based services. Consequently, their energy utilization has also become a necessary concern in cloud computing systems increasingly. Live migration, including several virtual machines (VMs) packed on in minimal physical machines (PMs) as virtual machines consolidation (VMC) technique, is an approach to optimize power consumption. In this article, we have proposed an energy-aware method for the VMC problem, which is called energy-aware virtual machines consolidation (EVMC), to optimize the energy consumption regarding the quality of service guarantee, which comprises: (1) the support vector machine classification method based on the utilization rate of all resource of PMs that is used for PM detection in terms of the amount' load; (2) the modified minimization of migration approach which is used for VM selection; (3) the modified particle swarm optimization which is implemented for VM placement. Also, the evaluation of the functional requirements of the method is presented by the formal method and the non-functional requirements by simulation. Finally, in contrast to the standard greedy algorithms such as modified best fit decreasing, the EVMC decreases the active PMs and migration of VMs, respectively, 30%, 50% on average. Also, it is more efficient for the energy 30% on average, resources and the balance degree 15% on average in the cloud.

中文翻译：

---

一种面向云计算的能量感知虚拟机整合方法：仿真与验证

由于各种基于互联网的服务，云系统已成为我们日常生活中必不可少的一部分。因此，它们的能源利用也日益成为云计算系统的必要关注点。实时迁移，包括作为虚拟机整合 (VMC) 技术打包在最小物理机 (PM) 中的多个虚拟机 (VM)，是一种优化功耗的方法。在本文中，我们针对VMC问题提出了一种能量感知方法，称为能量感知虚拟机整合（EVMC），以优化服务质量保证方面的能耗，包括：（1）支持基于用于PM检测的PM全部资源利用率的向量机分类方法，按负载量计算；(2) 改进的最小化迁移方法，用于 VM 选择；(3) 改进后的粒子群优化，用于 VM 放置。此外，该方法的功能需求的评估通过形式化方法和非功能性需求通过模拟呈现。最后，与标准的贪心算法（例如修改后的最佳拟合减少）相比，EVMC 分别减少了活动 PM 和 VM 的迁移，平均分别减少了 30% 和 50%。此外，云中能源平均30%，资源平均15%，平衡度更高效。该方法的功能需求评估通过形式化方法呈现，非功能需求通过模拟呈现。最后，与标准的贪心算法（例如修改后的最佳拟合减少）相比，EVMC 分别减少了活动 PM 和 VM 的迁移，平均分别减少了 30% 和 50%。此外，云中能源平均30%，资源平均15%，平衡度更高效。该方法的功能需求评估通过形式化方法呈现，非功能需求通过模拟呈现。最后，与标准的贪心算法（例如修改后的最佳拟合减少）相比，EVMC 分别减少了活动 PM 和 VM 的迁移，平均分别减少了 30% 和 50%。此外，云中能源平均30%，资源平均15%，平衡度更高效。