虚拟机迁移中一种新的物理主机异常状态检测算法

PHSDA: A new physical host status anomalous detection algorithm in virtual machine migration

摘要:

提出了一种新的物理主机异常状态检测算法PHSDA(Physical host status anomalous detection algorithm)。PHSDA算法包括两个阶段；在超负载检测中,采用一种迭代权重线性回归方法来预测物理资源的使用效率情况；在低负载检测中,利用多维物理资源的均方根来确定其资源使用阈值下限,避免异常状态的物理主机数量的增加; PHSDA检测算法配合迁移过程中后续的虚拟机选择策略和虚拟机放置策略,就可以形成一个全新的虚拟机迁移模型PHSDA-MMT-BFD。以CloudSim模拟器作为PHSDA的仿真环境。经PHSDA策略优化过后的新虚拟机迁移实验表明：比近几年的BenchMark迁移模型比较起来,可以很好的降低云数据中心的能量消耗,虚拟机迁移次数减少,云服务质量明显提高。

Abstract:

A new physical host status anomalous detection algorithm called PHSDA was proposed in this paper. PHSDA includes two phases, overloading host detection and under loading host detection. In overloading host detection, it used an iterative weighted linear regression method to determine two utilization thresholds and avoid performance degradation. In under loading host detection, PHSDA used a vector magnitude squared of multiple resources to consolidate active hosts. United with Subsequent strategy in virtual machine selection and virtual machine placement, a novel virtual machine migration model called PHSDA-MMT-BFD had been formed. PHSDA had been evaluated using CloudSim tools. Experimental results show that compared with benchmark VM migration models, energy consumption and numbers of VM migration had been reduced in PHSDA. The Qos of cloud service provider had been also promoted.