A QoS-Guaranteed Energy-Efficient VM Dynamic Migration Strategy in Cloud Data Centers
Hao Cao, Hongguang Sun, Min Sheng, Yan Shi, Jiandong Li · 2018
To reduce the energy consumption caused by cloud computing in cloud data centers, virtual machine (VM) dynamic migration technology has been widely used. With the VM dynamic migration technology, VMs can be migrated and consolidated dynamically, as such the number of active physical machines (PMs) is reduced, leading to the considerable energy saving. In this paper, we propose a QoS-Guaranteed energy-efficient VM dynamic migration strategy, by taking the user quality of service (QoS) requirements and system load into account. With a more accurate energy consumption model, we first study the effect of system load and user QoS differentiation on the total energy consumption, and then define the migration threshold as a function of the variance of user QoS requirements and system load. We observe that an adaptive migration threshold is beneficial to minimize the total energy consumption, which should be set to be decreased with the growth of system load to lower the variance of user QoS requirements within each PM. Simulation results show the effectiveness of our proposed strategy in energy saving by comparing with the existing algorithms.