Human-centric Computing and Information Sciences · 2016 · 70 citations · 17 references
ComputingComputer EngineeringComputer ScienceComputer ArchitectureOperating SystemsVirtualizationData Center VirtualizationEnergy-efficient Cloud ComputingPower ManagementServer VirtualizationVirtualization ArchitectureCooling System DesignData Center CoolingDatacenter VirtualizationPower ConsumptionPower Consumption OptimizationResource Efficiency DesignResource VirtualizationSystem Virtualization
Electric power consumption in data centers is largely driven by server cooling systems, and it rises as the number of equipment and servers increases. The proposed server operation system aims to lower power usage and CO₂ emissions by reducing equipment count and simplifying the system’s physical layout. Virtualization technology was employed during design and implementation to enhance resource efficiency of the system. Significant energy savings were achieved, and test‑bed experiments with a virtual machine confirmed that virtual hardware performs as efficiently as physical hardware.
Abstract According to current researches, much of the electric power is being consumed by the server cooling system at the Data Center. Moreover, the power consumption rate increases when the number of the equipments and servers expands. Thus, the proposed server operation system has been designed to decrease power consumption rate and CO 2 emission volume by minimizing the number of these equipments and simplifying the physical composition of the system. Virtualization technology was adopted in both designing and implementation phases to improve resource efficiency of the system. As a result, significant amount has been saved while constructing the server operation system in this paper. System’s performance has been evaluated using a virtual machine prior to its practical use through test bed experiments and the results confirms our expectation that the virtual hardwares will work as efficiently as actual ones.
17