项目名称: 云计算环境下数据中心的power capping关键问题研究
项目编号: No.61272460
项目类型: 面上项目
立项/批准年度: 2013
项目学科: 自动化技术、计算机技术
项目作者: 齐勇
作者单位: 西安交通大学
项目金额: 81万元
中文摘要: 在能源日趋紧张的今天,如何有效使用能源已成为各国战略的重点。而随着数据中心的能量消耗越来越大,其相关费用也成为云计算提供商的主要运营成本之一。power capping作为一种降低数据中心电力成本的重要方法,受到工业界和学术界的广泛关注。power capping通过控制数据中心的总能耗低于某个预先指定的值,从而减少供电基础设施的投入成本。以虚拟化为基础的云计算为power capping带来了新的挑战:负载的高度波动性要求能耗控制必须实时、动态的完成;虚拟化要求针对虚拟机这一逻辑单元进行功率分配和控制。本课题以降低数据中心运行成本为目标,研究云计算数据中心的power capping相关问题。内容包括:功耗预算(power budget)建模方法研究、服务级别的power budget动态分配算法研究、虚拟机的能耗测量模型和控制、电力市场环境下成本感知的动态power capping等
中文关键词: 云计算;数据中心;虚拟化;能耗限额;能耗模型
英文摘要: The power consumption of Datacenters increases dramatically. The cost related to power is becoming one of major cost of cloud providers. Power capping which is one of importants methods to reduce electricity cost of datacenters has receive much attention. Power capping reduces infrastructure cost through throttling the maximum power consumption. Vitualized datacenters bring new challenges to power capping. Power capping for virtualized datacenter has new challenges. First, power control should be online and dynamic due to workload fluctuation. Second, we should assign and throttle power of virtual machines. This proposal aims at cost reduction for datacenters and includes: research on power budget modelling, dynamic allocation of power budget at application level, measurement and control for power of virtual machine, and cost-aware power capping under electricity market environment. We provide theory and technology for green datacenters.
英文关键词: Cloud Computing;Data Center;Virtualization;Power Capping;Power Model