The Case for EnergyProportional Computing（EnergyProportional计算）.pdf

C O V E R F E A T U R E The Case for Energy-Proportional Computing Luiz André Barroso and Urs Hölzle Google Energy-proportional designs would enable large energy savings in servers, potentially doubling their efficiency in real-life use. Achieving energy proportionality will require significant improvements in the energy usage profile of every system component, particularly the memory and disk subsystems. nergy efficiency, a new focus for general-purpose trends could make energy a dominant factor in the total computing, has been a major technology driver cost of ownership.3 Besides the server electricity bill, in the mobile and embedded areas for some time. TCO includes other energy-dependent components such Earlier work emphasized extending battery life, as the cost of energy for the cooling infrastructure and Ebut it has since expanded to include peak power provisioning costs, specifically the data center infra- reduction because thermal constraints began to limit fur- structure’s cost. To a first-order approximation, both ther CPU performance improvements. cooling and provisioning costs are proportional to the Energy management has now become a key issue for average energy that servers consume, therefore energy servers and data center operations, focusing on the efficiency improvements should benefit all energy-depen- reduction of all energy-related costs, including capital, dent TCO components. operating expenses, and environmental impacts. Many Efforts such as the Climate Savers Computing Initiative energy-saving techniques developed for mobile devices () could help lower became natural candidates for tackling this new problem worldwide computer energy consumption by promoting space. Although servers clearly provide many parallels widespread adoption of h