Large data centers are well known for their high-energy intensity and have made dramatic efficiency improvements over the past decade. Small closet and room data centers have received much less attention, yet constitute a significant fraction of the total number of servers in the United States. The often makeshift, ad hoc nature of small data centers often results in little attention paid to energy efficiency and inadequate cooling equipment. The small physical footprint of these data centers, typically embedded within a larger building, makes it difficult to identify and target for efficiency measures. These conditions make small data centers notoriously inefficient relative to their larger counterparts. In this report, we present an analysis of small and midsize data centers in the US, drawing from surveys of commercial building stock. We find that servers in small data centers make up approximately 40% of installed server stock, with the vast majority of sites utilizing only 1-2 servers. We identify industries where small data centers are most prevalent, finding that the highest saturations are in medical, retail, office, and education sectors. Small data centers typically lack dedicated cooling equipment, often relying on building air conditioning and ventilation equipment for cooling. We further find that the type of cooling equipment used is highly correlated with the number of operational server racks, with less efficient cooling options used with fewer racks. We develop geospatial maps of small and midsize data centers to visually identify regions of high server concentration and calculate associated CO2 emissions. Small data centers consume 13 billion kWh of energy annually, emitting 7 million metric tons (MMT) of carbon dioxide–the equivalent emissions of approximately 2.3 coal-fired plants. We discuss efficiency measures that could be implemented and estimate potential energy and CO2 savings.

Servers’ field energy use remains poorly understood, given heterogeneous computing loads, configurable hardware and software, and operation over a wide range of management practices. This paper explores various characteristics of 1- and 2-socket volume servers that affect energy consumption, and quantifies the difference in power demand between higher-performing SPEC and ENERGY STAR servers and our best understanding of a typical server operating today. We first establish general characteristics of the U.S. installed base of volume servers from existing IDC data and the literature, before presenting information on server hardware configurations from data collection events at a major online retail website. We then compare cumulative distribution functions of server idle power across three separate datasets and explain the differences between them via examination of the hardware characteristics to which power draw is most sensitive. We find that idle server power demand is significantly higher than ENERGY STAR benchmarks and the industry-released energy use documented in SPEC, and that SPEC server configurations—and likely the associated power-scaling trends—are atypical of volume servers. Next, we examine recent trends in server power draw among high-performing servers across their full load range to consider how representative these trends are of all volume servers before inputting weighted average idle power load values into a recently published model of national server energy use. Finally, we present results from two surveys of IT managers (n=216) and IT vendors (n=178) that illustrate the prevalence of more-efficient equipment and operational practices in server rooms and closets; these findings highlight opportunities to improve the energy efficiency of the U.S. server stock.

We present extensive photometry at ultraviolet (UV), optical, and near-infrared (NIR) wavelengths, as well as dense sampling of optical spectra, for the normal Type Ia supernova (SN Ia) 2005cf. The optical photometry, performed at eight different telescopes, shows a 1σ scatter of ≲0.03 mag after proper corrections for the instrument responses. From the well-sampled light curves, we find that SN 2005cf reached a B-band maximum at 13.63 0.02 mag, with an observed luminosity decline rate Δm 15(B) = 1.05 0.03 mag. The correlations between the decline rate and various color indexes, recalibrated on the basis of an expanded SN Ia sample, yield a consistent estimate for the host-galaxy reddening of SN 2005cf, E(B - V)host = 0.10 0.03 mag. The UV photometry was obtained with the Hubble Space Telescope and the Swift Ultraviolet/Optical Telescope, and the results match each other to within 0.1-0.2 mag. The UV light curves show similar evolution to the broadband U, with an exception in the 2000-2500 spectral range (corresponding to the F220W/uvm2 filters), where the light curve appears broader and much fainter than that on either side (likely owing to the intrinsic spectral evolution). Combining the UV data with the ground-based optical and NIR data, we establish the generic UV-optical-NIR bolometric light curve for SN 2005cf and derive the bolometric corrections in the absence of UV and/or NIR data. The overall spectral evolution of SN 2005cf is similar to that of a normal SN Ia, but with variety in the strength and profile of the main feature lines. The spectra at early times displayed strong, high-velocity (HV) features in the CaII H&K doublet and NIR triplet, which were distinctly detached from the photosphere (v 10,000 km s-1) at a velocity ranging from 20,000 to 25,000 km s -1. One interesting feature is the flat-bottomed absorption observed near 6000 in the earliest spectrum, which rapidly evolved into a triangular shape and then became a normal SiII λ6355 absorption profile at about one week before maximum brightness. This premaximum spectral evolution is perhaps due to the blending of the SiIIλ6355 at photospheric velocity and another HV absorption component (e.g., an SiII shell at a velocity 18,000 km s -1) in the outer ejecta, and may be common in other normal SNe Ia. The possible origin of the HV absorption features is briefly discussed. © 2009. The American Astronomical Society.