Abstract
Data center power consumption is growing to unprecedented levels: the EPA estimates U.S. data centers will consume 100 billion kilowatt hours annually by 2011. Much of this energy is wasted in idle systems: in typical deployments, server utilization is below 30%, but idle servers still consume 60% of their peak power draw. Typical idle periods---though frequent---last seconds or less, confounding simple energy-conservation approaches.
In this article, we propose PowerNap, an energy-conservation approach where the entire system transitions rapidly between a high-performance active state and a near-zero-power idle state in response to instantaneous load. Rather than requiring fine-grained power-performance states and complex load-proportional operation from individual system components, PowerNap instead calls for minimizing idle power and transition time, which are simpler optimization goals. Based on the PowerNap concept, we develop requirements and outline mechanisms to eliminate idle power waste in enterprise blade servers. Because PowerNap operates in low-efficiency regions of current blade center power supplies, we introduce the Redundant Array for Inexpensive Load Sharing (RAILS), a power provisioning approach that provides high conversion efficiency across the entire range of PowerNap’s power demands. Using utilization traces collected from enterprise-scale commercial deployments, we demonstrate that, together, PowerNap and RAILS reduce average server power consumption by 74%.
- Amur, H., Nathuji, R., Ghosh, M., Schwan, K., and Lee, H.-H. S. 2008. IdlePower: Application-aware management of processor idle states. In Proceedings of the Workshop on Managed Many-Core Systems (MMCS’08).Google Scholar
- Barroso, L. A., Dean, J., and Hölzle, U. 2003. Web search for a planet: The google cluster architecture. IEEE Micro. 23, 2, 22--28. Google Scholar
Digital Library
- Barroso, L. A. and Hölzle, U. 2007. The case for energy-proportional computing. Computer 40, 12. Google Scholar
Digital Library
- Bash, C. and Forman, G. 2007. Cool job allocation: Measuring the power savings of placing jobs at cooling-efficient locations in the data center. In Proceedings of the USENIX Annual Technical Conference (USENIX’07). Google Scholar
Digital Library
- Bohrer, P., Elnozahy, E., Keller, T., Kistler, M., Lefurgy, C., and Rajamony, R. 2002. The case for power management in web servers. In Power Aware Computing. Google Scholar
Digital Library
- Chase, J., Anderson, D., Thakar, P., and Vahdat, A. 2001. Managing energy and server resources in hosting centers. In Proceedings of the 18th Symposium on Operating Systems Principles (SOSP’01). Google Scholar
Digital Library
- Delaluz, V., Kandemir, M., Vijaykrishnan, N., Sivasubramaniam, A., and Irwin, M. 2001. DRAM energy management using software and hardware directed power mode control. In Proceedings of the 7th International Symposium on High-Performance Computer Architecture (HPCA’01). IEEE Computer Society, 0159. Google Scholar
Digital Library
- Diniz, B., Guedes, D., Wagner Meira, J., and Bianchini, R. 2007. Limiting the power consumption of main memory. In Proceedings of the 34th Annual International Symposium on Computer Architecture (ISCA’07). ACM, New York, 290--301. Google Scholar
Digital Library
- ECOS and EPR. 2008. Efficient power supplies for data center. Tech. rep., ECOS and EPR.Google Scholar
- Elnozahy, E. N., Kistler, M., and Rajamony, R. 2003. Energy conservation policies for web servers. Proceedings of the USENIX Symposium on Internet Technologies and Systems. Google Scholar
Digital Library
- Fan, X., Weber, W.-D., and Barroso, L. A. 2007. Power provisioning for a warehouse-sized computer. In Proceedings of the 34th Annual International Symposium on Computer Architecture (ISCA’07). ACM, New York, 13--23. Google Scholar
Digital Library
- Hölzle, U. and Weihl, B. 2006. High-efficiency power supplies for home computers and servers. Tech. rep., Google.Google Scholar
- Huang, H., Shin, K. G., Lefurgy, C., and Keller, T. 2005. Improving energy efficiency by making DRAM less randomly accessed. In Proceedings of the International Symposium on Low Power Electronics and Design (ISLPED’05). ACM, New York, 393--398. Google Scholar
Digital Library
- Hynix. 2008. Hynix-DDR2-1Gb.http://www.hynix.com/datasheet/pdf/HY5PS1G4(8.16)31A(L)FP(Rev0.7).pdf.Google Scholar
- Intel. 2005. Intel Pentium M processor with 2-MB L2 cache and 533-MHz front side bus. http://download.intel.com/support/processors/mobile/pm/sb/30526202.pdf.Google Scholar
- Intel. 2007. Intel Pentium dual-core mobile processor. ftp://download.intel.com/design/mobile/datashts/31651903.pdf.Google Scholar
- Intel. 2008. Quad-core Intel Xeon processor 5400 series. http://www.intel.com/Assets/PDF/prodbrief/xeon-5400.pdf.Google Scholar
- Intel. 2009. Intel 64 and ia-32 architectures software developer’s manual volume 3b: System programming guide. http://www.intel.com/Assets/ja_JP/PDF/manual/253668.pdf.Google Scholar
- Laudon, J. 2006. UltraSPARC T1: A 32-threaded CMP for servers. Invited talk.Google Scholar
- Lebeck, A. R., Fan, X., Zeng, H., and Ellis, C. 2000. Power aware page allocation. SIGOPS Oper. Syst. Rev. 34, 5, 105--116. Google Scholar
Digital Library
- Lefurgy, C., Rajamani, K., Rawson, F., Felter, W., Kistler, M., and Keller, T. W. 2003. Energy management for commercial servers. IEEE Computer 36, 12. Google Scholar
Digital Library
- Lefurgy, C., Wang, X., and Ware, M. 2007. Server-level power control. In Proceedings of the 4th IEEE International Conference on Autonomic Computing (ICAC’07). Google Scholar
Digital Library
- Leigh, K. and Ranganathan, P. 2007. Blades as a general-purpose infrastructure for future system architectures: Challenges and solutions. Tech. rep. HPL-2006-182, HPLABS.Google Scholar
- Memcached 2010. Memcached - a distributed memory object caching system. http://www.memcached.org/.Google Scholar
- Micron. 2004. DDR2 SDRAM SODIMM. http://download.micron.com./pdf/datasheets/modules/ddr2/HTF16C128_256x64H.pdf.Google Scholar
- Miyoshi, A., Lefurgy, C., Hensbergen, E. V., Rajamony, R., and Rajkumar, R. 2002. Critical power slope: Understanding the runtime effects of frequency scaling. In Proceedings of the 16th Annual ACM International Conference on Supercomputing (ICS’02). Google Scholar
Digital Library
- Moore, J., Chase, J., Ranganathan, P., and Sharma, R. 2005. Making scheduling ‘cool’: Temperature-aware workload placement in data centers. In Proceedings of the USENIX Annual Technical Conference (USENIX’05). Google Scholar
Digital Library
- National Semiconductor. 2002. Introduction to power supplies. Tech. rep. AN-556, National Semiconductor.Google Scholar
- Padala, P., Zhu, X., Wanf, Z., Singhal, S., and Shin, K. 2007. Performance evaluation of virtualization technologies for server consolidation. Tech. rep. HPL-2007-59, HP/LABS.Google Scholar
- Pandey, V., Jiang, W., Zhou, Y., and Bianchini, R. 2006. DMA-aware memory energy management. In Proceedings of the 12th International Symposium on High-Performance Computer Architecture.Google Scholar
- Rasmussen, N. 2007. AC vs. DC power distribution for data centers. Tech. rep. #63, American Power Conversion.Google Scholar
- Samsung. 2008. SSD SATA 3.0Gbps 2.5 data sheet. http://www.samsung.com/global/system/business/semiconductor/product/2008/6/19/ 82925725_datasheet.pdf.Google Scholar
- Siddha, S., Pallipadi, V., and Ven, A. V. D. 2007. Getting maximum mileage out of tickless. In Proceedings of the LINUX Conference (LINUX’07).Google Scholar
- SMSC. 2008. LAN9420/LAN9420i single-chip ethernet controller with HP Auto-MDIX support and PCI interface. http://www.smsc.com/media/Downloads_Public/Data_Sheets/9420.pdf.Google Scholar
- Standard Performance Evaluation Corporation. 2005. Specweb2005. http://www.spec.org/web2005/.Google Scholar
- Standard Performance Evaluation Corporation. 2008. Specpower2008. http://www.spec.org/power_ssj2008/.Google Scholar
- Tolia, N., Wang, Z., Marwah, M., Bash, C., Ranganathan, P., and Zhu, X. 2008. Delivering energy proportionality with non energy-proportional systems -- optimizing the ensemble. In Proceedings of the USENIX Workshop on Power Aware Computing and Systems (HotPower’08). Google Scholar
Digital Library
- U.S. EPA. 2007a. Energy Star computer specification v. 4.0. Tech. rep., US EPA.Google Scholar
- U.S. EPA. 2007b. Report to congress on server and data center energy efficiency. Tech. rep., US EPA.Google Scholar
- U.S. Official Information Administration. 2008. Average retail price of electricity to ultimate customers by end-use sector, by state.Google Scholar
- Weber, W.-D. 2010. Energy-saving approaches for warehouse-scale computing. ISCAS Keynote. In Proceedings of the IEEE International Symposium on Circuits and Systems (ISCAS’10).Google Scholar
- Welch, P. D. 1964. On a generalized M/G/1 queuing process in which the first customer of each busy period receives exceptional service. Oper. Res. 12, 736--752.Google Scholar
Digital Library
- Wu, Q., Juang, P., Martonosi, M., Peh, L., and Clark, D. 2005. Formal control techniques for power-performance management. IEEE MICRO 5. Google Scholar
Digital Library
Index Terms
The PowerNap Server Architecture
Recommendations
PowerNap: eliminating server idle power
ASPLOS 2009Data center power consumption is growing to unprecedented levels: the EPA estimates U.S. data centers will consume 100 billion kilowatt hours annually by 2011. Much of this energy is wasted in idle systems: in typical deployments, server utilization is ...
PowerNap: eliminating server idle power
ASPLOS XIV: Proceedings of the 14th international conference on Architectural support for programming languages and operating systemsData center power consumption is growing to unprecedented levels: the EPA estimates U.S. data centers will consume 100 billion kilowatt hours annually by 2011. Much of this energy is wasted in idle systems: in typical deployments, server utilization is ...
PowerNap: eliminating server idle power
ASPLOS 2009Data center power consumption is growing to unprecedented levels: the EPA estimates U.S. data centers will consume 100 billion kilowatt hours annually by 2011. Much of this energy is wasted in idle systems: in typical deployments, server utilization is ...






Comments