Load-sensitive CPU Power Management for Web Search Engines
Catena M., Macdonald C., Tonellotto N.
search engines cpu frequency scaling power consumption
Web search engine companies require power-hungry data centers with thousands of servers to efficiently perform searches on a large scale. This permits the search engines to serve high arrival rates of user queries with low latency, but poses economical and environmental concerns due to the power consumption of the servers. Existing power saving techniques sacrifice the raw performance of a server for reduced power absorption, by scaling the frequency of the server's CPU according to its utilization. For instance, current Linux kernels include frequency governors i.e., mechanisms designed to dynamically throttle the CPU operational frequency. However, such general-domain techniques work at the operating system level and have no knowledge about the querying operations of the server. In this work, we propose to delegate CPU power management to search engine-specific governors. These can leverage knowledge coming from the querying operations, such as the query server utilization and load. By exploiting such additional knowledge, we can appropriately throttle the CPU frequency thereby reducing the query server power consumption. Experiments are conducted upon the TREC ClueWeb09 corpus and the query stream from the MSN 2006 query log. Results show that we can reduce up to ~24% a server power consumption, with only limited drawbacks in effectiveness w.r.t. a system running at maximum CPU frequency to promote query processing quality.
Source: 38th International ACM SIGIR Conference on Research and Development in Information Retrieval, pp. 751–754, Santiago del Cile, 09-13/08/2015
[2] A. Z. Broder, D. Carmel, M. Herscovici, A. So er, and J. Zien. E cient query evaluation using a two-level retrieval process. In CIKM, 2003, pages 426{434.
[3] D. Brodowski. CPU frequency and voltage scaling code in the Linux kernel: Linux CPUFreq. https://www.kernel. org/doc/Documentation/cpu-freq/index.txt, Visited: 2015-02-16.
[4] S. Buttcher, C. L. A. Clarke and G. V. Cormack. Information retrieval: Implementing and evaluating search engines, 2010, MIT Press.
[5] G. Chowdhury. An agenda for green information retrieval research. Information Processing & Management, 48(6):1067{1077, 2012.
[6] A. Freire, C. Macdonald, N. Tonellotto, I. Ounis, and F. Cacheda. A Self-adapting Latency/Power Tradeo Model for Replicated Search Engines. In WSDM, 2014, pages 13{22.
[7] M. Harchol-Balter. Performance Modeling and Design of Computer Systems: Queueing Theory in Action. Cambridge University Press, 2013.
[8] N. Hidalgo, E. Rosas, V. Gil-Costa, and M. Marin. Assessing Energy E ciency in ISP and Web Search Engine Collaboration. In WAINA, 2014, pages 299{304.
[9] U. Hoelzle and L. A. Barroso. The Datacenter As a Computer: An Introduction to the Design of Warehouse-Scale Machines. Morgan and Claypool Publishers, 1st edition, 2009.
[10] E. Kayaaslan, B. B. Cambazoglu, R. Blanco, F. P. Junqueira, and C. Aykanat. Energy-price-driven Query Processing in Multi-center Web Search Engines. In SIGIR, 2011, pages 983{992.
[11] D. Lo, L. Cheng, R. Govindaraju, L. A. Barroso and C. Kozyrakis. Towards energy proportionality for large-scale latency-critical workloads. In ISCA, 2014, pages 301{312.
[12] M. Marin, V. Gil-Costa, and C. Gomez-Pantoja. New Caching Techniques for Web Search Engines. In HPDC, 2010, pages 215{226.
[13] C. Macdonald, R. McCreadie, R. Santos, and I. Ounis. From Puppy to Maturity: Experiences in Developing Terrier. In OSIR Workshop, 2012.
[14] F. B. Sazoglu, B. B. Cambazoglu, R. Ozcan, I. S. Altingovde, and O. Ulusoy. A Financial Cost Metric for Result Caching. In SIGIR, 2013, pages 873{876.
[15] E. Schurman and J. Brutlag. Performance related changes and their user impact. In Proc. Velocity, 2009, page 1.
[16] F. Silvestri. Mining Query Logs: Turning Search Usage Data into Knowledge. Foundations and Trends in IR, Now Publishers Inc., 2010.
[17] D. C. Snowdon, S. Ruocco, and G. Heiser. Power Management and Dynamic Voltage Scaling: Myths and Facts. In PARC Workshop, 2005.
[18] A. Teymorian, O. Frieder, and M. A. Maloof. Rank-energy Selective Query Forwarding for Distributed Search Systems. In CIKM, 2013, pages 389{398.
[19] N. Tonellotto, C. Macdonald, and I. Ounis. E cient and e ective retrieval using selective pruning. In WSDM, 2013, pages 63{72.
[20] S. Vigna. Quasi-succinct indices. In WSDM, 2013, pages 83{92.
Metrics
Back to previous page
@inproceedings{oai:it.cnr:prodotti:344447,
title = {Load-sensitive CPU Power Management for Web Search Engines},
author = {Catena M. and Macdonald C. and Tonellotto N.},
doi = {10.1145/2766462.2767809},
booktitle = {38th International ACM SIGIR Conference on Research and Development in Information Retrieval, pp. 751–754, Santiago del Cile, 09-13/08/2015},
year = {2015}
}
0000-0002-7427-1001ISTI Repository
DOIAlso available from
