Gordon: using flash memory to build fast, power-efficient clusters for data-intensive applications

ABSTRACT

As our society becomes more information-driven, we have begun to amass data at an astounding and accelerating rate. At the same time, power concerns have made it difficult to bring the necessary processing power to bear on querying, processing, and understanding this data. We describe Gordon, a system architecture for data-centric applications that combines low-power processors, flash memory, and data-centric programming systems to improve performance for data-centric applications while reducing power consumption. The paper presents an exhaustive analysis of the design space of Gordon systems, focusing on the trade-offs between power, energy, and performance that Gordon must make. It analyzes the impact of flash-storage and the Gordon architecture on the performance and power efficiency of data-centric applications. It also describes a novel flash translation layer tailored to data intensive workloads and large flash storage arrays. Our data show that, using technologies available in the near future, Gordon systems can out-perform disk-based clusters by 1.5× and deliver up to 2.5× more performance per Watt.

REFERENCES

• 1: http://hadoop.apache.org/core/.
• 2: http://www.fusionio.com/.
• 3: http://xtreview.com/addcomment-id-4801-view-Intel-atom-1.6-Ghzbenchmark. html.
• 4: http://laptoping.com/intel-atom-benchmark.html.
• 5: http://www.vmware.com/.
• 6: DRAMeXchange. http://www.dramexchange.com/.
• 7: Open nand flash interface specification 1.0. http://www.onfi.org/documentation.html.
• 8: Open nand flash interface specification 2.0. http://www.onfi.org/documentation.html.
• 9: Samsung k9f8g08uxm flash memory datasheet.
• 10: Samsung debuts 64gbit mlc nand flash memory. EE Times Asia, October 2007. http://www.eetasia.com/ART 8800485916 499486 NP 11c4687c.HTM.
• 11: New intel centrino atom processor technology ushers in 'best internet experience in your pocket', 2008. Intel Press release.
• 12: N. Agrawal, V. Prabhakaran, T. Wobber, J. D. Davis, M. Manasse, and R. Panigrahy. Design tradeoffs for ssd performance. June 2008.
• 13: A. Birrell, M. Isard, C. Thacker, and T. Wobber. A design for high-performance flash disks. Technical Report MSR-TR-2005-176, Microsoft Research, December 2005.
• 14: Li-Pin Chang, On efficient wear leveling for large-scale flash-memory storage systems, Proceedings of the 2007 ACM symposium on Applied computing, March 11-15, 2007, Seoul, Korea [doi>10.1145/1244002.1244248]
• 15: Hung-chih Yang , Ali Dasdan , Ruey-Lung Hsiao , D. Stott Parker, Map-reduce-merge: simplified relational data processing on large clusters, Proceedings of the 2007 ACM SIGMOD international conference on Management of data, June 11-14, 2007, Beijing, China [doi>10.1145/1247480.1247602]
• 16: C. T. Chu, S. K. Kim, Y. A. Lin, Y. Yu, G. R. Bradski, A. Y. Ng, and K. Olukotun. Map-reduce for machine learning on multicore. In B. Schölkopf, J. C. Platt, and T. Hoffman, editors, NIPS, pages 281--288. MIT Press, 2006.
• 17: Jeffrey Dean , Sanjay Ghemawat, MapReduce: simplified data processing on large clusters, Proceedings of the 6th conference on Symposium on Opearting Systems Design & Implementation, p.10-10, December 06-08, 2004, San Francisco, CA
• 18: David DeWitt , Jim Gray, Parallel database systems: the future of high performance database systems, Communications of the ACM, v.35 n.6, p.85-98, June 1992 [doi>10.1145/129888.129894]
• 19: D. J. Dewitt , S. Ghandeharizadeh , D. A. Schneider , A. Bricker , H. -I. Hsiao , R. Rasmussen, The Gamma Database Machine Project, IEEE Transactions on Knowledge and Data Engineering, v.2 n.1, p.44-62, March 1990 [doi>10.1109/69.50905]
• 20: D. Economou, S. Rivoire, C. Kozyrakis, and P. Ranganathan. Fullsystem power analysis and modeling for server environments. June 2006.
• 21: Xiaobo Fan , Wolf-Dietrich Weber , Luiz Andre Barroso, Power provisioning for a warehouse-sized computer, Proceedings of the 34th annual international symposium on Computer architecture, June 09-13, 2007, San Diego, California, USA
• 22: Y. Fukuzumi, Y. Matsuoka, M. Kito, M. Kido, M. Sato, H. Tanaka, Y. Nagata, Y. Iwata, H. Aochi, and A. Nitayama. Optimal integration and characteristics of vertical array devices for ultra-high density, bitcost scalable flash memory. Electron Devices Meeting, 2007. IEDM 2007. IEEE International, pages 449--452, 10-12 Dec. 2007.
• 23: G. Ganger, B. Worthington, and Y. Patt. Disksim. http://www.pdl.cmu.edu/DiskSim/.
• 24: Sanjay Ghemawat , Howard Gobioff , Shun-Tak Leung, The Google file system, ACM SIGOPS Operating Systems Review, v.37 n.5, December 2003
• 25: Goetz Graefe, Encapsulation of parallelism in the Volcano query processing system, ACM SIGMOD Record, v.19 n.2, p.102-111, Jun. 1990
• 26: T. R. Halfhill. Intel's tiny atom. Microprocssor Report, April 2008.
• 27: A. Huffman. Onfi: Leading the way to higher performance. http://www.onfi.org/docs/ComputexDRAMeXchange.pdf.
• 28: Intel. Increasing data center density while driving down power and cooling costs, June 2006. White paper.
• 29: Intel. Quad-core intel xeon processor 3200 series datasheet, 2007.
• 30: Intel. Intel system controller hub datasheet, 2008. http://download.intel.com/design/chipsets/embedded/datashts/319537.pdf.
• 31: Michael Isard , Mihai Budiu , Yuan Yu , Andrew Birrell , Dennis Fetterly, Dryad: distributed data-parallel programs from sequential building blocks, ACM SIGOPS Operating Systems Review, v.41 n.3, June 2007
• 32: Dawoon Jung , Yoon-Hee Chae , Heeseung Jo , Jin-Soo Kim , Joonwon Lee, A group-based wear-leveling algorithm for large-capacity flash memory storage systems, Proceedings of the 2007 international conference on Compilers, architecture, and synthesis for embedded systems, September 30-October 03, 2007, Salzburg, Austria [doi>10.1145/1289881.1289911]
• 33: S.-M. Jung, J. Jang, W. Cho, H. Cho, J. Jeong, Y. Chang, J. Kim, Y. Rah, Y. Son, J. Park, M.-S. Song, K.-H. Kim, J.-S. Lim, and K. Kim. Three dimensionally stacked nand flash memory technology using stacking single crystal si layers on ild and tanos structure for beyond 30nm node. Electron Devices Meeting, 2006. IEDM '06. International, pages 1--4, 11-13 Dec. 2006.
• 34: D. Kwak, J. Park, K. Kim, Y. Yim, S. Ahn, Y. Park, J. Kim, W. Jeong, J. Kim, M. Park, B. Yoo, S. Song, H. Kim, J. Sim, S. Kwon, B. Hwang, H. kyu Park, S. Kim, Y. Lee, H. Shin, N. Yim, K. Lee, M. Kim, Y. Lee, J. Park, S. Park, J. Jung, and K. Kim. Integration technology of 30nm generation multi-level nand flash for 64gb nand flash memory. VLSI Technology, 2007 IEEE Symposium on, pages 12--13, 12-14 June 2007.
• 35: Andrew W. McNabb , Christopher K. Monson , Kevin D. Seppi, MRPSO: MapReduce particle swarm optimization, Proceedings of the 9th annual conference on Genetic and evolutionary computation, July 07-11, 2007, London, England [doi>10.1145/1276958.1276991]
• 36: Micron. Micron ddr3 sdram mt41j256m8 datasheet rev'd, 2008. http://download.micron.com/pdf/datasheets/dram/ddr3/2Gb DDR3 SDRAM.pdf.
• 37: Y. Park, J. Choi, C. Kang, C. Lee, Y. Shin, B. Choi, J. Kim, S. Jeon, J. Sel, J. Park, K. Choi, T. Yoo, J. Sim, and K. Kim. Highly manufacturable 32gb multi-level nand flash memory with 0.0098 μ2 cell size using tanos(si - oxide - al2o3 - tan) cell technology. Electron Devices Meeting, 2006. IEDM '06. International, pages 1--4, 11-13 Dec. 2006.
• 38: Colby Ranger , Ramanan Raghuraman , Arun Penmetsa , Gary Bradski , Christos Kozyrakis, Evaluating MapReduce for Multi-core and Multiprocessor Systems, Proceedings of the 2007 IEEE 13th International Symposium on High Performance Computer Architecture, p.13-24, February 10-14, 2007 [doi>10.1109/HPCA.2007.346181]
• 39: R. Schuetz, H. Oh, J.-K. Kim, H.-B. Pyeon, S. Przybylski, and P. Gillingham. Hyperlink nand flash architecture for mass storage applications. Non-Volatile Semiconductor Memory Workshop, 2007 22nd IEEE, pages 3--4, 26-30 Aug. 2007.
• 40: A. Shah. PC World, February 2008. http://www.pcworld.com/article/id,142684--page,1/article.html.
• 41: http://www.supermicro.com/products/system/.
• 42: D. Woodhouse. Jffs2: The journalling flash file system, version 2. http://sources.redhat.com/jffs2/.

INDEX TERMS

Primary Classification:
  C. Computer Systems Organization
  C.5 COMPUTER SYSTEM IMPLEMENTATION
      C.5.5 Servers

Additional Classification:
  C. Computer Systems Organization
  C.3 SPECIAL-PURPOSE AND APPLICATION-BASED SYSTEMS
  C.4 PERFORMANCE OF SYSTEMS

General Terms:
Performance

Keywords:
cluster architecture, data centric, flash memory, solid-state storage