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

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.

[1]  Donovan A. Schneider,et al.  The Gamma Database Machine Project , 1990, IEEE Trans. Knowl. Data Eng..

[2]  Goetz Graefe,et al.  Encapsulation of parallelism in the Volcano query processing system , 1990, SIGMOD '90.

[3]  David J. DeWitt,et al.  Parallel database systems: the future of high performance database systems , 1992, CACM.

[4]  GhemawatSanjay,et al.  The Google file system , 2003 .

[5]  Sanjay Ghemawat,et al.  MapReduce: Simplified Data Processing on Large Clusters , 2004, OSDI.

[6]  Kunle Olukotun,et al.  Map-Reduce for Machine Learning on Multicore , 2006, NIPS.

[7]  Kinam Kim,et al.  Three Dimensionally Stacked NAND Flash Memory Technology Using Stacking Single Crystal Si Layers on ILD and TANOS Structure for Beyond 30nm Node , 2006, 2006 International Electron Devices Meeting.

[8]  Kinam Kim,et al.  Highly Manufacturable 32Gb Multi -- Level NAND Flash Memory with 0.0098 μm2 Cell Size using TANOS(Si - Oxide - Al2O3 - TaN) Cell Technology , 2006, 2006 International Electron Devices Meeting.

[9]  Christos Kozyrakis,et al.  Full-System Power Analysis and Modeling for Server Environments , 2006 .

[10]  Kevin D. Seppi,et al.  MRPSO: MapReduce particle swarm optimization , 2007, GECCO '07.

[11]  Y. Iwata,et al.  Optimal Integration and Characteristics of Vertical Array Devices for Ultra-High Density, Bit-Cost Scalable Flash Memory , 2007, 2007 IEEE International Electron Devices Meeting.

[12]  Douglas Stott Parker,et al.  Map-reduce-merge: simplified relational data processing on large clusters , 2007, SIGMOD '07.

[13]  Christoforos E. Kozyrakis,et al.  Evaluating MapReduce for Multi-core and Multiprocessor Systems , 2007, 2007 IEEE 13th International Symposium on High Performance Computer Architecture.

[14]  Jin-Ki Kim,et al.  HyperLink NAND Flash Architecture for Mass Storage Applications , 2007, 2007 22nd IEEE Non-Volatile Semiconductor Memory Workshop.

[15]  Wolf-Dietrich Weber,et al.  Power provisioning for a warehouse-sized computer , 2007, ISCA '07.

[16]  Heeseung Jo,et al.  A group-based wear-leveling algorithm for large-capacity flash memory storage systems , 2007, CASES '07.

[17]  Michael Isard,et al.  A design for high-performance flash disks , 2007, OPSR.

[18]  Li-Pin Chang,et al.  On efficient wear leveling for large-scale flash-memory storage systems , 2007, SAC '07.

[19]  Jangho Park,et al.  Integration Technology of 30nm Generation Multi-Level NAND Flash for 64Gb NAND Flash Memory , 2007, 2007 IEEE Symposium on VLSI Technology.

[20]  Yuan Yu,et al.  Dryad: distributed data-parallel programs from sequential building blocks , 2007, EuroSys '07.

[21]  Rina Panigrahy,et al.  Design Tradeoffs for SSD Performance , 2008, USENIX Annual Technical Conference.