SBIR Phase I: SSD In-Situ Processing

SBIR 第一阶段：SSD 原位处理

• 批准号: 1548968
• 负责人: Vladimir Alves
• 金额: $ 15万
• 依托单位: NGD Systems, Inc.
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2016
• 资助国家: 美国
• 起止时间: 2016-01-01 至 2016-06-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1548968&HistoricalAwards=false
• 关键词: SBIR; Phase; SSD; Situ; Processing

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project will fundamentally change what a storage device can do, and give storage a third capability that was not addressed by existing storage technology - the ability to actually process the data under the explicit control of the user. For the computation to take place, only the computational request and the resulting data need to transfer over the storage interface, reducing the interface traffic and the required power. The advent of Big Data and the increasing use of Hyperscale Server technology have resulted in the creation of an additional storage tier that is different from traditional enterprise storage. This new tier requires significantly larger capacity with lower cost and lower operating power, and yet must still exhibit enterprise reliability. This combination of characteristics cannot be serviced by existing technologies, and execution with large data sets typical of Big Data results in inefficient solutions. The information being stored represents the large, unstructured data mined by today's companies for key information and trends that help dictate corporate direction, advertising, and monetization. Future applications include real-time distributed video and image processing, genome sequencing and mining of any unstructured Big Data.This Small Business Innovation Research (SBIR) Phase I project explores the Big Data paradigm shift where processing capability is pushed as close to the data as possible. The in-situ processing technology pushes this concept to the absolute limit, by putting the computational capability directly into the storage itself and eliminating the need to move the data to main memory before processing. The technology innovation begins with a solid foundation of an enterprise SSD tailored for the needs of modern Data Centers. Key technology that will be added to support these capabilities include hardware-assisted quality of service control, low-cost TLC/3D-TLC NAND Flash enablement through the use of advanced ECC, and a proprietary elastic Flash Translation Layer to support extremely large capacity drives. The final element added to this foundation will be the ability to perform computations directly on the data with the addition of specialized in-situ processing aided by hardware accelerators. To make this disruptive solution as non-invasive as possible, a level of system software is needed to make adoption as seamless as possible.

此小型企业创新研究（SBIR）第一阶段项目的广泛影响/商业潜力将从根本上改变存储设备的功能，并为存储提供现有存储技术无法解决的第三种功能-在用户的明确控制下实际处理数据的能力。 为了进行计算，只有计算请求和结果数据需要通过存储接口传输，从而减少了接口流量和所需的功率。大数据的出现和超大规模服务器技术的日益广泛使用，导致创建了一个不同于传统企业存储的额外存储层。这一新层需要更大的容量、更低的成本和更低的运行功耗，但仍必须表现出企业级的可靠性。现有技术无法满足这些特征的组合，并且使用大数据典型的大型数据集执行会导致低效的解决方案。 存储的信息代表了当今公司挖掘的大型非结构化数据，这些数据用于帮助确定公司方向、广告和货币化的关键信息和趋势。未来的应用包括实时分布式视频和图像处理、基因组测序和任何非结构化大数据的挖掘。这个小企业创新研究（SBIR）第一阶段项目探索了大数据范式的转变，其中处理能力尽可能接近数据。原位处理技术将这一概念推向了绝对的极限，将计算能力直接放入存储本身，并消除了在处理之前将数据移动到主存储器的需要。技术创新始于为满足现代数据中心需求而量身定制的企业SSD奠定坚实的基础。为支持这些功能而添加的关键技术包括硬件辅助服务质量控制、通过使用高级ECC实现的低成本TLC/3D-TLC NAND闪存支持，以及支持超大容量驱动器的专有弹性闪存转换层。添加到这个基础上的最后一个元素将是直接对数据执行计算的能力，并在硬件加速器的帮助下添加专门的原位处理。 为了使这种破坏性的解决方案尽可能地非侵入性，需要一定程度的系统软件来使采用尽可能无缝。