CSR: Small: Heterogeneous Storage Systems with Emerging Technologies for Solving Big Data Problems

CSR：小型：利用新兴技术解决大数据问题的异构存储系统

• 批准号: 1812537
• 负责人: David Du
• 金额: $ 49.91万
• 依托单位: University of Minnesota-Twin Cities
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2018
• 资助国家: 美国
• 起止时间: 2018-08-01 至 2023-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1812537&HistoricalAwards=false
• 关键词: CSR; Small; Heterogeneous; Storage; Systems

The past decade has witnessed tremendous advances in computing, communication and storage technologies. With the unprecedented connectivity provided by Internet, many new data-driven applications have emerged and are being developed. The information technology (IT) infrastructure of this new era calls for low cost and high-performance storage, flexible ways of processing data for decision making and information retrieval, and keeping data for extremely long durations. This project intends to address certain aspects of these challenges. That is, 1) developing heterogeneous storage systems with low cost and high performance based on emerging storage technologies and 2) leveraging active storage devices for supporting decision making and information retrieval. An active storage device is a storage device having some additional, but limited processing power and can process data locally on the device. Active storage devices can be used to effectively filter out the unnecessary data from storage to memory to support decision making and information retrieval. This project will encourage innovative and creative thinking for integrated solutions that combine emerging storage technologies/devices to build storage systems supporting decision making and information retrieval, enhancing the fundamental understanding of large scale data management in big data era, and quickly developing prototypes to demonstrate the capabilities of these designs. This project is also closely collaborated with a number of industrial companies through NSF Industry-University Cooperative Research Center for Research in Intelligent Storage (CRIS). The outcomes of the project can be quickly adopted by storage industry. The expected project outcomes include fostering the advancement of science and technology especially for data-driven type of scientific research, making society more efficient with big data solutions/applications, and creating better ways of managing and migrating huge amount of available data with better performance and lower cost. Through collaboration with storage industry, the project will provide an ideal hands-on learning and development environment to teach both graduate and undergraduate students important system building and experimental skills that are critical for today's IT workforce. The project outcomes will be incorporated into classroom teaching of both course projects and the core courses in computer science and data science programs for both graduate and under-graduate students. The outcomes of the project include managing algorithms for new storage technologies/devices, new storage architectures, different storage systems designs, new data models, information access methods/algorithms and new ways to deliver information and making decisions. The research outcomes will be deposited and available to public in http://cris.cs.umn.edu.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

过去十年见证了计算、通信和存储技术的巨大进步。随着互联网提供前所未有的连接性，许多新的数据驱动应用程序已经出现并正在开发中。新时代的信息技术（IT）基础设施需要低成本、高性能的存储、灵活的决策和信息检索数据处理方式以及极长的数据保存时间。 该项目旨在解决这些挑战的某些方面。即，1）基于新兴存储技术开发低成本、高性能的异构存储系统；2）利用主动存储设备支持决策和信息检索。 主动存储设备是具有一些附加但有限的处理能力并且可以在设备上本地处理数据的存储设备。主动存储设备可以有效地过滤掉存储到内存中不必要的数据，以支持决策和信息检索。该项目将鼓励集成解决方案的创新和创造性思维，结合新兴存储技术/设备来构建支持决策和信息检索的存储系统，增强对大数据时代大规模数据管理的基本理解，并快速开发原型来演示这些设计的能力。 该项目还通过NSF智能存储产学合作研究中心（CRIS）与多家工业公司密切合作。项目成果可快速被存储行业采用。预期的项目成果包括促进科学技术的进步，特别是数据驱动型科学研究的进步，通过大数据解决方案/应用程序使社会更加高效，以及创建更好的方法来管理和迁移大量可用数据，并具有更好的性能和性能。成本更低。 通过与存储行业的合作，该项目将提供一个理想的实践学习和开发环境，向研究生和本科生传授对当今 IT 员工至关重要的重要系统构建和实验技能。该项目成果将纳入课程项目以及研究生和本科生计算机科学和数据科学项目核心课程的课堂教学中。该项目的成果包括新存储技术/设备的管理算法、新存储架构、不同存储系统设计、新数据模型、信息访问方法/算法以及传递信息和做出决策的新方式。 研究成果将在 http://cris.cs.umn.edu 上存放并向公众开放。该奖项反映了 NSF 的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。

项目成果

Sliding Look-Back Window Assisted Data Chunk Rewriting for Improving Deduplication Restore Performance

滑动回溯窗口辅助数据块重写以提高重复数据删除恢复性能

• 发表时间: 2019-01
• 期刊: Usenix conference on File and Storage Technologies
• 作者: Zhichao Cao; Shiyong Liu
• 通讯作者: Shiyong Liu

IMG-DNA: Approximate DNA Storage for Images

IMG-DNA：图像的近似 DNA 存储

• 发表时间: 2021-06
• 期刊: International Systems and Storage Conference (SYSTOR’21
• 作者: Bingzhe Li; Li Ou
• 通讯作者: Li Ou

TransKV: A Networking Support for Transaction Processing in Distributed Key-value Stores

TransKV：分布式键值存储中事务处理的网络支持

• DOI: 10.1109/bigdataservice52369.2021.00011
• 发表时间: 2021-08
• 期刊: IEEE International Conference on Big Data Computing Service and Applications (BigDataService
• 作者: Eldakiky, Hebatalla;Du, David Hung
• 通讯作者: Du, David Hung

ChewAnalyzer: Workload-Aware Data Management Across Differential Storage Pools

ChewAnalyzer：跨差异存储池的工作负载感知数据管理

• 发表时间: 2018-01
• 期刊: IEEE MASCOTS
• 作者: Xiongzi Ge; Xuchao Xie
• 通讯作者: Xuchao Xie

Can We Store the Whole World Data in DNA-Storage?

我们可以将全世界的数据存储在 DNA 存储器中吗？

• 发表时间: 2020-07
• 期刊: Usenix HotStorage 2020
• 作者: Bingzhe Li; Nae
• 通讯作者: Nae