SHF: Small: ASF: An Adaptive Scaling Framework for High Scalability of XOR-Based RAID Systems

SHF：小型：ASF：基于 XOR 的 RAID 系统的高可扩展性的自适应扩展框架

• 批准号: 1702474
• 负责人: Xubin He
• 金额: $ 16.24万
• 依托单位: Temple University
• 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: 2016-08-15 至 2019-01-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1702474&HistoricalAwards=false
• 关键词: SHF; Small; ASF; Adaptive; Scaling

Explosive growth in data volume, heterogeneity, and complexity imposes unprecedented challenges in data analysis and organization in data centers. RAID particularly XOR-based RAID plays an important role to provide both reliability and high performance storage services for these data centers. However, they suffer from problems on the scalability issue due to multi-folded factors, including: heterogeneous RAID layouts and various erasure codes, high overhead of existing scaling process to significantly downgrade the storage performance, and lack of bidirectional scaling support. The objective of this project is to address the scalability challenge for storage systems in large data centers. This project designs novel techniques to exploit XOR-based parity codes to achieve efficient scaling, develops a series of scalable XOR-based erasure codes to bridge the relations among heterogeneous RAID forms for interoperability, and integrates various erasure codes in a framework to provide a unified user interface for RAID scaling. The research will facilitate the storage systems to effectively tackle the scalability challenge in large data centers. It cohesively investigates how to utilize and integrate proper RAID scaling technologies to improve the scalability of storage systems. This project will yield fundamental contributions to research on efficient big data management in high-performance data centers, and greatly contribute to enabling high-performance computing systems to stay on track of their historic scaling, and hence benefit numerous real-life data-intensive applications such as biology, chemistry, earth science, health care, and so on.

数据量、异构性和复杂性的爆炸性增长给数据中心的数据分析和组织带来了前所未有的挑战。 RAID（尤其是基于 XOR 的 RAID）在为这些数据中心提供可靠性和高性能存储服务方面发挥着重要作用。然而，由于多种因素，它们在可扩展性问题上遇到了问题，包括：异构 RAID 布局和各种纠删码、现有扩展过程的高开销导致存储性能显着降低，以及缺乏双向扩展支持。该项目的目标是解决大型数据中心存储系统的可扩展性挑战。该项目设计了新颖的技术来利用基于 XOR 的奇偶校验码来实现高效扩展，开发了一系列可扩展的基于 XOR 的纠删码来桥接异构 RAID 形式之间的关系以实现互操作性，并将各种纠删码集成在一个框架中以提供统一的用于 RAID 扩展的用户界面。该研究将有助于存储系统有效应对大型数据中心的可扩展性挑战。它集中研究了如何利用和集成适当的 RAID 扩展技术来提高存储系统的可扩展性。该项目将为高性能数据中心高效大数据管理的研究做出基础性贡献，并为高性能计算系统保持历史性扩展做出巨大贡献，从而使众多现实生活中的数据密集型应用受益例如生物学、化学、地球科学、医疗保健等。

项目成果

Chameleon: An Adaptive Wear Balancer for Flash Clusters

Chameleon：闪存集群的自适应磨损平衡器

• 发表时间: 2018-05
• 期刊: the 32nd IEEE International Parallel and Distributed Processing Symposium (IPDPS
• 作者: Zhao, N;Anwar, A;Cheng, Y;Salman, M;Li, D;Wan, J;Xie, C;He, X;Wang, F;Butt, A
• 通讯作者: Butt, A

RAFI: Risk-Aware Failure Identification to Improve the RAS in Erasure-coded Data Centers

RAFI：风险意识故障识别以改进纠删码数据中心的 RAS

• 发表时间: 2018-07
• 期刊: The USENIX Annual Technical Conference (ATC
• 作者: Fang, J;Wan, S;He, X
• 通讯作者: He, X

SELF: A High Performance and Bandwidth Efficient Approach to Exploiting Die-Stacked DRAM as Part of Memory

SELF：一种利用裸片堆叠 DRAM 作为内存一部分的高性能和带宽高效方法

• DOI: 10.1109/mascots.2017.23
• 发表时间: 2017-09
• 期刊: and Simulation of Computer and Telecommunication Systems (MASCOTS
• 作者: Guo, Yuhua;Liu, Qing;Xiao, Weijun;Huang, Ping;Podhorszki, Norbert;Klasky, Scott;He, Xubin
• 通讯作者: He, Xubin

DREAM: Data Representation Aware of Damage to Extend the Lifetime of MLC NAND Flash Memory

DREAM：感知损坏的数据表示以延长 MLC NAND 闪存的使用寿命

• 发表时间: 2018-07
• 期刊: the 10th USENIX Workshop on Hot Topics in Storage and File Systems (HotStorage '18
• 作者: Ye, T;Wan, S;He, X;Xiao, W;Xie, C
• 通讯作者: Xie, C

A Set-aware Key-Value Store on Shingled Magnetic Recording Drives with Dynamic Band

具有动态频带的叠瓦式磁记录驱动器上的设置感知键值存储

• 发表时间: 2018-05
• 期刊: the 32nd IEEE International Parallel and Distributed Processing Symposium (IPDPS
• 作者: Yao, T;Tan, Z;Wan, J;Huang, P;Zhang, Y;Xie, C;He, X
• 通讯作者: He, X