CAREER: Study of Scalable, High-Performance Distributed Shared Storage Systems

职业：可扩展、高性能分布式共享存储系统的研究

• 批准号: 9984852
• 负责人: Yiming Hu
• 金额: --
• 依托单位: University of Cincinnati Main Campus
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2000
• 资助国家: 美国
• 起止时间: 2000-03-15 至 2005-01-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9984852&HistoricalAwards=false
• 关键词: CAREER; Study; Scalable; Performance; Distributed

It is well known that storage systems are becoming one of the biggest obstacles for many high-performance applications, due to the ever-widening speed gap between CPU and disks. The goal of this research is to overcome the performance limitation of traditional storage systems by developing and studying novel high performance, scalable, distributed storage systems. The research in this project will develop architectures of high-performance distributed storage systems that are scalable with a number of disks, with easy-to-use Application Programming Interfaces and high resource utilization. Moreover, data placing algorithms, which can perform dynamical load balancing and performance optimization, will be studied. To further improve the I/O performance, the project will design distributed I/O cache architectures that can significantly reduce I/O read and write cost and benefit a wide range of workloads. Educational activities will include: developing new graduate-level courses that emphasize specially on the close interaction between computer architectures and operating systems; teaching paper reading, critical analysis and presentation skills for students; and integrating research and education.

众所周知，由于CPU和磁盘之间不断扩大的速度差距，存储系统正成为许多高性能应用的最大障碍之一。 这项研究的目的是通过开发和研究新型的高性能，可扩展的分布式存储系统来克服传统存储系统的性能限制。该项目中的研究将开发具有许多磁盘可扩展的高性能分布式存储系统的体系结构，并具有易于使用的应用程序编程接口和高资源利用。此外，将研究可以执行动态负载平衡和性能优化的数据放置算法。 为了进一步提高I/O性能，该项目将设计分布式的I/O缓存体系结构，这些架构可以大大降低I/O读写成本并受益于广泛的工作量。 教育活动将包括：开发新的研究生级课程，这些课程特别强调计算机架构与操作系统之间的紧密相互作用；教学纸阅读，批判性分析和学生的演讲技巧；以及整合研究和教育。