Research Initiation Award: Designing Hierarchical Storage Systems for Large Capacity and Low Latency

该项目将研究将新的存储技术（如闪存RAM和螺旋扫描磁带）集成到存储层次结构中，以利用每种技术的优势，同时掩盖其弱点。磁盘阵列的突破性工作已经完成；在这个项目中，重点将放在层次结构的其他领域，如固态存储器和磁带。例如，本研究将探讨如何增加固态存储器的有效容量以降低其较高的成本。这项研究将分三个阶段进行：(1)测量新的存储设备并跟踪当前的I/O工作负载；(2)设计和模拟新的层次结构和算法；(3)实现了分层存储系统的原型，并对其在实际使用中的运行情况进行了测量。

This project will investigate the integration of new storage technologies, such as flash RAM and helical scan tapes, into the storage hierarchy to take advantage of each technology's strengths while masking its weaknesses. Ground-breaking work in disk arrays has already been accomplished; in this project, the emphasis will be on other areas of the hierarchy, such as solid-state memories and magnetic tapes. For example, this research will investigate how to increase the effective capacity of solid-state memories to mitigate their higher cost. This research will take place in three stages: (1) measure new storage devices and trace current I/O workloads; (2) design and simulate new hierarchical architectures and algorithms; (3) implement a prototype hierarchical storage system and measure its operation under real use.