DC:Small: Energy-aware Coordinated Caching in Cluster-based Storage Systems

点击翻译按钮获取中文摘要

As the computing capacity increases rapidly in large-scale cluster computing platforms, power management becomes an increasingly important concern. This project focuses on the research of reducing disk and memory power consumption through energy-aware cooperative caching in cluster-based storage systems. The project leverages I/O characteristics of scientific applications and dynamic power management features of disk drives and memory chips to reduce I/O energy consumption. This project involves three components: (1) investigate program context based pattern detection to predict I/O activities in the operating systems, (2) investigate disk energy aware cooperative cache management schemes, and (3) prototype the management schemes and incorporate into cluster-based file systems. This project has broader impact through its contributions to the energy-aware computing, graduate education, and undergraduate education via an existing NSF-REU site award.