This National Science Foundation (NSF) Computer and Information Science and Engineering (CISE) Federal Grant Award ID 2521613, valued at $133,980.00, supports the University of Delaware's research project titled "EAGER: EXPLORING AUTOMATIC OPTIMIZATION OF MULTI-TIERED HPC STORAGE SYSTEMS VIA PRACTICAL REINFORCEMENT LEARNING." The project aims to develop automated middleware for high-performance computing (HPC) storage systems to help scientists achieve optimal data access performance without manual effort. The researchers plan to leverage machine learning techniques, specifically reinforcement learning, to intelligently and adaptively schedule and coordinate I/O requests, with a focus on two main areas: 1) reinforcement learning-based data placement for high storage utilization, and 2) reinforcement learning-based I/O coordination for shared storage. This research is expected to produce software infrastructure that can work with existing storage components and minimize impacts on both scientific applications and HPC systems. The award period runs from October 1, 2024, to June 30, 2025.
Mod # | Description | Reason For Modification | Federal Obligation (Click to sort descending) | Date (Click to sort ascending) |
|---|---|---|---|---|
| Not listed | $134.0k | 3/4/25 |