This three-year, $324,000 Project Grant from the National Science Foundation's Computer and Information Science and Engineering program will fund the development of a new Multi-Layer Error Detection architecture to improve the reliability of petabyte-scale file transfers. Researchers from Arizona State University and Boston University will design and implement MLED, leveraging features of the Fabric testbed that are unavailable on other networks. MLED aims to eliminate detectable errors during file transfers and provably reduce the risk of undetected errors through a layered approach parameterized by policy requirements. The researchers will develop theoretical models of MLED, design a software-defined implementation for Fabric, and evaluate MLED's performance to understand design tradeoffs when transferring very large files. This work supports the NSF's priorities of advancing networked computing systems and managing massive datasets, with the potential to improve scientific collaborations that generate terabytes of data daily.
Generated 1/7/24, 12:12 AM