Parallel I/O and Storage Research

The I/O demands of many applications have outpaced the capabilities of I/O subsystems, especially for extreme scale systems.  This situation is expected to only get worse as the community moves toward exascale systems.  The ORNL Future Technologies Group has been investigating these extreme-scale I/O challenges.  We study application I/O demands (including how best to determine these demands), how well current and expected I/O solutions meet these demands, and how to address the gap between application demands and achievable performance. 

Recent Publications

MapReduce with Communication Overlap (MaRCO)

LACIO: A New Collective I/O Strategy for Parallel I/O Systems

Probabilistic Communication and I/O Tracing with Deterministic Replay at Scale

For more I/O- and storage-related publications, see the full list of Future Technologies Group publications.


Philip C. Roth

Seyong Lee

Weikuan Yu (joint with Auburn University)

Yong Chen (postdoc with the Future Technologies group in 2010, now at Texas Tech University)


Send any questions regarding our research in this direction to us using our contact page.