HEC: File System Tracing, Replaying, Profiling, and Analysis on HEC Systems

Suny At Stony Brook, Stony Brook NY

Investigators

Abstract

The PIs propose to develop scalable tools and techniques that will work on large clusters to identify I/O performance bottlenecks, visualize them for cluster users for ease of analysis. Thiese techniques will conduct large scale tracing and replaying, collecting vital information useful to analyze the cluster's performance given a specific application. The PIs will use automated and user driven feedback to raise or lower the level of tracing on individual cluster nodes to (1) zoom in/ on hotspots and (2) trade off information accuracy vs. overheads. Extensive investigations will go toward ensuring the accuracy of the information.