Abstract | ||
---|---|---|
Concurrent accesses to the shared storage resources in current HPC machines lead to severe performance degradation caused by I/O contention. In this study, we identify some key challenges to efficiently handling interleaved data accesses, and we propose a system-wide solution to optimize global performance. We implemented and tested several I/O scheduling policies, including prioritizing specific applications by leveraging burst buffers to defer the conflicting accesses from another application and/or directing the requests to different storage servers inside the parallel file system infrastructure. The results show that we mitigate the negative effects of interference and optimize the performance up to 2x depending on the selected I/O policy. |
Year | DOI | Venue |
---|---|---|
2016 | 10.1109/eScience.2016.7870922 | 2016 IEEE 12th International Conference on e-Science (e-Science) |
Keywords | Field | DocType |
I/O Interference,Parallel File Systems,I/O Policies,I/O Staging,Burst Buffers | File system,Computer science,Scheduling (computing),Server,Input/output,Bandwidth (signal processing),Interference (wave propagation),Burst buffer,Distributed computing | Conference |
ISSN | ISBN | Citations |
2325-372X | 978-1-5090-4274-6 | 2 |
PageRank | References | Authors |
0.37 | 17 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Anthony Kougkas | 1 | 9 | 2.83 |
matthieu dorier | 2 | 131 | 13.91 |
Robert Latham | 3 | 134 | 8.57 |
Robert Ross | 4 | 2717 | 173.13 |
Xian-he Sun | 5 | 1987 | 182.64 |