A nine year study of file system and storage benchmarking
ACM Transactions on Storage (TOS)
Towards an I/O tracing framework taxonomy
PDSW '07 Proceedings of the 2nd international workshop on Petascale data storage: held in conjunction with Supercomputing '07
Preserving time in large-scale communication traces
Proceedings of the 22nd annual international conference on Supercomputing
Towards realistic file-system benchmarks with CodeMRI
ACM SIGMETRICS Performance Evaluation Review
Generating realistic impressions for file-system benchmarking
FAST '09 Proccedings of the 7th conference on File and storage technologies
WorkOut: I/O workload outsourcing for boosting RAID reconstruction performance
FAST '09 Proccedings of the 7th conference on File and storage technologies
ScalaTrace: Scalable compression and replay of communication traces for high-performance computing
Journal of Parallel and Distributed Computing
Pattern Matching and I/O Replay for POSIX I/O in Parallel Programs
Euro-Par '09 Proceedings of the 15th International Euro-Par Conference on Parallel Processing
Generating realistic impressions for file-system benchmarking
ACM Transactions on Storage (TOS)
CloudCmp: shopping for a cloud made easy
HotCloud'10 Proceedings of the 2nd USENIX conference on Hot topics in cloud computing
CloudProphet: towards application performance prediction in cloud
Proceedings of the ACM SIGCOMM 2011 conference
Extracting flexible, replayable models from large block traces
FAST'12 Proceedings of the 10th USENIX conference on File and Storage Technologies
Proceedings of the Twenty-Fourth ACM Symposium on Operating Systems Principles
ACM SIGOPS 24th Symposium on Operating Systems Principles
ROOT: replaying multithreaded traces with resource-oriented ordering
Proceedings of the Twenty-Fourth ACM Symposium on Operating Systems Principles
Flash caching on the storage client
USENIX ATC'13 Proceedings of the 2013 USENIX conference on Annual Technical Conference
Automatic identification of application I/O signatures from noisy server-side traces
FAST'14 Proceedings of the 12th USENIX conference on File and Storage Technologies
| Hi-index | 0.00 |
//TRACE is a new approach for extracting and replaying traces of parallel applications to recreate their I/O behavior. Its tracing engine automatically discovers internode data dependencies and inter-I/O compute times for each node (process) in an application. This information is reflected in per-node annotated I/O traces. Such annotation allows a parallel replayer to closely mimic the behavior of a traced application across a variety of storage systems. When compared to other replay mechanisms, //TRACE offers significant gains in replay accuracy. Overall, the average replay error for the parallel applications evaluated in this paper is below 6%.