Abstract | ||
---|---|---|
Aiming to solve the performance-degradation problem when multiple computing nodes are in use in mission-critical batch systems (so-called "grid-batch" systems), a new performance-prediction method that focuses on metadata management for file input/output (I/O) control and performance degradation in case of concurrent I/O streams is proposed. To enhance the accuracy of the prediction, this I/O-performance prediction method models metadata management time as a function of number of files and models performance degradation as a probabilistic function of sequential I/O throughput and random I/O throughput. According to an evaluation of the proposed method, the difference between actual and estimated execution time is 14.0%. In contrast, as for the storage/network-based conventional method, the difference is 36.5%. These results demonstrate that the target prediction error, namely, within 20%, was accomplished with the proposed method, which can therefore be considered effective in predicting the performance of grid-batch systems. |
Year | DOI | Venue |
---|---|---|
2011 | 10.1016/j.procs.2011.04.237 | PROCEEDINGS OF THE INTERNATIONAL CONFERENCE ON COMPUTATIONAL SCIENCE (ICCS) |
Keywords | Field | DocType |
grid-batch, I/O performance prediction, metadata management, concurrent I/O streams, distributed file system | Data mining,Computer science,Input/output,Batch processing,Throughput,Probabilistic logic,Mission critical,Metadata management,Performance prediction,Grid | Journal |
Volume | ISSN | Citations |
4 | 1877-0509 | 0 |
PageRank | References | Authors |
0.34 | 7 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Toshihiko Kashiyama | 1 | 10 | 2.01 |
Tomohiro Hanai | 2 | 1 | 0.70 |
Yoshio Suzuki | 3 | 0 | 0.34 |
Ken Naono | 4 | 8 | 5.74 |