Data Prefetching Based on Long-Term Periodic Access Patterns
Data prefetching is a technique allowing to retrieve data which will be most likely needed in a near future, before the actual demand. Considerable research was devoted to this technique, however, it is typically based on short-term data access patterns. We propose to predict future accesses based on long-term periodic pattern mining. Human activity in many areas, and thus many real-world business processes appear to have natural periods: they may have day, week and/or month periods. Discovering of such periods in I/O (or higher level) activity logs should allow to build a prefetch predictor, which is aware of data accesses not only in a near future, but in a far future perspective as well, and thus able to make more reasonable prefetch decisions. In this work we investigate the algorithm for mining periodic long-term access patterns, and discuss issues involved in building a prefetch system, which integrates predictor based on discovering these patterns with a prefetch cost model.
Unable to display preview. Download preview PDF.
- 1.Agrawal, R., Srikant, R.: Fast algorithms for mining association rules in large databases. In: Proceedings of the 20th International Conference on Very Large Data Bases, VLDB 1994, pp. 487–499. Morgan Kaufmann Publishers Inc., San Francisco (1994)Google Scholar
- 2.Amer, A., Long, D.D.E., Paris, J.-F., Burns, R.C.: File access prediction with adjustable accuracy. In: 21st IEEE International Proceedings of the Performance, Computing, and Communications Conference, PCC 2002, pp. 131–140. IEEE Computer Society, Washington, DC (2002)Google Scholar
- 3.Cao, P., Felten, E.W., Karlin, A.R., Li, K.: A study of integrated prefetching and caching strategies. In: Proceedings of the 1995 ACM SIGMETRICS Joint International Conference on Measurement and Modeling of Computer Systems, SIGMETRICS 1995/PERFORMANCE 1995, pp. 188–197. ACM, New York (1995)CrossRefGoogle Scholar
- 4.Griffioen, J., Appleton, R.: Reducing file system latency using a predictive approach. In: Proceedings of the USENIX Summer 1994 Technical Conference on USENIX Summer 1994 Technical Conference, USTC 1994, vol. 1, p. 13. USENIX Association, Berkeley (1994)Google Scholar
- 6.Li, Z., Chen, Z., Srinivasan, S.M., Zhou, Y.: C-miner: Mining block correlations in storage systems. In: Proceedings of the 3rd USENIX Conference on File and Storage Technologies, FAST 2004, pp. 173–186. USENIX Association, Berkeley (2004)Google Scholar