Prediction and Grouping

Description

Most systems treat each data request as an independent event. However, such requests in a computer system are driven by programs and user behaviour, and are therefore far from random. We have conducted research into multiple aspects of predicting data access behavior, and the identification and exploitation of such behavior to produce: predictive caches, improved access predictors, and the automated grouping of related data. This work is ongoing, and has also been extended to problems in mobile data management, file migration and power conservation in mobile storage applications.

Status

Despite having resulted in new avenues of research on mobile storage management, work is ongoing on file and data access prediction. Most recently we have developed new access predictors that incorporate machine learning technique to automatically improve prediction quality and reduce the chances of mispredictions.

We are currently investigating the use of machine learning to adaptively place files in non-hierarchical storage systems.

Publications

• Karl Brandt, Darrell D. E. Long, Ahmed Amer, Predicting When Not To Predict, Proceedings of the 12th Annual IEEE/ACM International Symposium on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems (MASCOTS '04), October 2004.
• Purvi Shah, Jehan-François Pâris, Ahmed Amer, Darrell D. E. Long, Identifying Stable File Access Patterns, Proceedings of the Twelfth NASA Goddard/Twenty First IEEE Conference on Mass Storage Systems and Technologies (MSST '04), April 2004.
• Jehan-François Pâris, Ahmed Amer, Darrell D. E. Long, A Stochastic Approach to File Access Prediction, Proceedings of the International Workshop on Storage Network Architecture and Parallel I/O (SNAPI '03), September 2003.
• Ahmed Amer, Alison Luo, Newton Der, Darrell D. E. Long, Alexander Pang, Visualizing Cache Effects on I/O Workload Predictability, Proceedings of the International Performance Conference on Computers and Communication (IPCCC '03), April 2003.
• Gary Whittle, Jehan-François Pâris, Ahmed Amer, Darrell D. E. Long, Randal Burns, Using Multiple Predictors to Improve the Accuracy of File Access Predictors, Proceedings of the Twentieth IEEE/Eleventh NASA Goddard Conference on Mass Storage Systems and Technologies (MSST '03), April 2003.
• Ahmed Amer, Darrell D. E. Long, Randal Burns, Group-based management of distributed file caches, Proceedings of the 22nd International Conference on Distributed Computing Systems (ICDCS '02), July 2002.
• Ahmed Amer, Darrell D. E. Long, Jehan-François Pâris, Randal Burns, File access prediction with adjustable accuracy, Proceedings of the International Performance Conference on Computers and Communication (IPCCC '02), April 2002.
• Tsozen Yeh, Darrell D. E. Long, Scott A. Brandt, Using program and user information to improve file prediction performance, Proceedings of the International Symposium on Performance Analysis of Systems and Software (ISPASS '01), November 2001, pages Proceedings.
• Tsozen Yeh, Darrell D. E. Long, Scott A. Brandt, Performing file prediction with a program-based successor model, Proceedings of the 9th International Symposium on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems (MASCOTS '01), August 2001, pages 193–202.
• Ahmed Amer, Darrell D. E. Long, Aggregating caches: A mechanism for implicit file prefetching, Proceedings of the 9th International Symposium on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems (MASCOTS '01), August 2001, pages 293–301.
• Tsozen Yeh, Darrell D. E. Long, Scott A. Brandt, Caching Files with a Program-based Last N Successors, Workshop on Caching, Coherency and Consistency (WC3 '01), June 2001.
• Tsozen Yeh, Darrell D. E. Long, Scott A. Brandt, Conserving Battery Energy through Making Fewer Incorrect File Predictions, IEEE Workshop on Power Management for Real-Time and Embedded Systems at the IEEE Real-Time Technology and Applications Symposium, May 2001.
• Ahmed Amer, Darrell D. E. Long, Noah: Low-cost file access prediction through pairs, Proceedings of the 20th IEEE International Performance, Computing and Communications Conference (IPCCC '01), April 2001, pages 27–33.
• Tsozen Yeh, Darrell D. E. Long, Scott A. Brandt, Increasing predictive accuracy through limited prefetching, Proceedings of Communications Networks and Distributed Systems Modeling and Simulation (CNDS 2002), January 2001, pages 131–138.
• Thomas Kroeger, Darrell D. E. Long, Design and implementation of a predictive file prefetching algorithm, Proceedings of the 2001 USENIX Annual Technical Conference, January 2001, pages 105–118.
• Carlos Maltzahn, Kathy Richardson, Dirk Grunwald, Reducing the Disk I/O of Web Proxy Server Caches, USENIX Annual Technical Conference, June 1999.
• Carlos Maltzahn, Kathy Richardson, Dirk Grunwald, James Martin, On Bandwidth Smoothing, 4th International Web Caching Workshop (WCW'99), March 1999.
• Thomas Kroeger, Darrell D. E. Long, The case for efficient file access pattern modeling, Proceedings of the 7th IEEE Workshop on Hot Topics in Operating Systems (HotOS-VII), March 1999, pages 14–19.
• Thomas Kroeger, Darrell D. E. Long, Jeffrey C. Mogul, Exploring the bounds of web latency reduction from caching and prefetching, Proceedings of the USENIX Symposium on Internet Technologies and Systems (USITS '97), December 1997, pages 13–22.
• Thomas Kroeger, Darrell D. E. Long, Predicting file-system actions from prior events, Proceedings of the Winter 1996 USENIX Technical Conference, January 1996, pages 319–328.