Check out the new USENIX Web site.

Reducing File System Latency using a Predictive Approach

James Griffioen
griff@dcs.uky.edu

Randy Appleton
randy@dcs.uky.edu

Department of Computer Science
University of Kentucky
Lexington, KY 40506

Abstract

Despite impressive advances in file system throughput resulting from technologies such as high-bandwidth networks and disk arrays, file sys- tem latency has not improved and in many cases has become worse. Consequently, file system I/O remains one of the major bottlenecks to operating system performance[10].

This paper investigates an automated predic- tive approach towards reducing file latency. Automatic Prefetching uses past file accesses to predict future file system requests. The objec- tive is to provide data in advance of the request for the data, effectively masking access laten- cies. We have designed and implement a system to measure the performance benefits of automatic pre- fetching. Our current results, obtained from a trace-driven simulation, show that prefetching results in as much as a 280% improvement over LRU especially for smaller caches. Alternatively, prefetching can reduce cache size by up to 50%.

Download the full text of this paper in ASCII (39,561 bytes) and POSTSCRIPT (106,382 bytes) form.

To Become a USENIX Member, please see our Membership Information.