Improving File System Performance via Predictive Caching

Despite impressive advances in file system throughput resulting from technologies such as high-bandwidth networks and disk arrays, file system latency has improved little and in several cases has become worse. Consequently, file system I/O remains one of the major bottlenecks to operating system performance [Ous90]. This paper investigates an automated predictive approach that improves file cache performance, utilizes increased device bandwidth, and reduces file system latency. Automatic Prefetching uses past file accesses to predict future file system requests. The objective is to effectively masking access latencies. In short, we want to transform the file cache from a repository of recently accessed data into a staging area for data about to be accessed. We have designed and implemented a system to measure the performance benefits of automatic prefetching. Our current results show that prefetching results in over 60% improvement in file system performance compared to LRU, especially..