IMPLEMENTATION AND PERFORMANCE OF INTEGRATED APPLICATION-CONTROLLED FILE CACHING, PREFETCHING, AND DISK SCHEDULING

As the performance gap between disks and microprocessors continues to increase, effective utilization of the file cache becomes increasingly important. Application-controlled file caching and prefetching can ap ply application-specific knowledge to improve file cache management. H owever, supporting application-controlled file caching and prefetching is nontrivial because caching and prefetching need to be integrated c arefully, and the kernel needs to allocate cache blocks among processe s appropriately. This article presents the design, implementation, and performance of a file system that integrates application-controlled c aching, prefetching, and disk scheduling. We use a two-level cache man agement strategy. The kernel uses the LRU-SP (Least-Recently-Used with Swapping and Placeholders) policy to allocate blocks to processes, an d each process integrates application-specific caching and prefetching based on the controlled-aggressive policy, an algorithm previously sh own in a theoretical sense to be nearly optimal. Each process also imp roves its disk access latency by submitting its prefetches in batches so that the requests can be scheduled to optimize disk access performa nce. Our measurements show that this combination of techniques greatly improves the performance of the file system. We measured that the run ning time is reduced by 3% to 49% (average 26%) for single-process wor kloads and by 5% to 76% (average 32%) for multiprocess workloads.