Clump: Improving File System Performance Through Adaptive Optimizations

Recent years have seen a meteoric rise in microprocessor performance. Overall computer system performance, however, has not kept pace with that of its most glamorous component. One key bottleneck in system performance is le system I/O 17]. This paper investigates methods for improving le system performance by reducing or hiding disk latency. These methods monitor le accesses and infer relationships about the accessed les. Based on this continuously updated information, the system makes decisions about how to improve le system performance. Two optimization techniques are examined: reorganizing the disk to colocate related les and prefetching les in anticipation of future requests. An algorithm for selecting les to prefetch based on second-order correlations despite the presence of noise is presented. Cluster-based disk reorganization was unable to achieve any beneet over traditional block and cylinder shuf-ing 21, 24, 23, 1]. Our prefetching systems were able to predict les to prefetch with over 50% accuracy.