Adaptive caching in a distributed file system
Michael W. Condry, Swee Boon Lim · 1996
Effective file system caching reduces local disk accesses and remote file server accesses significantly. Traditional file systems use fixed strategies to control caching. This thesis shows that a file system with adaptive caching achieves better performance than traditional file systems. Our file system implements multiple caching strategies and permits performance tuning through customized caching strategies. It adapts to the computing environment by selecting strategies suitable for the environment. It observes file accesses and uses the observed behaviors to anticipate and predict future behaviors. It adapts to different file access behaviors by modifying caching strategies. It does not depend on the application or the user for caching hints but will utilize hints when provided. Experiments with two large workloads having distinct file access characteristics show that adaptive file caching consistently outperforms non-adaptive caching. Adaptive file caching can reduce runtime by 36.6%, cache misses by 20.6%, and network load by 24.2%. In addition, this work also includes innovations in file system architecture. They include continuations for highly-concurrent asynchronous remote accesses, and zombies for efficient memory reclamation.