An Improved Adaptive Cache Prefetch Algorithm

Abstract

I/O scheduling algorithm has a crucial impact on the performance of disk arrays. in the disk array system, how to improve I/O transfer rate has been a critical issue. This paper presents an improved adaptive cache prefetch algorithm to solve low prefetch performance problem of multi-user access, caused by the memory address dispersion. Using sliding window algorithm based on the queue to get a stable hit rate, and using capacity dual-cache structure to shorten vacated cache time. in case of multi-user request, the improved adaptive cache prefetch algorithm can get a stable hit rate and reduce the parallel access cache time to achieve high performance prefetch, and fast adaptive response. the experimental analysis shows that the I/O data transmission rate for improved adaptive algorithm system is higher than the basic adaptive prefetch algorithm by 60%, while the average response time is reduced by 60%. to solve the low performance of adaptive algorithm for multi-user prefetch, improved adaptive cache prefetch algorithm is ideal for multi-user environment's I/O performance optimization.