Skip to Main Content
Most operating systems manage buffer caches for buffering I/O blocks, because I/O processing is slower than CPU processing. Application programs request I/O processing from files. In order to improve the performance of I/O processing, a buffer cache should be managed with regard to both blocks and files. This paper proposes an I/O buffer cache mechanism based on the frequency of file usage. This mechanism calculates the importance of each file. Then, blocks of important files are stored in a protected space. The blocks stored in the protected space are given priority for caching. We evaluated the proposed mechanism by kernel make processing. The results show that the proposed mechanism improves the processing time by 18 s (5.7%) as compared to the LRU algorithm.