File I/O optimization based on statistics
Yu Wang · Jisuanji yingyong yanjiu · 2011
This paper analyzed the following items:the files size distribution and space occupation in the popular operating systems such as Windows and Linux;the block sizes distribution of accesses requests in a network file services environment;the layout of downloading or online playback activities of video clip files on the Internet.It verified the necessity and feasibility of optimizing I/O performance of swap file,small size files and frequently visited files based on the statistics.This paper proposed a new architecture and scheduling algorithm of hybrid external storage system composed of iRAM,solid state disk and zoned-disk.Experimental results show the new architecture can reduce the I/O response time,increase the transmission rate and improve system performance.