Pattern-Guided File Compression with User-Experience Enhancement for Log-Structured File System on Mobile Devices

Mobile applications exhibit unique file access patterns, often involving random accesses of write-mostly files and readonly files. The high write stress of mobile applications significantly impacts on the lifespan of flash-based mobile storage. To reduce write stress and save space without sacrificing user-perceived latency, this study introduces FPC, file access pattern guided compression. FPC is optimized for the random-writes and fragmented-reads of mobile applications. It features dual-mode compression: Foreground compression handles write-mostly files for write stress reduction, while background compression packs random-reading file blocks for boosted read performance. FPC exploits the out-of-place updating design in F2FS, a log-structured file system for mobile devices, for the best effect of the proposed dual-mode compression. Experimental results showed that FPC reduced the volume of total write traffic and executable file size by 26.1% and 23.7% on average, respectively, and improved the application launching time by up to 14.8%.