A Strategy of Small Files Caching Based on Access Temperature

Large-scale random access to small files can cause some hot data that are accessed much more frequently than others. How to promote hot data access becomes the key to improve the overall performance of the storage system. Considering reading more than writing to small files and the time locality and frequency dynamics of small files access, this paper proposes a caching strategy based on access temperature and introduces a temperature factor model to fulfill time-dependent dynamic management of file access frequency. On this basis, the management of cache memory and cache replacement is used to achieve efficient caching of hot files. Experimental results show that the strategy based on access heat could improve the access performance of massive files.