A Study of Data Classification for Hybrid Storage System Based on Genetic Algorithm

Abstract

The explosive growth in data storage system so that people demand for large capacity and fast access to the increasingly urgent. At present, able to bear this responsibility is the rapidly growing solid state disk (SSD) and the traditional hard disk drive (HDD). The capacity of HDD is large, but its random reading speed may not have greatly improved. The random reading speed of SSD is a hundred of times faster than HDD, but its ectopic updating, limited number of erase, and expensive price had restricted its popularization. In recent years, the development of hybrid storage system consisting of SSD and HDD was rapid. In order to raise Capability/price Ratio of the hybrid storage system, we must classify the storage of data according to the characteristics of SSD and HDD. This paper presented a data classification method based on genetic algorithm, It gave full play to the advantages of the rapid random reading speed of SSD, the large capacity and in situ update of HDD in accessing a large amount of data in hybrid storage system, and it restrained the inferiority of limited number of erase on SSD and slower random reading speed on HDD. Index Terms - Hybrid storage, Data classification, File heat, Genetic algorithm