En-Stor: Energy-Aware Hybrid Mobile Storage System using Predictive Prefetching and Data Mining Engine

In this paper, we proposed an energy-aware data mining predictive prefetching technique for hybrid storage systems called En-Stor that uses data mining predictive prefetching to save energy. We used as an example of the hybrid storage systems mobile hard disk drives (MHDDs) and solid-state disks (SSDs). As the SSDs are much more energy-efficient than MHDDs, aggressive prefetching for data from MHDDs will enable them to be in the standby mode as much as possible in order to save power. En-Stor differs from existing energy-aware prefetching techniques in two ways. First, En-Stor is implemented in hybrid storage system where MDDS and SSDs are used. Second, it used data mining predictive prefetching techniques to prefetch the data from MDDs to SSDs to increase the standby time of the MDDs, hence reduce the energy consumption. The data mining predictive prefetching techniques will also increase the performance of the system because most of the requested data will be stored in the SSDs, which offer much faster access time than the MDDs. A simulator was created to evaluate the performance of the En-Stor. Our results show that En-Stor reduces the power consumption of the mobile disk drives by at least 85% when compared with non En-Stor system.