Fast Configuration of MEMS-Based Storage Devices for Streaming Applications

摘要

An exciting class of storage devices is emerging: the class of Micro-Electro-Mechanical storage Systems (MEMS). Properties of MEMS-based storage devices include high density, small form factor, and low power. The use of this type of devices in mobile infotainment systems, such as video cameras is not at all obvious. We must explore their configuration and assess their benefit with respect to existing devices, such as Flash. In this paper, we study the configuration of the data layout of MEMS-based storage devices for predominately streaming applications. The configuration targets: the timing performance, energy consumption, and the capacity. We show that formatting the data layout based on knowledge of the expected streaming workload results in devices that consume at least 22% less energy than Flash and perform comparably. We present a fast and effective configuration method to format the data layout. This method exploits the bimodal distribution of the request size in streaming applications. Using the fast method, we present a design rule for streaming environments: reducing the amount of prefetching of streaming data allows to reach configurations with small trade-offs between the design targets.