Workload-based Configuration of MEMS-Based Storage Devices for Mobile Systems

Because of its small form factor, high capacity, and expected low cost, MEMS-based storage is a suitable storage technol-ogy for mobile systems. However, flash memory may out-perform MEMS-based storage in terms of performance, and energy-efficiency. The problem is that MEMS-based stor-age devices have a large number (i.e., thousands) of heads, and to deliver peak performance, all heads must be deployed simultaneously to access each single sector. Since these de-vices are mechanical and thus some housekeeping informa-tion is needed for each head, this results in a huge capacity loss and increases the energy consumption of MEMS-based storage with respect to flash. We solve this problem by proposing new techniques to lay out data in MEMS-based storage devices. Data layouts represent optimizations in a design space spanned by three parameters: the number of active heads, sector parallelism, and sector size. We explore this design space and show that by exploiting knowledge of the expected workload, MEMS-based devices can employ all heads, thus delivering peak performance, while decreasing the energy consumption and compromising only a little on the capacity. Our exploration shows that MEMS-based storage is competitive with flash in most cases, and outperforms flash in a few cases