Paper Title: Using an Emerging Memory Cache with a High-Speed Client SSD

Paper Abstract: As PCIe bus speeds move from 8 Gb/s (Gen 3) to 16 Gb/s (Gen 4) or 32 Gb/s (Gen 5), NAND-flash based SSDs can no longer keep up, even with high-density devices and large numbers of parallel channels (more than 8). A solution is to add a high-speed cache made from an emerging memory technology such as MRAM or 3D XPoint which offers much faster writes. However, the emerging memories are expensive so the cache has to be fairly small. Issues include the added complexity of the controller, the types of emerging memories that are supported, and the cache algorithm.

Paper Author: Stanley Huang, Director, SSD Product Marketing, Silicon Motion

Author Bio: Stanley Huang is Director SSD Product Marketing at Silicon Motion, where he leads the company’s highly successful SSD product planning, product marketing, and project team. His team has helped make Silicon Motion a leading supplier in the SSD controller industry. He has over 8 years experience in the SSD market and was a board member of the CompactFlash Association. He earned an MSEE at the National Tsinghua University (Taiwan) and has presented at several previous Flash Memory Summits.