Paper Title: Self-Adaptive NAND Flash DSP

Paper Abstract: Driven by the growing demands of high-capacity, high-performance and low-cost non-volatile storage, NAND flash technology scales aggressively in terms of bit cost. With planar NAND reaching its scaling limits, 3D NAND technology was developed to provide far higher storage density by vertically stacking. On the other hand, with more bits in a NAND cell, QLC is on the way. Despite the impressive growth of bit density, NAND flash memories increasingly subject to worse raw storage reliability. Affected by too many process and usage factors, the NAND flash characterization of the threshold voltage distribution varies in its life time, which makes it a challenging issue to NAND flash controller. Maxio’s self-adaptive NAND DSP technology is proposed to dynamically track the NAND characterization. Besides the static NAND modeling from the big data of NAND analysis, this DSP algorithm on-the-fly learns the NAND threshold voltage distribution from NAND reads and dynamically adjusts the NAND characterization. With more accurate NAND threshold voltage distribution, combined with strong LDPC codes, Maxio’s self-adaptive NAND DSP technology can significantly improve NAND memory reliability, extend NAND life time and shorten read response time.

Paper Author: Wei Xu, Vice President, Maxio Technology

Author Bio: Dr. Wei Xu serves as Vice President of R&D for Maxio Technology. He is responsible for leading the engineering group that develops the innovative solid-state disk (SSD) controller products. Before joining Maxio, he was a Design Manager at Marvell Semiconductor. Dr. Xu received a Bachelor of Science degree and a Master of Science degree from Fudan University in Shanghai, and Ph.D. in Electric Engineering from Rensselaer Polytechnic Institute (RPI) in US.