Unified Solid-State-Storage Architecture with NAND Flash Memory and ReRAM that Tolerates 32× Higher BER for Big-Data Applications

📋 论文概要

该论文提出了一种统一固态存储（USSS）架构，通过反向镜像（RM）、错误降低合成（ERS）、页面RAID和错误掩蔽（EM）四种技术，使NAND闪存的可接受原始误码率（ABER）提升32倍，或有效延长耐用性4.2倍、数据保留时间34倍，以解决大数据应用中高误码率对固态硬盘可靠性的挑战。

💡 主要创新点

• 提出统一固态存储架构，结合NAND闪存和ReRAM，通过四种技术协同提升错误容忍度。
• 引入反向镜像和错误降低合成技术，有效降低原始误码率对系统性能的影响。
• 采用页面RAID和错误掩蔽策略，进一步增强数据可靠性和纠错能力。

🏷 关键词

📄 原文摘要

Unified solid-state storage (USSS) provides high error tolerance with four techniques: reverse-mirroring (RM), error-reduction synthesis (ERS), page-RAID, and error-masking (EM). The acceptable raw bit-error rate (ABER) of NAND flash memory is enhanced by 32×, or endurance or data-retention time effectively extends by 4.2 or 34×, respectively. ABER is defined to realize BER after ECC below 10-15. Big-data enterprise storage is escalating demand for SSD, because of SSD’s high speed, low power and small form factor. A conventional storage system has four basic levels of hierarchy: server, disk array, SSD, and NAND flash. ECC and redundancy provide reliability. For performance, in addition to reliability, stripe and mirror everything (S.A.M.E.) has been proposed [1] in which RAID-1 (mirroring) is applied to disk arrays. Big-data tends to become copied among different layers. Since controllers at each level are optimized independently of each

👥 作者与机构

Shuhei Tanakamaru1,2, Masafumi Doi1, Ken Takeuchi1

Chuo University, Tokyo, Japan, 2University of Tokyo, Tokyo, Japan