Abstract:Dynamic random access memory (DRAM) is being upgraded iteratively, and as a result, its transmission rate and bandwidth are rising quickly. Simultaneously, as the DRAM process has advanced, the storage cell size has decreased and cell integration has improved within each device, leading to a significant boost in storage capacity and density. DRAM has been widely utilized as a crucial storage component in personal computers, mobile devices, servers, and data centers because of these benefits. However, data reliability is greatly hampered by DRAM’s vulnerability to single-bit, row, and column errors, which result in data loss and corruption as well as the possibility of system crashes and downtime. Error correction codes (ECC) are used by DRAM to protect data and increase reliability, but because large capacity DRAM is more prone to multi-bit errors of cross-chip. Traditional error correction strategies are unable to keep up with the demand for multi-bit errors of cross-chip. Therefore, a crucial problem that needs to be solved is the design of an ECC strategy with robust error correction capabilities. A high order ECC scheme with stronger error correcting capability is developed at a higher firmware layer without changing the hardware architecture to address reliability issues brought by DRAM multi-bit errors of cross-chip. The higher order ECC technique is then used to gain a stronger error correction capability while minimizing the latency overhead when an uncorrectable error is discovered by rank-level ECC (RECC). The error correction performance of the proposed high order ECC algorithm is evaluated and verified using simulation experiments in terms of both error correction capability and encoding/decoding latency. Simulation results show that compared with existing ECC schemes, the proposed high order ECC scheme for DRAM reduces latency by 69% and storage overhead by 5.56%. The proposed high order ECC method has significant research implications and is useful in preventing data loss and enhancing DRAM reliability.

SCORE: A Novel Scheme to Efficiently Cache Overlong ECCs in NAND Flash Memory.

MScache: A buffer management scheme based on page-level address mapping for NAND-flash SSD

Exploit Asymmetric Error Rates of Cell States to Improve the Performance of Flash Memory Storage Systems

An elastic error correction code technique for NAND flash-based consumer electronic devices

A Machine Learning-empowered Cache Management Scheme for High-Performance SSDs

Asymmetric Error Rates of Cell States Exploration for Performance Improvement on Flash Memory Based Storage Systems

Data Storage Time Sensitive ECC Schemes for MLC NAND Flash Memories

Lifetime adaptive ECC in NAND flash page management.

CDF-LDPC: A New Error Correction Method for SSD to Improve the Read Performance.

DEPS: Exploiting a Dynamic Error Prechecking Scheme to Improve the Read Performance of SSD.

Loss is Gain

LDPC-in-SSD: making advanced error correction codes work effectively in solid state drives.

DECC: Differential ECC for Read Performance Optimization on High-Density NAND Flash Memory

CooECC: A Cooperative Error Correction Scheme to Reduce LDPC Decoding Latency in NAND Flash

NCache: A Machine-Learning Cache Management Scheme for Computational SSDs

High-Performance and Flexible Design Scheme with ECC Protection in the Cache

Product Code Schemes for Error Correction in MLC NAND Flash Memories

Flexible product code-based ECC schemes for MLC NAND Flash memories.

Realizing Unequal Error Correction for Nand Flash Memory at Minimal Read Latency Overhead

Improving DRAM Reliability Using a High Order Error Correction Code

Overclocking Nand Flash Memory I/O Link in Ldpc-Based Ssds