On the design of SSRS and RS codes for enhancing the integrity of information storage in NAND flash memories

• Published in: IEEE access Vol. 11; p. 1
• Main Authors: Achala, G, Shripathi Acharya, U, Srihari, Pathipati
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.01.2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Bit error rate; channel coding; Codes; Computer architecture; Data processing; Data storage; Decoding; error correction code; Flash memories; flash memory; Flash memory (computers); Information storage; Integrity; Memory devices; Microprocessors; NAND flash; RBER; Reed-Solomon codes; SSRS codes; Storage capacity; Symbols; Synthesis; UBER; Voltage
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2023.3294544

The revolution in the field of information processing systems has created a huge demand for reliable and enhanced data storage capabilities. This demand is being met by advances in channel coding algorithms along with upward scaling of the capacities of hardware devices. NAND Flash memory is a type of non-volatile memory. Scaling of the size of flash memories from Single Level Cell (SLC) devices to Multilevel cell (MLC) devices has increased the storage capacity. However, these multi-bit per cell architectures are characterized by significantly higher Raw Bit Error Rate (RBER) values when compared with SLC architectures. The requirement of low Undetected Bit Error Rate (UBER) values has motivated us to synthesize powerful channel codes for enhancing the integrity of information Storage in multi-level NAND Flash Memory devices. This paper describes the synthesis of novel Subfield Subcodes of Reed Solomon Codes (SSRS) and Reed-Solomon (RS) codes which are matched to multi-bit per cell architectures. UBER values have been calculated for each of the synthesized codes described in this paper. This allows the determination of the performance and the improvement in data storage integrity brought by using these codes. We have shown that the synthesized SSRS and RS codes can provide very low UBER even when the corresponding RBER values are appreciable. As RS codes permit the detection and correction of a greater number of errors for a given code length, their performance is superior to that of SSRS codes. This improved performance is obtained at the cost of greater complexity of encoding and decoding processes.