Comparison Of Two Encoding Methods in RAID6 Storage Architecture and Analysis of Improvement of EVENODD Encodings

Abstract

This paper provides a comparative analysis of two widely used error correction codes: RS (Reed-Solomon) and EVENODD. The analysis covers principles, algebraic forms, computational complexities, minimum Hamming distances, decoding processes, and specific application scenarios. RS encoding is known for its versatility, allowing for the flexible addition of parity symbols. However, it comes with a high computational cost due to polynomial evaluations. On the other hand, EVENODD encoding offers simplicity and computational efficiency through mere XOR operations but has limitations on the number of parity symbols it can generate. The paper explores the algebraic representations of both codes, discusses their minimum Hamming distances, and evaluates their decoding processes. Additionally, it analyzes the computational requirements of EVENODD under small write operations and provides a comprehensive comparison of the computational complexities of RS and EVENODD, aiding practitioners in choosing the most suitable error correction scheme for specific applications. A seeming limitation in our building process is because of the number of information disks must fall into a prime number. Nevertheless, if the preferred quantity of data sectors isn't a prime number, we can easily posit the existence of additional disks filled entirely with zeros, and this won't impact the encoding and decoding procedures.