The Original View of Reed-Solomon Coding and the Welch-Berlekamp Decoding Algorithm

Abstract

Reed-Solomon codes are a class of maximum distance separable error correcting codes with known fast error correction algorithms. They have been widely used to assure data integrity for stored data on compact discs, DVDs, and in RAID storage systems, for digital communications channels such as DSL internet connections, and for deep space communications on the Voyager mission. The recent explosion of storage needs for "Big Data'' has generated renewed interest in large storage systems with extended error correction capacity. Reed-Solomon codes have been suggested as one potential solution. This dissertation reviews the theory of Reed-Solomon codes from the perspective taken in Reed and Solomon's original paper on them. It then derives the Welch-Berlekamp algorithm for solving certain polynomial equations, and connects this algorithm to the problem of error correction. The discussion is mathematically rigorous, and provides a complete and consistent discussion of the error correction process. Numerous algorithms for encoding, decoding, erasure recovery, error detection, and error correction are provided and their computational cost is analyzed and discussed thus allowing this dissertation to serve as a manual for engineers interested in implementing Reed-Solomon coding.