Exploring and evaluating various data structures within RAID 6 architecture: An application-oriented study

Abstract

With the development of modern technology, people have more and more chance to use large storage systems. It is important to keep the reliability of the modern storage system. There are different architectures to achieve the goal. Redundant Array of Independent Disks (RAID) architectures are one efficient way to recover the storage system from disk failures. From different RAID data structures, RAID 6 refers to use two additional parity disks to allow the users to recover from up to two disk failures. However, there are different ways to perform the RAID 6. For example, the EVENODD code uses an exclusive OR (XOR) operation to calculate parity. It has low storage requirements and simple computation. The Row Diagonal Parity (RDP) code is an upgraded version of the EVENODD code. Effectively reduce the computational consumption of parity-check encoding. On the other hand, the Reed-Solomon code has an efficient recovery algorithm and a quantitative calculation process. Plus, there are other implementation methods with their advantage and limitations for the RAID 6 architecture. To assist the application of RAID6, this paper aims to analyze, implement, and apply different RAID6 structures. The methodology of the paper is the exclusive literature review of published paper in the field in recent 10 to 20 years.