Using Markov models to estimate the reliability of RAID architectures

RAID stands for Redundant Array of Independent Disks. It is a storage technology that combines multiple disk drive components into a logical unit. RAID is now used as an umbrella term for computer data storage schemes that can divide and replicate data among multiple physical drives. Reliability and availability are big concerns for any data storage system, including RAID architectures. In this paper, an alternative method estimating the reliability of various RAID architectures is presented, using the Markov models approach. The reliability analysis of various RAID architectures is performed using CARE-Computer Aided Reliability Engineering, an industrial software tool. Estimating the reliability of RAID architectures and combining them with the reliabilities of other elements of the system, it allows estimating the overall reliability of the system. Evaluating reliability and performance of RAID enable enterprises to identify the appropriate disk configurations, as part of their overall IT strategies.