Data Loss in RAID-5 Storage Systems with Latent Errors

摘要

Storage systems employ redundancy and recovering schemes to protect against device failures and latent sector errors, and to enhance reliability. The effectiveness of these schemes has been evaluated based on the Mean Time to Data Loss (MTTDL) and the Expected Annual Fraction of Data Loss (EAFDL) metrics. The reliability degradation due to device failures has been assessed in terms of both these metrics, but the adverse effect of latent errors has been assessed only in terms of the MTTDL metric. This article addresses the issue of evaluating the amount of data losses caused by latent errors. It presents a methodology for obtaining MTTDL and EAFDL of RAID-5 systems analytically in the presence of unrecoverable or latent errors. A theoretical model capturing the effect of independent latent errors and device failures is developed, and closed-form expressions are derived for the metrics of interest.