SPA: On-Line Availability Upgrades for Parity-based RAIDs through Supplementary Parity Augmentations

In this paper, we propose a simple but powerful on-line availability upgrade mechanism, Supplementary Parity Augmentations (SPA), to address the availability issue for parity-based RAID systems. The basic idea of SPA is to store and update the supplementary parity units on one or a few newly augmented spare disks for on-line RAID systems in the operational mode, thus achieving the goals of improving the reconstruction performance while tolerating multiple disk failures and latent sector errors simultaneously. By applying the exclusive OR operations appropriately among supplementary parity, full parity and data units, SPA can reconstruct the data on the failed disks with a fraction of the original overhead that is proportional to the supplementary parity coverage, thus significantly reducing the overhead of data regeneration and decreasing recovery time in parity-based RAID systems. In particular, SPA has two supplementary-parity coverage orientations, SPA Vertical and SPA Diagonal, which cater to user’s different availability needs. The former, which calculates the supplementary parity of a fixed subset of the disks, can tolerate more disk failures and sector errors; whereas, the latter shifts the coverage of supplementary parity by one disk for each stripe to balance the workload and thus maximize the performance of reconstruction during recovery. The SPA with a single supplementary-parity disk can be viewed as a variant of but significantly different from the RAID5+0 architecture in that the former can easily and dynamically upgrade a RAID5 system to a RAID5+0like system without any change to the data layout of the RAID5 system. Our extensive trace-driven simulation study shows that both SPA orientations can significantly improve the reconstruction performance of the RAID5 system while SPA Diagonal significantly improves the reconstruction performance of RAID5+0, at an acceptable performance overhead imposed in the operational mode. Moreover, our reliability analytical modeling and Sequential Monte-Carlo simulation demonstrate that both SPA orientations consistently more than double the MTTDL of the RAID5 system and improve the reliability of the RAID5+0 system noticeably.