Chip-Level RAID with Flexible Stripe Size and Parity Placement for Enhanced SSD Reliability

The move from SLC to MLC/TLC flash memory technology is increasing SSD capacity at lower cost, but at the cost of sacrificing reliability. An approach to remedy this loss is to employ the RAID architecture with the chips that comprise SSDs. However, using the traditional RAID approach may result in negative effects as the total number of writes is increased due to the parity updates. In this paper, we describe Elastic Striping and Anywhere Parity (eSAP)-RAID, a RAID scheme that allows flexible stripe sizes and parity placement. Using performance and lifetime models that we derive of SSDs employing RAID-5 and eSAP-RAID, we show that eSAP-RAID brings about significant performance and reliability benefits by reducing parity writes compared to RAID-5. We also implement these schemes in SSDs using DiskSim with SSD Extension and validate the models using realistic workloads. We also discuss policies such as dynamic stripe sizing and selective data protection that exploits the flexible nature of eSAP. We show that through such policies particular reliability enhancement goals can be met.

1 Introduction

We make three contributions in this paper. First, we present reliability analyses of two RAID schemes; one, the well known RAID-5 scheme and two, eSAP (elastic Striping and Anywhere Parity), a novel scheme that was previously proposed [1], [2] and which we describe in detail later. Next, through analyses and extensive experiments using the DiskSim with SSD Extension platform, we show that eSAP provides reliability better than RAID-5 by significantly reducing the parity overhead and, eventually, reduced P/E cycles for both MLC and TLC based SSDs. Finally, we show the flexible usability of eSAP through two examples, dynamic stripe sizing and selective data protection. Dynamic stripe sizing is a means to adjust the stripe size to improve reliability as wear increases, while selective data protection is a means to target particular data for parity protection. For the later, in particular, we devise a scheme, which we call Meta-eSAP, where reliability of metadata such as the superblock of the file system is enhanced by applying eSAP to just these types of data. We provide quantitative analyses that show that the two methods can improve reliability in particular ways.