Using Flash to Tolerate Track Failures in RAID

Author

Zheng Chen ; McBride, Allen

Author_Institution

Sch. of Comput. & Eng., Univ. of Electron. Sci. & Technol. of China, Chengdu, China

fYear

2014

fDate

9-11 Jan. 2014

Firstpage

234

Lastpage

235

Abstract

RAID systems are designed to tolerate disk failures, but latent track failures can prevent recovery. Erasure codes can tolerate latent track failures through schemes such as RAID 6. With traditional erasure codes, this resilience often requires high space or computation overhead. In this paper, we introduce Plus-One codes, which use a plug-and-play flash drive to add to a pre-existing RAID system the ability to tolerate one additional track failure more efficiently than using a different RAID design to tolerate that failure. Augmenting a RAID system using a Plus-One code requires little additional space or computation overhead. Plus-One codes can also be used to add flash memory to RAID cards for greater efficiency and transparency.

Keywords

RAID; disc drives; error correction codes; failure analysis; flash memories; RAID 6; RAID cards; RAID systems; computation overhead; disk failure; erasure codes; flash memory; latent track failures; plug-and-play flash drive; plus-one codes; tolerate track failures; Drives; Encoding; Fault tolerance; Fault tolerant systems; Flash memories; Galois fields; Universal Serial Bus; Plus-One code; RAID; track failure tolerance;

fLanguage

English

Publisher

ieee

Conference_Titel

High-Assurance Systems Engineering (HASE), 2014 IEEE 15th International Symposium on

Conference_Location

Miami Beach, FL

Print_ISBN

978-1-4799-3465-2

Type

conf

DOI

10.1109/HASE.2014.41

Filename

6754612