A Cost Effective Fault-Tolerant Scheme for RAIDs

The rapid progress in mass storage technology has made itpossible for designers to implement large data storage systems for avariety of applications. One of the efficient ways to build largestorage systems is to use RAIDs as basic storage modules. In general,the data can be recovered in RAIDs only when one error occurs. But inlarge RAIDs systems, the fault probability will increasewhen the number of disks increases, and the use of disks with bigstorage capacity will cause the recovering time to prolong, thusthe probability of the second disk's fault will increase. Therefore, itis necessary to develop methods to recover data when two or moreerrors have occurred. In this paper, a fault tolerant scheme is proposed basedon extended Reed-Solomon code, a recovery procedure is designed to correctup to two errors which is implemented by software and hardwaretogether, and the scheme is verified by computer simulation. In thisscheme, only two redundant disks are used to recover up to two disks'fault. The encoding and decoding methods, and theimplementation based on software and hardware are described. Theapplication of the scheme in software RAIDs that are built in clustercomputers are also described. Compared with the existing methods such asEVENODD and DH, the proposed scheme has distinct improvement inimplementation and redundancy.