SPARING FOR REDUNDANT DISK ARRAYS

One of the main advantages of a redundant disk array architecture is t hat it provides fault tolerance against disk drive failures. This impr ovement in reliability can be further enhanced if spare drives are als o added to the array since a failed drive can be expeditiously replace d. Furthermore, even though data can be reconstructed from the other d rives of a redundant array in the event of a drive failure, performanc e is degraded substantially in this mode of operation. Clearly it is d esirable to get out of this degraded mode of operation as quickly as p ossible. Again, having spare drives in the array will facilitate that. The purpose of this paper is to study some of the issues related to s paring in a redundant disk array. In particular, we will try to unders tand the effect on reliability of different sparing schemes. We will a lso examine the economic trade-offs of having spare drives in a system .