RAID5 PERFORMANCE WITH DISTRIBUTED SPARING

Distributed sparing is a method to improve the performance of RAID5 di sk arrays with respect to a dedicated sparing system with N + 2 disks (including the spare disk), since it utilizes the bandwidth of all N 2 disks. We analyze the performance of RAIDS with distributed sparing in normal mode, degraded mode, and rebuild mode in an OLTP environmen t, which implies small reads and writes. The analysis in normal mode u ses an M/G/1 queuing model, which takes into account the components of disk service time. In degraded mode, a low-cost approximate method is developed to estimate the mean response time of fork-join requests re sulting from accesses to recreate lost data on the failed disk. Rebuil d mode performance is analyzed by considering an M/G/1 vacationing ser ver model with multiple vacations of different types to take into acco unt differences in processing requirements for reading the first and s ubsequent tracks. An iterative solution method is used to estimate the mean response time of disk requests, as well as the time to read each disk, which is shown to be quite accurate through validation against simulation results. We next compare RAIDS performance in a system 1) w ithout a cache; 2) with a cache; and 3) with a nonvolatile storage (NV S) cache. The last configuration, in addition to improved read respons e time due to cache hits, provides a fast-write capability, such that dirty blocks can be destaged asynchronously and at a lower priority th an read requests, resulting In an improvement in read response time. T he small write penalty is also reduced due to the possibility of repea ted writes to dirty blocks in the cache and by taking advantage of dis k geometry to efficiently destage multiple blocks at a time.