A RAID reconfiguration scheme for gracefully degraded operations

One distinct advantage of Redundant Array of Independent Disks (RAID) is fault tolerance. But the performance of a disk array in degraded mode is so poor that no one uses the RAID after failure. Continuous operation of RAID in degraded mode is very important in many real time applications, which can not be interrupted in providing continuous services. In this paper, we propose an efficient architectural reconfiguration scheme to enhance the performance of RAID-5 in degraded mode, called reconfigurable RAID-5. It reconfigures RAID-5 to RPTD-0 in degraded mode. Using this scheme, the calculation of the failure data and the generation of parity in writing the new data to the failed disk can be reduced. It also alleviates the small write problem for RAID-5 in degraded mode. We use the phase parallel model to analyze the total execution time of the RAID-5 and of the reconfigurable RAID-5. Through theoretical analysis and benchmark test, we find the performance of the reconfigurable RAID-5 can be 200 times better than conventional RAID-5.