Improvement in Recovering Speed and Availability by Virtualizing Autonomous Disks

Since the recent increase of disk capacity enlarges the influence of disk failures with lengthening the suspension of service, the speedup of recovery process becomes more important to derive the high availability. We have proposed the Autonomous Disks capable of managing data, such as allocat- ing data with balancing load, handling faults, and recovering failures. In this paper, we propose to improve staggered data allocation used in the Autonomous Disks to shorten the recovery period. We parallelize the recovery process by introducing virtual disks, and consider the effect of the number of virtual disks within an Autonomous Disk.