Redundant disk storage system

A storage system including a plurality of interrelated storage devices enabling recovery and a guaranteed response time. When a data request comes from a host computer, and when the number of failed disk devices shows a change in a device state management table, a determination is made as to how many of the disk devices need be read. Such a determination is based on an actual reading redundancy that indicates how many of the disk devices are allowed to be in failure, at the time of data reading, for data recovery. The actual reading redundancy is the sum of the number of failed disk devices and a predetermined reading redundancy, which indicates how many of the disk devices are yet allowed to be fail whilst enabling data recovery. Then, the thus determined number of disk devices are selected in accordance with a selection factor, and a selection result is written into a read disk management table. Then, a reading process is executed with respect to the target disk devices written in the table.