Disk drive arrays are receiving attention from computer researchers looking for higher performance mass storage. RAID (Redundant Array of Inexpensive Disks) appears to offer benefits over storage systems based on large format disks. This paper investigates disk drive and disk support hardware failures as factors in the data integrity of RAID. Through the use of redundancy in the disk array, support hardware failures can be eliminated as a major factor in the data integrity of RAID. This paper also presents some practical considerations in the design of a RAID prototype and discusses how such a prototype could be constructed at Berkeley in the near future. The prototype will most likely be based on 5.25 inch disk drives and the Small Computer System Interface (SCSI) bus.