RAID: A Personal Recollection of How Storage Became a System

a graduate student) and I first defined the acronym RAID, or redundant arrays of inexpensive disks, in a paper we wrote in 1987.1 The RAID idea was that it was feasible to achieve significantly higher levels of storage reliability from possibly very large numbers of lower-cost and lower-reliability smaller disk drives, which were then emerging for personal computers. It accomplished this by presenting the image of single logical disk, yet arranging the underlying physical disks into arrays. Data would be partitioned and redun-dantly spread across these. Figure 1 summarizes the case we made at the time. We were not the first to think of the idea of replacing what Patterson described as a slow large expensive disk (SLED) with an array of inexpensive disks. For example