RAID Was Named in a Grad Student Paper to Mock Big Iron

In 1988, three researchers at UC Berkeley — David Patterson, Garth Gibson, and Randy Katz — published "A Case for Redundant Arrays of Inexpensive Disks (RAID)." The title was deliberate provocation. At the time, IBM's mainframe storage strategy rested on large, proprietary hard drives that cost tens of thousands of dollars each. Patterson and his co-authors proposed that you could beat those drives in performance and reliability by combining many cheap commodity disks in parallel.

They invented a companion acronym to make the contrast obvious: SLED, for Single Large Expensive Disk. The paper argued that an array of small, inexpensive disks could match or exceed a SLED's throughput while surviving individual disk failures — provided you added redundancy schemes like striping and parity.

The paper defined five RAID levels (0 through 4), each with different tradeoffs between capacity, performance, and fault tolerance. A sixth, RAID 5, was added shortly after and became the most widely deployed in enterprise environments because it distributed parity across all drives rather than dedicating one.

By the early 1990s, vendors selling enterprise storage arrays had a problem with the word "Inexpensive" — it sounded cheap in a market where buyers associated price with quality. The Storage Networking Industry Association quietly redefined the 'I' as "Independent," and that version appears in most modern documentation. The original paper still says what it said in 1988: the whole point was to beat IBM with hardware that cost less.