Although you might think it's what to use to eliminate the bugs in poorly written software, it stands for Redundant Array of Inexpensive (or Independent) Disks (or Drives). RAID technology includes several methods of ganging together two or more off-the-shelf hard drives to create bigger, faster, and safer storage subsystems that you can use just like the single drives you're used to. Six types of RAID have been defined named Level 0 to Level 5, all involving different data storage schemes.
RAID storage divides data across two or more drives for faster read and write times. It uses a speed-enhancing technique known as "striping" to break up data onto different drives. Most of the time it involves the additional techniques of "mirroring" and "parity" to provide backup data in case of a drive failure.
Mirroring protects data by storing at least one extra copy of the data on a separate disk in the array, writing all data at least twice. This method is safe, but not as fast as the other methods. Parity can be thought of like this: it's as if the process creates a multi-part equation from your data and stores one part of the equation on each data drive and the solution on a parity drive. If a data drive fails, the system can recalculate the equation to reveal the missing bits. But if two drives fail, not enough information remains for the parity scheme to recover your data. Replacing a drive in case of failure is called swapping.
The simplest, most popular, and least expensive RAID implementation is Level 0, which involves striping only, providing maximum speed and performance, but, contrary to its name, providing no redundancy, since no mirroring or parity is used. According to vendors and industry analysts, about 80 percent of RAID arrays sold are configured for Level 0. How the data in a RAID array is broken up is controlled by RAID software or by a hardware controller built into a circuit board in the RAID enclosure or on an adapter board you install in your computer.
K.O.P.C.Online home > table of contents > RAID