RAID (data management)

From Conservapedia - Reading time: 3 min

---

RAID (Redundant Array of Inexpensive Disks or Drives, or Redundant Array of Independent Disks) is a class of data storage architecture intended to reduce the risk of data loss. Since storage media will inevitably fail eventually for one reason or another, RAID attempts to provide redundancy for active data which will enable users to survive drive failures with all of their data intact. The exception to this in RAID 0, which provides no redundancy, but offers added functionality as long as there are no failures.

Striping[edit]

An integral part of all standard RAID architectures except for RAID 1 is striping. Striping is the process of breaking data into chunks, and spreading these chunks across multiple drives. So, if a single massive file were to be placed on a striped array, multiple small segments of that file would be deposited on each drive. When that file is then read in its entirety, all of the drives would go to work simultaneously, loading the needed data chunks from each drive and assembling them.

RAID types[edit]

(popular types in bold)

RAID type	Function	Notes
RAID 0	Stripes data across all drives in an array, making many physical drives one virtual drive. This offers both convenience and speed, since the drive array looks and acts like one drive, while the read/write load is distributed across multiple drives	This offers no redundancy, so a single drive failure can destroy much if not all of the data therein.
RAID 1	Mirroring: the contents of one drive are copied bit for bit onto another drive. In case of failure, there is already a complete copy ready to use.	This is the most "expensive" kind of RAID, since half (or more, if non-equivalent drive sizes are used) is lost to redundancy.
RAID 2
RAID 3	All but one drive uses RAID 0. The last drive is dedicated to storing a parity bit from each sequential bit across all stripes.	Offers single-drive redundancy
RAID 4
RAID 5	Uses one drives-worth of space out of the array (striped across all drives) to store a parity bit from each sequential bit across all stripes. When any one drive fails, the contents of that lost drive can be slowly calculated out using the remaining drives and the parity data.	This only offers single-drive redundancy, so if a second drive fails (something not uncommon during a restoration, due to the increased load) all data may be lost. With increasing hard drive sizes, this is increasingly a problem which is driving many people to RAID 6.
RAID 6	Uses two drives-worth of space out of the array (striped across all drives) to store parity data from each sequential bit across all stripes. When one or two drives fail, the contents of that lost drive(S) can be slowly calculated out using the remaining drives and the parity data.	This cost more storage space than RAID 5, but offers added protection against a second failure during the restoration process.
RAID 10	Uses RAID 1 to mirror a RAID 0 stripe	Sometimes called "RAID 1+0" since RAID 1 is used to mirror each drive in the RAID 0 stripe. The cost of this is roughly equivalent to that of RAID 1, but this provides better overall performance at larger scale.

^[1][2]

Uses[edit]

A form of RAID is almost always used in commercial data centers. Typically, a RAID architecture is set up when the servers are first installed. One or more "hot spares" (empty drives which are connected, powered, and ready to be used) are also attached, so in case of a drive failure, the server can automatically begin replicating the lost data onto one of these. A technician will then go and replace the failed drive with a "cold spare" (a spare drive which is not connected or powered) to take the place of the host spare which has been used.
As data storage demands keep increasing, individuals are also finding that they either must use relatively large storage devices, or upload everything over the internet to someone else's servers ("the cloud"). Many users will only have a single copy of their data, or only make periodical backups. Since the cloud manages RAID automatically, users typically do not need to worry about data loss. However, individuals with large amounts of data who wish to store it themselves often turn to RAID arrays of their own. Hard drives can be configured to use RAID within some desktop computers (when the hardware permits) or using external enclosures. Some individuals are even using RAID on flash drives,^[3] even though this reduces the life expectancy of each individual drive.

References[edit]