RAID is the acronym for Redundant Array of Inexpensive (or Independent) Disks. RAID is reliable fast accessible storage media.

RAID’s are the most commonly used short-term storage device technology.  A RAID consists of several magnetic storage disks, whereby the data is spread or “striped” over multiple devices to improve performance and increase reliability.  In case of a single disk failure, information can be re-generated using (redundant) parity information stored on another disk.

There are different levels of RAID’s, depending on usage of the redundancy:

• Level 0 provides data striping over multiple disks, thereby improving performance but not redundant data use.  Additionally, all data will be lost in the event of a disk error.
• Level 1 provides disk mirroring.  This is somewhat costly because all data is duplicated, requiring twice the amount of disk storage.
• Level 2 is rarely implemented because of the overhead, but stores parity information on more than one disk.  Otherwise it is the same as level 3.
• Level 3 is the same of level 0, but reserves one dedicated disk for error correction data in the form of parity data.  It therefore has some level of fault tolerance and good performance.
• Level 4 still has a single parity disk, but stripes large blocks of data over the disks, increasing read performance.
• Level 5 provides data striping at the byte level and also distributes the error correction information over the disks, providing good fault tolerance and good performance.
• Level 0+1 is a combination of level 0 and 1 and provides data striping as well as mirroring, again requiring twice the amount of storage capacity.

RAID 5 is the most commonly used type, especially in a SAN or NAS. Note that a RAID is not totally safe: if there is a simultaneous failure of multiple disks, one can still loose information. A RAID therefore always needs to be implemented hand-in-hand with an appropriate duplicate, redundant, and/or back-up device.

RAID level 5 is the most commonly used for PACS applications.