Quick links

RAID ABCs
Types of RAID
Other RAID variants
Technical Support

RAID ABCs

Redundant Array of Inexpensive Disks, or RAID, is a number of hard disks which act as an extremely reliable single disk drive. RAID uses several techniques used in RAID as explained below.

Spanning / Software Striping - Splitting information and writing it across multiple physical disk drives. RAID 0 utilizes this technique

Mirroring - Duplication of data from one disk drive to another.

Duplexing - Duplicates the disk drive as well as the disk controller.

Deferred - Data is cached in cache memory and writes to the hard disk drive as the disk drive becomes available.

Hot Swapping - Failed disk drives can be replaced and data can be placed back onto the disk drive while the remainder of the system is in operation

Hot Sparing - Disk drive is automatically initialized into the array when another fails.

Spindle Synchronization - Synchronization of the rotation of all disk drives in the array allowing information to be written all at once.

TYPES OF RAID

RAID 0

Software stripping and Block Interleave (minimum 2 drives needed )

Data is written to each drive in succession, each block going to the next available drive thus the data is distributed across the array drives (striping) providing faster operation and less chance that one drive will get overloaded with data requests. The volume can of course be much larger than any single drive. Since no redundancy is provided, the failure of a single drive will bring the system down.
Fastest and most efficient array type but offers no fault-tolerance.

RAID 1

Disk Mirroring and Duplexing ( minimum 2 drives needed )

Drives are used in pairs and all data is written identically to both drives. Each drive can be duplexed by being connected to its own interface controller. The failure of one drive will not bring down the system, instead the other drive will continue to operate. Of course, two drives are now used for the equivalent storage capacity of one drive.
There is no performance gain with this level.
The array of choice for performance-critical, fault-tolerant environments. In addition, RAID-1 is the only choice for fault-tolerance if no more than two drives are desired.

RAID 2 - Data striping and bit interleave

Data is written across each drive in succession, one bit at a time. Checksum data is recorded in a separate drive. This method is very slow for disk writes.
Seldom used today since ECC is embedded in almost all modern disk drives.

RAID 3

Data striping with bit interleave and parity checking

This level is similar to lever 2 but more reliable. Data striping is done across the drives, one byte at a time. Usually 4 or 5 drives are used providing very high data transfer rates. One drive is dedicated to storing parity information. The failure of a single drive can be compensated by using the parity drive to reconstruct the failed drive contents. Since the parity drive is accessed on every write operation, the writing of data tends to be slower. The failure of two drives or more can be a problem.
Can be used in data intensive or single-user environments which access long sequential records to speed up data transfer, however, RAID-3 does not allow multiple I/O operations to be overlapped and requires synchronized-spindle drives in order to avoid performance degradation with short records.

RAID 4

Block interleave data striping with parity checking

As in level 3, RAID 4 uses a single parity drive and block data striping like in RAID 0. The drives in this RAID level function individually, with an individual drive reading a block of data. A failure of the controller will of course be catastrophic.
Offers no advantages over RAID-5 and does not support multiple simultaneous write operations.

RAID 5

Block interleave, data striping with distributed check-data on all drives

The one to use for NetWare.
Parity information is distributed across all drives. RAID 5 efficiency goes up as the number of disks increases. You can use hot spares to rebuild a failed drive on "the fly".
The best choice in multi-user environments which are not write performance sensitive. However, at least three, and more typically five drives, are required for RAID-5 arrays.

RAID 6

Extention to RAID 5 which adds a log structured file system providing a mapping between a disk drives physical sectors and their logical representation. As information is written it will be placed to sequential physical disk sectors.

RAID 10

Stripped array whose segments are RAID 1 arrays and containing the same fault tolerance as RAID 1.

High I/O rates are achieved by stripping RAID 1 segments.

Excellent solution for those considering RAID 1 as RAID 10 provides an additional write performance, however, it is a very expensive solution.

RAID 53

Implemented as striped RAID 0 array whose segments are RAID 3 arrays. RAID 53 also contains the same Fault tolerance and overhead as RAID 3.

Excellent solution for those considering RAID 3 as RAID 53 provides additional write performance, however, it is a very expensive solution and requires all drives to have the same synchronization.