RAID 0

RAID 0 distributes data across a number of disks

• Each block of data is written to a separate disk drive
• All disks must be of the same size
• Not fault tolerant

RAID 1

RAID 1 places exact copies of data on a set of disks

• duplicate data on several disks 100% redundancy
• no data rebuild necessary in case of disc failure
• highest disc overhead of all RAID types

RAID 2

Bit striping with dedicated parity disc

• parity calculated and stored on a dedicated disc
• high transfer rates for sequential reads
• poor performance for multiple requests
• in case of failure data is rebuild from remaining discs
• if more than one disc fails data is lost
• write transactions slower due to parity that has to be calculated
• rarely used

RAID 3

Byte striping with dedicated parity disc

• parity calculated and stored on a dedicated disc
• high transfer rates for sequential reads
• poor performance for multiple requests
• in case of failure data is rebuild from remaining discs
• if more than one disc fails data is lost
• write transactions slower due to parity that has to be calculated
• rarely used

RAID 4

Block striping with dedicated parity disc

• parity calculated and stored on a dedicated disc
• in case of failure data is rebuild from remaining discs
• if more than one disc fails data is lost
• write transactions slower due to parity that has to be calculate

RAID 5

Block level striping with distributed parity

• parity calculated and stored across all discs disc failures have lesser impact on performance
• disc failures have lesser impact on performance
• in case of failure data is rebuild from remaining discs
• if more than one disc fails data is lost
• fast read transactions
• write transactions slower due to parity that has to be calculated

RAID 6

Block level striping with dual distributed parity

• Each block of data is written to a separate disk drive
• All disks must be of the same size
• Not fault tolerant

Nested RAID

Nested RAID combines two or more RAIDs. This allows to achieve better performance and reliability.

Controller redundancy

To achieve a truly available RAID, the RAID controller should also be redundant to prevent RAID failure caused by controller failure. An additional benefit of a redundant RAID controller is enhanced performance and the possibility of load balancing.

Hot spare discs

Unused hard drives in the disc array, which will automatically be used to replace a failed discs. This allows to limit RAID vulnerability.