RAID Redundant Array of Independent (or Inexpensive) Disks

RAID is short for Redundant Array of Independent (or Inexpensive) Disks.

It is a category of disk drives that employ two or more drives in combination with fault tolerance and performance management systems.

For example, RAID 0 implements a striped disk array in which the data is broken down into blocks and each block is written to a separate disk drive.

The enhancement in performance of RAID over a regular single drive comes from spreading the I/O load across many channels and drives. And, best performance is achieved when data is striped across multiple controllers with only one drive per controller.

Storing data redundantly and in multiple paces in a balanced way is the part that improves overall performance of the data system.

Where RAID is used

RAID disk drives are used frequently on servers. They are not generally necessary for personal computers, unless the personal computer is used for some data consuming applications where data speed and/or fault tolerance is critical, such as video production and editing, pre-press applications, etc.

RAID levels

There is a number of different RAID levels:

• Level 0: Striped Disk Array without Fault Tolerance: RAID Level 0 provides data striping, that is spreading out blocks of each file across multiple disk drives, but it offers no redundancy. This improves performance but does not deliver fault tolerance. It is not a "True" RAID because it is NOT fault-tolerant. The failure of just one drive will result in all data in an array being lost. RAID Level 0 requires a minimum of 2 drives to implement. It is easy to implement because it has a very simple design.
  
  
• Level 1: Mirroring and Duplexing: RAID Level 1 provides disk mirroring. Level 1 provides twice the read transaction rate of single disks and the same write transaction rate as single disks. 100% redundancy of data means that in case of a disk failure, data just needs to be copied to the replacement disk. Transfer rate per block is equal to that of a single disk.
  
  
• Level 2: Error-Correcting Coding: RAID Level 2 is rarely used. This stripes data at the bit level rather than the block level. The advantage is that of a relatively simple controller design compared to RAID levels 3,4 & 5. The disadvantage is a very high ratio of ECC disks to data disks with smaller word sizes which is inefficient. Also, entry level cost very high.
  
  
• Level 3: Bit-Interleaved Parity: RAID Level 3 provides byte-level striping with a dedicated parity disk. It has a very high read data transfer rate and a very high write data transfer rate. The disadvantage is that the transaction rate is equal to that of a single disk drive at best.
  
  
• Level 4: Dedicated Parity Drive: RAID Level 4 is a commonly used implementation of RAID. It provides block-level striping (like Level 0) with a parity disk. If a data disk fails, the parity data is used to create a replacement disk. The advantage is a very high read data transaction rate. It has a low ratio of ECC (parity) disks to data disks which means high efficiency. A disadvantage is that the parity disk can create write bottlenecks. It also exhibits the worst write transaction rate and write aggregate transfer rate.
  
  
• Level 5: Block Interleaved Distributed Parity: RAID Level 5 provides data striping at the byte level and also stripe error correction information. This results in excellent performance and good fault tolerance. It offers the highest read data transaction rate, medium write data transaction rate, and a very good aggregate transfer rate. The low ratio of ECC (parity) disks to data disks means high efficiency.
  
  
• Level 6: Independent Data Disks with Double Parity: RAID Level 6 provides block-level striping across a set of drives, just like in RAID 5, and allows for additional fault tolerance by using a second independent distributed parity scheme (dual parity). RAID 6 provides for an extremely high data fault tolerance and can sustain multiple simultaneous drive failures. The main disadvantage is that the controller overhead to compute parity addresses is extremely high.
  
  
• Level 10: A Stripe of Mirrors: RAID Level 10 is not really one of the original RAID levels, it is implemented as a striped array whose segments are RAID 1 arrays. Multiple RAID 1 mirrors are created, and a RAID 0 stripe is created over these. RAID 10 has the same fault tolerance as RAID level 1, and it also has the same overhead for fault-tolerance as mirroring alone.
  
  
• Level 0+1: A Mirror of Stripes: RAID Level 0+1 consists of two RAID 0 stripes where a RAID 1 mirror is created over them that is used for both replicating and sharing data among disks. This is a good solution for sites that need high performance but are not concerned with achieving maximum reliability. RAID 0+1 should not be confused with RAID 10. RAID 0+1 has the same fault tolerance as RAID level 5, and it has the same overhead for fault-tolerance as mirroring alone.

RAID is a powerful data storage system.