Redundant Array of Independent Disks (RAID) Standards
Airfare Daily Deals eCigarettes Eyeglasses Hotels Jewelry Online Backup Online Dating Online Printing Online Tickets Skin Care Textbook Rentals Vitamins Web Hosting Weddings
Find thousands of shopping-related forums
SEARCH

Redundant Array of Independent Disks (RAID) Standards

The Redundant Array of Independent Disks (RAID) standards are a set of vendor-independent specifications for fault-tolerant configurations on multiple-disk systems. If one or more of the disks fails, data can be recovered from the remaining disks. RAID can be implemented through operating system software, but hardware-based RAID implementations are more efficient and are more widely deployed. There are several RAID levels, each of which provides a different combination of features and efficiencies. RAID levels are identified by number; RAID 0, RAID 1, and RAID 5 are the most common.

The Redundant Array of Independent Disks (RAID) standards are a set of vendor-independent specifications for fault-tolerant configurations on multiple-disk systems. If one or more of the disks fails, data can be recovered from the remaining disks. RAID can be implemented through operating system software, but hardware-based RAID implementations are more efficient and are more widely deployed. There are several RAID levels, each of which provides a different combination of features and efficiencies. RAID levels are identified by number; RAID 0, RAID 1, and RAID 5 are the most common. The original RAID specifications were titled Redundant Array of Inexpensive Disks. As the disk cost of RAID implementations has become less of a factor, the term “Independent” disks have been widely adopted instead. The original RAID standards were developed in 1987, but the specification has grown to include additional RAID types that are variations on or combination of the original standards. The following levels summarize the original RAID types:

Striping (RAID level 0): RAID level 0 implements striping, which the process of spreading data across multiple drives. Striping can dramatically improve read and write performance. Striping provides no fault tolerance; however, because the data is spread across multiple drives, if any one of the drives fails, you will loss all your data. You must have at least two physical disk drives to implement striping. Windows Servers provide built-in support for RAID level 0.

Mirroring or Duplex (RAID level 1): In RAID level 1, data from an entire partition is duplicated on two identical drives by either mirroring or duplex. In mirroring, the two disks share a drive controller. In duplex, each disk has its own drive controller, so the controller card is not a single point of failure. Data is written to both halves of the mirror simultaneously. Windows Servers support both mirroring and duplex across SCSI, ESDI, and IDE disk drives.

Striping with Error Correction Code (RAID level 2): Striping with error correction code spreads data, bit by bit, across multiple drives. Error correction code information is built from the bits and stored on a separate drive.

Striping with Partly on A Single Drive (RAID level 3): In a RAID 3 level system, data is spread byte by byte across multiple drives. A RAID level 3 systems required at least three, and usually no more than five, drives. RAID level 3 system provides both performance and fault tolerance enhancements. In multi-drive systems, files can be written or read faster than in a single drive system.

Striping by Block with Parity on a Single drive (RAID level 4): A RAID level 4 systems spreads data block by block across multiple drives. A block refers to whatever the block size is on the disks. Usually, blocks are groups of 1 to 16 disk sectors. In a separate drive the all parity information stored. A RAID level 4 systems uses at least two, and usually no more than five, drives. RAID level 4 systems provide both read-performance and fault tolerance.

Striping with Parity Spread Across Multiple Drives (RAID level 5): RAID level 5 systems spread data byte by byte across multiple drives, with parity information also spread across multiple drives. You need at least three physical disk drives. If one drive falls, the parity information on the remaining drives can be used to reconstruct the data. With RAID 5, disk performance is enhanced because more than one read and write can occur simultaneously.

However, the parity calculations create some write performance overhead.

 

Need an answer?
Get insightful answers from community-recommended
experts
in Hardware Engineering & Design on Knoji.
Would you recommend this author as an expert in Hardware Engineering & Design?
You have 0 recommendations remaining to grant today.
Comments (0)
ARTICLE DETAILS
RELATED ARTICLES
ARTICLE KEYWORDS