The Computer Revolution/Hardware/RAID

From Wikibooks, open books for an open world
Jump to: navigation, search

RAID (redundant array of independent disks) is a storing data method on more than one hard drives. It can be used to increase performance and also most often used to protect critical data on a storage server. Usually RAID involves recording duplicate copies of stored data, they can be used when necessary and to reconstruct lost data. It helps to increase the fault tolerance, such as a system crash of a storage system. An example or a RAID, uses striping which spreads files over two or more hard drives. Another RAID technique is disk mirroring, which is when data is written to two duplicate hard drives simultaneously.

Raid uses disk striping which can spread files over 2 or more hard drives. Disk striping allows for better performance, and retrieve data from multiple hard drives. RAID also uses disk mirroring which allows writing of two hard drives at the same time, this allows a back up if one hard drive should fail. If a hard drive fails disk mirroring allows the system to switch to another hard drive without loosing any data. RAID is currently working with a new storage system called Drobo, which connects to a computer, and external hard drive. This system requires no upgrades, and are easy to manage and repair.

RAID Designs[edit]

RAID with two disks mirrored from drive 0 to drive 1

There are several different RAID designs or levels that use different combinations of RAID techniques, in example:


  • RAID 0 uses disk striping, which spreads files over two or more hard drives
  • RAID 1 technique is disk mirroring, where data is written to two duplicate hard drives simultaneously.
  • RAID 5


Disk Mirroring[edit]

The objective of disk mirroring is to increase fault tolerance. For example, if one hard drive fails, the system will instantly switch to the other hard-drive without any loss of data.

While RAID1 (mirroring) was originally developed as a way to improve reliability no matter what the impact on performance, it turns out that modern N-disk RAID1 implementations have up to N times faster read performance than a single disk, and the same write performance as a single disk.[1][2][3][4][5]


RAID 5[edit]

RAID 5 provides data striping at the byte level and also stripe error correction information. Raid 5 is very popular because it is relatively fast and provides superior data protection, also known as a high fault tolerance. If any drive in the array fails the data is automatically recovered by the system as it rebuilds using the parity data with its associated mathematical code. Parity is essentially the final result of a comparison between sets of data written to the separate drives of the array. The requisite code to rebuild comes from the parity bits and mathematical calculations performed using a simple formula which is encompassed in the logic used in an XOR table. Using parity increases the reliability of the system to enable a full, error-less backup. RAID 5 will be more expensive to use, partly because it requires a minimum of three hard drives for it to work and partly because one whole drive is needed to enable the data needed to repair a data loss. RAID 5 using a 4 drive setup is very common and its capacity is the sum of the space available on Disk1+Disk2+Disk 3. Disk 4 provides the redundancy needed to provide the high fault tolerance of the RAID 5 system. Therefore, if all the drives are 3TB each, the maximum data capacity of the system is 9TB. Failure of any one drive in the system is of no consequence -- as the parity bits are recorded on each drive and can be rebuilt from there.

RAID 5 array using 4 disc drives

References used:

http://www.serverwatch.com/tutorials/article.php/3618421/Understanding-RAID.htm

http://www.webopedia.com/TERM/R/RAID.html

http://www.youtube.com/watch?v=LTq4pGZtzho&feature=related

Setup[edit]

Often a system administrator puts together all the hardware first, containing only blank hard drives; then turns the system on and sets up RAID as part of the operating system installation.[6][7][8][9]

Further reading[edit]

  1. Tucker Balch. "10 Surprising Facts About RAID". 2012.
  2. "Speed: RAID 1 Mirroring". 2014.
  3. "Does RAID 1 have faster slower or the same data transfer speed as a single drive would have?".
  4. "RAID-1 vs single drive performance".
  5. "How much slower is a RAID 1 hard drive compared to no RAID?".
  6. "Installation: Ubuntu Software RAID".
  7. "RAID setup".
  8. "Using mdadm to Configure RAID-Based and Multipath Storage".
  9. "How to install CentOS 5 & 6 on a software partitionable RAID1.".