In 1987 the term Raid had the meaning of Redundant Array of Inexpensive Disks and it was intended to define a technology allowing users to have a high level of reliability storage with very inexpensive disks by arranging them into arrays for redundancy.
Within the time, marketing came up to the term and now it is known as Redundant Array of Independent Disks because marketers wanted to dissociate the word “Inexpensive” so that the user won’t expect to pay little for a RAID.
Now, RAID is used as a term for the system that divides and replicates data among different disk drives in order to achieve a good level or reliability. There are different architectures for a RAID and they are differenced by the number after the word RAID. This way now you can find RAID 0, 1, 2, 3, 4, 5, 6, 1 + 0 and 0 + 1,
Except of the RAID 0 method, all the rest are redundant which means that if a disk completely fails, all the data will be stored in other disk. This can be done by mirroring the disk. With this technique, all what is stored in a drive will also be stored in another one. Striping is another technique used in RAID so that the data is split across several disks so that it won’t be lost so easily and also error correction. The different schemes for RAID (this is 0, 1, 2, 3, 4, 5, 6, 1 + 0 and 0 + 1) use any of these techniques or a combination of them and is necessary to use two or more disks at the same time in the system.
The major advantage of using a RAID is that even if a disk fails, the data won’t be lost and will still be stored in another disk so that you can then replace the failing disk by another one. Most array systems need to switch the system off in order to change a disk, but some other systems needing high availability (such as servers for example) allow the hot swapping feature so that a disk may be replaced without having to turn all the system off.
RAID will need more computation to read and write information and this may overload the processor in cheaper systems. This is because most OS allow now the RAID function by software and then all the necessary computations for the maintenance of the RAID are made by the microprocessor. In more expensive RAID systems, this is made by hardware and a controller completely independent from the processor does all the computations needed, leaving the microprocessor free for other computations.
Although it may look as a good alternative to backup, it is not. RAID systems prevent failures of electrical kind for example or to system malfunctions, but it won’t work in the case of a software corruption or a virus infecting the system so it is only recommended for systems where the data being used at the moment can not be lost and not to store old data which may be needed in the future.
