An Introduction To RAID
What is RAID?
Depending upon whom you speak to RAID is an acronym for either:
Redundant Array of Inexpensive Disks or,
Redundant Array of Independent Disks
Whilst there is a basis in the historical documentation of RAID systems for the first there is an argument that the latter is a more apt description of a RAID system.
Whichever way you interpret the acronym the idea of a RAID system is the same; to provide a cheap, efficient method of storing data on hard disk drives, whilst reducing the overall likelihood of data loss by providing redundancy for failure of one or more members of the RAID system or Array.
This redundancy is provided either by a mirroring process, whereby the data is written in duplicate to any number of hard drives, through to the use of various mathematical methods, from Exclusive Disjunction or XOR in RAID 5 systems to RAID 6 systems which build upon this with extra error checking Reed Solomon code to provide for further drive loss prevention.
Support
Most modern operating systems, including, Microsoft Windows, Mac OS X, almost all versions of Linux as well as Unix and Solaris systems support software RAID setups, however, it is often far more efficient when implemented through a dedicated array controller where any redundancy calculation can be achieved in hardware rather than software.
There is also RAID 0, whilst RAID 0 arrays do share some similarities with a RAID system such as multiple hard disk drives and block striping, RAID 0 is not a true form of RAID as there is no redundancy against failure. If you lose a drive within a striped array, you will lose access to the entire volume and of course all of you data.
History
The concept of the true RAID system originated in the late 1980s from the University of California, Berkeley, where David A. Patterson directed a team that later came to develop the basis of all modern RAID arrays.
Prior to this IBM pioneered a process of mirroring data developing what would become RAID 1, IBM were also responsible for systems that were identical in concept to RAID 5 systems, independently and almost concurrently with Patterson's team.
RAID Controllers
Drives in a RAID system must be controlled by either dedicated hardware or by software.
Hardware solutions come in many forms, each offers a slightly different implementation of the specific RAID level that they support. Historically RAID cards were SCSI, IDE RAID controllers exist but are less common and most modern controllers are now either SATA or SAS (Serial Attached SCSI).
Most modern operating systems also support the creation of RAID arrays in software. Microsoft Windows supports RAID levels 0, 1 and 5, but support varies between versions, with only Microsoft Windows Server versions supporting all 3.
Mac OS X also supports 0, 1 and 5, however, this support is available throughout all versions of the OS via Disk Utility. Linux support via LVM (Logical Volume Manager) is extended to almost all modern OSs.
When a RAID is configured in software there are additional overheads for the main system processor - it is normally not the most efficient method of implementing RAID.
Modern Usage
RAID is widely used in servers and systems where data loss is not desired or cannot be tolerated. Most workgroups, networks, web servers and archiving systems use some method of redundancy alongside a strong data backup policy to ensure that data cannot be lost, however, not all RAID systems are created equal and not all file Servers are as secure as they seem.