|
When interpreted literally, it means storing information across an array of relatively low cost hard disk drives (HDDs). It is generally considered to be "Technology that combines numbers of such inexpensive HDDs into a single HDD."
RAID is the use of multiple disks to manage HDD data using a range of different techniques. These are typically divided into 6 levels ; RAID 0, RAID 1, RAID 2, RAID 3, RAID 4, RAID 5. They all differ in terms of data deployment and the type of redundancy offered.
The technology was initially proposed by a study group of Professor. Patterson at the University of California. When they presented their paper, the "I" in RAID represented "Inexpensive". This was because they were seeking ways to use the inexpensive HDDs that were on the market, in their study, instead of the proprietary and rather exclusive disk devices commonly used at that time to improve disk drive speed and reliability.
But today, as expensive dedicated HDDs have all but disappeared from the market, the "I" in RAID has come to represent "Independent".
It has also become popular to mix and match the various RAID level technologies to provide more specific cost reductions and performance enhancements.
Both software RAID and hardware RAID are available for installation.
Software RAID can be implemented through features that combine multiple disk devices connected directly to a host computer (typically via a SCSI interface) and regards them as a single logical memory device. This feature introduced with the operating systems Windows NT/2000 is commonly used.
With Hardware RAID, a control component, independent of the host CPU implements RAID. The two most popular methods of Hardware RAID are via PCI bus connection to the host computer using a card, or integrated with the disk drive and connected to the host computer via fibre channel or SCSI.
Hardware RAID is by far the most common method in full-fledged server systems, as it places no additional processing burdens on the server.
|