Redundant Array of Independent Disks (RAID) is nothing more than a configuration which uses two or more HD’s together for a particular purpose, usually for backup purposes, superior performance, or both. This age has come to desktops around about five years as a board feature of the motherboard, it was previously only possible via PCI controller cards. Almost all motherboards today have support for some RAID systems usually RAID0 and RAID1. Top-line motherboards can support RAID3 or RAID0 +1, but it is quite unusual. The performance of onboard RAID controllers is still well below the PCI cards, but for a home user is more than enough. Just remember that if you want to buy a new motherboard and RAID is something essential to you, it is always recommended to choose cards with Nvidia chipset.
Fortunately or unfortunately, Intel has yet to put a driver onboard as efficient as that of its competitor. Do not know if the standardization of motherboards with onboard RAID controllers that happened every day, the line that defines what is a desktop and a server that is or becomes more tenuous by the simple fact that manufacturers needed something to differentiate themselves from competitors. As was said initially emerged in RAID servers in order to improve stability and fault tolerance. But this is just one of more than 10 possible configurations of a RAID storage system, we’ll discuss about them below.
Types of conventional RAID :
RAID 0 : This is the configuration or applications-oriented enthusiasts who need speed, because the increase in machine performance is tremendous. In this configuration the controller will be in charge of dividing and distributing the data amongst all HD’s set up for RAID0. Thus two HD’s in RAID0 even have a nominal gain of 100% on a normal hard disk and an effective gain of about 30% which is still a lot. Striping mode it handles all the HD’s that way as a single HD, adding the capacity of all volumes and treating as a single disc.
For example: If you have two disks of 250GB, this configuration you will see as just a disk of 500GB. The biggest disadvantage with this setup is that, as all information is divided between the volumes, if by chance some volume is corrupted or damaged, all data will be lost and the configuration must be re-done.
RAID 1 : Known as mirroring, is a setting dedicated exclusively for backups purpose. It is used to mirror all data from one disk to another while maintaining an identical copy in real time of everything that happens in the disks primary, if the primary disk dies, or gets corrupt you won’t lose anything. For this, it is recommended to use HD’s of the same capacity, otherwise flush by the capacity of HD mirrored. And the HD backup can not be smaller than the main HD.
Example: If the principal is the 250GB backup also has to be at least 250GB. If the backup has 300GB, 50GB lose, it simply will not use. If you have two 250GB HD’s as major, need other two HD’s at least 250GB as backup.
RAID 2 : It is a way which is no longer used nowadays. RAID 2 is to embed error correction codes in each cluster of data recorded. However, all current drives systems comes with built-in error correction, making the system obsolete.
RAID 3 : This is a very complex system as well as requires too much capital to maintain it. RAID 3 requires 4 HD’s (actually may be less, but only justified with 4). There, the first three disc’s work just like RAID 0, all information is divided and saved as if it were a single disc. The difference is that the fourth disc serves as a recovery disk, it will save information in the other 3 disk so that if one of the first 3 disk’s die, you can recover the data from the recovery disk (ie. Disk 4). With it you receives high performance, because you have three disks in striping mode and still have the fault tolerance of these disks. Basically, it is perfect for large companies where money doesn’t matters.
RAID 4 : It is similar to RAID 3, the only difference is their way of recording data. While the RAID 3 splits the data into smaller chunks and writes in various HD’s, RAID 4 saves in large blocks. Writing in large blocks you get in a lot of speed or reading or writing data, adds up an advantage of allowing you to quickly access a larger quantity of information. In some applications, RAID 4 is on leave to RAID 3 but are quite rare and for very specific purposes.
RAID 5 : The basic idea behind RAID 5 is same as RAID 3, use of striping and fault tolerance, but works quite differently. In this configuration instead of you having a single drive with data recovery, the system spreads the data on all other HD’s. That way you don’t require a dedicated disk to that, as information is on all disks. If a disk fell, with the help of other disks you can restore it. Only two records fall you will have problems. There is a loss of performance compared to RAID 3 however, it is a more robust solution that allows greater flexibility. It is also worth remembering that you will spend a larger space, because content of recovery that was once intended for a single disk is spread to others.
So, if you have three 20GB disks in the system, you will have only 40GB free, since 20GB will be distributed among the disks so that you can recover them in case of failure. Although this system operates with three HD’s, the ideal is five. RAID 5 is great to gain speed without sacrificing fault tolerance and ideal for those companies having limited budget.
RAID 6 : RAID 6 is the setting more new and relatively rare to be found. It follows the same idea of RAID 5 but uses double-byte para for recovery. Thus, the system has fault tolerance of up to two HD’s. It is by far the conventional RAID system more expensive, which requires a minimum of 4 HD’s, but is recommended only up 6. If before with RAID 5 you had to recover the contents spread in other HD’s, here in RAID 6 you have twice the content. So if you have 10 HD’s 20GB/HD of this system, it will have 160GB free, 40GB (ie. two disks) will be distributed among the rest. RAID 6 is ideal for high risk who requires optimum performance.
