All About RAID

RAID is the short form of Redundant Array of Inexpensive Disks, where Inexpensive is often replaced with Independent. A few years back when the disks had less storage capacity with expensive prices, a couple of disks were mounted to create a larger virtual disk, and it was much more logical and easily administered.

All the disks are all placed in a confined area and if one disk crashes then a very useful amount of data can be lost. In order to avoid this is a technique called ‘Mirroring’ is taken into consideration. The data from all the disks is copied to another bulk of disks, so that if there is any corrupted disk then it would be replaced with its mirror disk. This is done for the complete integrity of your mission-critical data.

RAID Types

Although there are many types of RAID arrays, but the ones explained below are the important of all.

Concatenation:

This type of RAID array is called ’Simple’ RAIDs. It is a group of disks that are connected together tightly. Data is distributed in a linear pattern on the disk. Data is written a sequential order i-e, first disk is filled and then the text gets written on to the next disk and so on. All this is done by the volume manager, who is the in charge of the RAID. If we have less data as compared to the total size of all the disks combined then it is useless for us.

RAID-0:

It is also known as ‘striping’. Striping is much similar to concatenation, but data processing of RAID-0 is much better than its predecessor. Whenever we write some data it gets stored across all the disks. So when there is a request for data retrieval or placement it is executed much faster. Stripe width and columns are the things which make this type important.
-Stripe Width: It specifies the amount of data to be written on one particular disk.
-Column: It refers to each disk. A major flaw of RAID-0 is that if a single disk crashes than the whole system can collapse, since the data is stripped across all the disks.

RAID-1:

This type of RAID is called ‘Mirroring’. Though it’s not actually a RAID but sure is an important component of RAID. The logic behind this technique is to create a mirror image of your RAID. So that if a particular disk crashes, then the mirror will play its part. Mirroring is an excellent way of protecting the business critical data, but it comes for a good amount of money.

RAID-0+1:

This type combines Stripping and Mirroring together. We place a bunch of disks, which acts like mirrors to save the stripped data. One important thing to know is that the mirror has to be the same type of RAID as the original RAID. If we want to make a mirror copy of the simple RAID, then our mirror will be simple RAID and vice versa.

RAID-5:

RAID-5 is the best solution for increasing disk space and disk redundancy by leaps and bounds. It is similar to RAID-1 as it contains columns and stripe-widths. The difference appears when we write data. The data is written to the multiple disks at the same time, parity is also written along the data.

So even if a single disk fails then no need to worry since the data is saved on the remaining disks. Though we get quite pleased with the results but all goes in vain when we think about I/O bandwidth.