This article aims to explain the basics of a useful technology for the proper functioning (at computer) of a company, because with this technology we are guaranteed safe storage (storage that avoids loss of information vital to the company) in addition to also ensure the speed for copying and accessing information contained on the disks in our server (as the RAID level chosen).
There are 7 levels of RAID, but only 4 are the most used.
RAID 0: This is known as "stripping." This level of RAID provides speed in copying our files because writing in these server disks (RAID function for this we need at least two discs) is divided by different disks. This is because, if data were recorded on a single disc, this process would be done sequentially. With RAID, the data for each disc is recorded at the same time. The more records we possess, the greater the writing speed . This level provides fault tolerance because there is no redundancy. This means that a failure in one of the disks can cause loss of information. This type of RIAD is advised to companies that work with applications of imaging and video.
RAID 1: It is known as "mirror". RAID 1 works by adding parallel disks to the main disk on your computer. Thus, if for example, a computer has 2 discs, you can apply more than one disk for each, totaling 4. The disks that were added work as a copy of the first. Thus, if the primary disk receives data, the added disc also gets the data. It’s why it have the name mirror because the disk becomes an almost identical copy of another. Thus, if one disk fails, another can immediately take over the operation and continue to provide the information. The consequence of this case is that the recording of data is slower because it’s performed twice. For this reason, a very common implementation of RAID 1 is its use in File servers.
RAID 2: This type of RAID adopts a mechanism to detect failures in hard drives to run in memory. Thus, all disks in the array are being "monitored" by the mechanism. Currently, RAID 2 is rarely used, since virtually all new hard disks come with mechanisms for fault detection implemented.
RAID 3 and RAID 4: RAID's are designed to give security in recording files on the server, because these RAID's use the method of parity (we have a record solely to ensure the continuity of the recording if there is any failure in another disk). The only difference between RAID 3 and RAID 4 has to do with the speed of calculation of parity (RAID 4 is the evolution of RAID 3). You need at least 3 disks to use this type of RAID.
RAID 5: This RAID is very similar to RAID 4 except for the fact that parity isn’t intended for a single disc, but for the entire array. This makes the recording of data faster because it isn’t necessary to access to a parity disk in each recording. Nevertheless, as parity is distributed among the disks, RAID 5 tends to have a little less speed than RAID 4 but is more reliable because it also provides the disk failure for parity. RAID 5 is the highest used, which provides satisfactory results in applications not very heavy. This level requires at least 3 disks to work.
RAID 0 +1: RAID 0 +1 is a combination of levels 0 (Stripping) and 1 (Mirror), where the data are divided among the disks to improve performance, but also use other disks to duplicate information. Thus it’s possible to use the optimum level 0 with the redundancy level 1. However, it takes at least 4 drives to mount a raid of this type. These characteristics make the RAID 0 + 1 the fastest and safest one, but it’s the most expensive to deploy.
