Server 101: RAID under the scope

Fault tolerance is one of the concepts that is put into consideration when designing an infrastructure. In a nutshell fault tolerance is the ability for a device or a system to operate, even when it is faulty. One of the technologies that is based on the concept is fault tolerance is RAID.

What is RAID?

RAID which stands for Redundant Array of Independent Drives is technology that allow multiple hard disk drives to work as a single entity. As a result, characteristics of these hard drives change based on its levels up used and it is mostly implemented in servers. There are multiple setups exists, but in this post, the main three RAID levels will be covered. The main three levels are level 0, level 1 and level 5.


Data distribution in RAID 0 (Courtesy of Wikimedia)

RAID level 0 or known as RAID zero or striping is a set up consists of two or more hard drives, where data is spit across all hard drives as seen in the picture above. Such setup allows data to be read and written times faster than the specs of each hard drive. However, in case of disk failure or data corruption in a hard drive, all data will be about useless.


Data Distribution in RAID 1
Data Distribution in RAID 1 (Courtesy of Wikimedia)

RAID level 1 or known as mirroring or RAID 1 is a setup consists of two or more hard drives, where data is duplicated across all hard drives. As a result, data remain safe if a one hard drive is intact. On the other side, RAID 1 has its share of disadvantages. First, despite the amount of hard drives are added to the setup, only the capacity of the smallest hard drive will be used.  Second, RAID 1 doesn’t provide any additional performance than standalone hard drive.


Data Distribution in RAID 5
Data Distribution in RAID 5 ( Courtesy of Wikimedia)

RAID 5 that is also known as Striping with distributed sets is a setup consists of three or more hard drives. Unlike other levels, RAID 5 mechanism saves the piece of data on a hard drive and keep certain calculations on others and these calculations enable data to be recovered. Due to its mechanism RAID 5 provides continuous operation, even with a failed drive in the pool. Moreover, these calculations can lead toward a faster operation.

Hemo’s Take

RAID is a fun topic to discuss. IT pros can argue which RAID is the best RAID. Also, another argument can take place regards software based RAID versus dedicated RAID controller based RAID setup. Other storage experts might argue that RAID 5 should not be used and RAID 6 should be used instead. To be honest, what really matter is to know why and where use to use each RAID level. What matters is which RAID level is the most beneficial and most suitable for the requirements.