Understanding RAID Levels

Spread the love

RAID or Redundant Array of Independent Disks is a method of increasing storage performance and making it fault tolerance. Basically RAID combines two or more hard disks into a group which seen by the OS as a single volume. There are several ways to configure RAID depending on the situation and the goals. Understanding RAID levels is part of storage optimization strategy. One can utilize RAID to improve the overall service experience by ensuring optimum storage operation and increase service availability by enabling quick recovery from failure at the storage level.

Understanding RAID Levels & Choosing the Right Strategy

There are four levels of RAID that are commonly used in the industry: RAID 0, RAID 1, RAID 5, and RAID 10. These levels are supported by almost every RAID controller, be it hardware or software. Below are the explanation of each level:

RAID level 0 usually called striping. It basically splits the data across multiple disks (minimum 2). Data operation is handled by multiple disks instead of one, thus increasing the I/O performance and improving the overall server operation. The disks forming the array needs to be the same size and speed. The total capacity is the sum of all disks capacity, for example: combining two 1TB disks in RAID 0 mode will result in total capacity of 2TB (1TB + 1TB). There is no redundancy at this level, because when one of the hard disks fail the whole array will also fail.

Understanding RAID Levels

Disk requirement: 2 (minimum), Performance: Excellent, Redundancy: None, Cost: $$
When to choose this level: RAID 0 is best used for non-critical system which involves intense read/write activity and bigger storage, for example: video editing, image rendering, etc.

RAID level 1 is called mirroring, because it literally creates the duplicate of the data in one disk to another disk. RAID 1 is focusing on redundancy instead of performance. RAID 1 requires exactly 2 disks with same size and speed in the array. When one of the disks fail, the other can still operate. There is no addition to the capacity at this level, the total capacity stay the same as if there is only one disk. For example: combining two 1TB disks in RAID 1 mode will result in total capacity of 1TB.


Disk requirement: 2, Performance: Great, Redundancy: Excellent, Cost: $$
When to choose this level: RAID 1 is best used for critical system which requires continuous operation and where storage downtime is not acceptable, such as: file server, financial database, etc

RAID level 5 is striping with parity. RAID 5 also splits data across multiple disks like RAID 0, but it adds parity checksum on each disk. Parity checksum is a data that can be used by the system to recreate missing data from one of the other disks should it experiencing failure. The way it adds parity checksum makes it required at least 3 disks to create the array, but a RAID 5 array can survive when one of the disks failed without any data loss. The total storage capacity of RAID 5 array is number of the disks minus one (N-1). RAID 5 array is more fault tolerance than RAID 0 and has bigger storage capacity than RAID 1.


Disk requirement: 3 (minimum), Performance: Great, Redundancy: Great, Cost: $$$
When to choose this level: RAID 5 is a balanced option that provides both performance and redundancy while still maintaining the cost efficiency. It is best used for system which requires space and intense read activity.

RAID level 10 is combination of mirroring and striping. This is why RAID 10 can also be referred as RAID 1+0 or hybrid RAID. It provides the advantage of RAID 1 (fault tolerance) and RAID 0 (increase performance). RAID 10 works by mirroring the data to the secondary disk then split them evenly across the remaining disks. With the way it works, it requires minimum 4 disks and the number of the disks in the array should be even. Total storage capacity of a RAID 10 array is number of the disks divided by two (N/2).


Disk requirement: 4 (minimum — must be even), Performance: Excellent, Redundancy: Excellent, Cost: $$$$
When to choose this level: RAID 10 is the most expensive option among the others, however it provides more redundancy and smoother read/write operation compared to RAID 5. It is very recommended for every critical system, especially in enterprise level.


After understanding RAID levels, administrator can determine which one that fit in with the condition and requirement. However, it is important to note that RAID is not a replacement for backup solution. RAID is a good way to increase storage performance and enable quick recovery from disk failure, but having a backup solution will help to secure the data against accident or disaster.

The following two tabs change content below.

Arranda Saputra

ITIL Certified, CCNA, CCDA, VCP6-DCV, MCSA Administering Windows Server 2012
I am IT practitioner in real life with specialization in network and server infrastructure. I have years of experience in design, analysis, operation, and optimization of infrastructure solutions for enterprise-scaled network. You can send me a message on LinkedIn or email to arranda.saputra@outlook.com for further inquiry regarding stuffs that I wrote or opportunity to collaborate in a project.

Latest posts by Arranda Saputra (see all)

scroll to top