What is a RAID? RAID levels 0 to 7 simply explained! (2024)

In the areas of data storage and management, redundancy (storing content multiple times, preferably on individual storage media) and performance are of great importance. This is where RAID technology, short for Redundant Array of Independent Disks, comes into play. RAID systems have proven themselves in ensuring data integrity, availability and performance for various applications, from personal computers to enterprise-class servers. Below you will find information about RAID in general and the different levels, each designed for data security, performance or both.

What is a RAID system?

A RAID system is a method of organizing data across multiple hard drives. It is used to optimize redundancy, performance, or both at the same time. Rather than relying on a single hard drive, a RAID system spreads data across multiple independent drives, minimizing the likelihood of data loss. Because if one of the storage media fails, the data is still available on the intact media. The RAID levels briefly explained below offer different approaches to combine the benefits of redundancy and performance, depending on the needs of the user.

RAID level 0

RAID level 0, also known as "striping", is primarily intended to improve performance. Data is evenly distributed across multiple drives, resulting in faster read and write speeds. However, RAID 0 offers no redundancy, which means that failure of a single disk can result in complete data loss. So it still makes sense to rely on backup solutions such as external hard drives, cloud storage and the like so that you can continue to access your own files and programs in the event of a system failure.

• RAID 0 focuses on: Accelerating disk access
• Minimum number of hard drives: 2

RAID level 1

RAID level 1, also called "mirroring", places the emphasis on redundancy. This means that the content should be available multiple times so that it can still be used from a second storage medium if one storage medium fails. To do this, the data is mirrored on two or more identical drives. If one of the drives fails, the data on the other storage remains intact and accessible. RAID 1 offers a high level of data security, but in comparison to RAID 0 it sacrifices a bit in terms of storage utilization and overall performance.

• RAID 1 focuses on: Data security by copying all content
• Minimum number of hard drives: 2

RAID level 2

RAID level 2 was mainly used in mainframe computers and is no longer relevant today. In addition to the advantages of other RAID versions, the complicated implementation in (existing) systems also plays a role here. RAID 2 was intended to provide both performance and data redundancy, and it achieved this using techniques such as "Hamming code bits" (error-correcting data units) and "bit-level striping". This is where an “Error Correcting Code” or “Error Checking and Correcting” (both abbreviated to ECC) is used.

• RAID 2 focuses on: Protection against hard drive failure
• Minimum number of hard drives: 2

RAID level 3

In relation to the acquisition costs, the RAID level 3 is supposed to provide a very high performance gain and at the same time a high level of data security. In addition to several hard disks for storing files and their copies, an additional memory is used - the parity disk. However, this does not contain any further data copies, but the bit sum of the data disks. The bits of each data disk are added together and then checked whether the sum is even or odd. Even sums are noted with bit value 0, odd sums with bit value 1. So, compared to the previous levels, RAID 3 introduces a dedicated disk for error correction.

• RAID 3 focuses on: Secure processing of large files
• Minimum number of hard drives: 3

RAID level 4

RAID level 4 is similar to RAID 3 but uses "block-level striping" instead of "byte-level striping". Larger blocks of data are therefore stored with information about the data disks. The dedicated parity drive is involved in all read and write processes, which makes it necessary to focus on its maximum speeds. So RAID 4 only offers high speed if this also applies to the parity disk. Permanent access can also wear out this control memory more quickly. Over time, replacing the fastest possible storage medium will result in comparatively high costs.

• RAID 4 focuses on: Fast reading and writing of smaller files
• Minimum number of hard drives: 3

RAID level 5

RAID level 5 is one of the most commonly used levels alongside RAID 0, RAID 1 and the combination of RAID 0+1. The RAID 5 storage combination also uses block-level striping, but distributes the parity across all drives. This reduces potential bottlenecks and can improve overall performance. In the event of a drive failure, the data can be recovered from the remaining drives and the parity information stored there. Another advantage is, of course, that there is no need for a fast hard disk, which tends to be replaced more frequently, and is only used for parity information. RAID 5 only fails if two or more storage devices fail at the same time, which means high data security.

• RAID 5 focuses on: Reading and writing smaller files, without a parity disk
• Minimum number of hard drives: 3

RAID level 6

RAID level 6 builds on RAID 5 but offers double parity information. This allows protection against a simultaneous failure of up to two drives. The additional redundancy ensures increased data security, but is usually at the expense of writing and/or computing power. Especially since no parity bits are used, but a code for correcting single-bit errors. Various options for implementing RAID 6 can mitigate one or the other disadvantage, i.e. provide more writing speed or more performance.

• RAID 6 focuses on: Higher data security than previous levels
• Minimum number of hard drives: 4

RAID level 7

RAID level 7 is not an official designation. The individual methods used for RAID 7 are not included in the official RAID standards. Rather, the term refers to an individually adapted RAID configuration that meets specific user or company requirements.

RAID combinations and variations with individual advantages

As you can see, various factors come together when creating RAID systems to protect data. Some of the levels mentioned require a minimum number of hard drives, which also leads to a certain minimum of costs. These add up depending on the number, amount of memory, read and write speed and exchange rate or risk of failure. In addition, there is the overall speed factor and the effect on the computing power for data and parity distribution. By adapting the individual levels or a combined use, some areas can be optimized.

There are therefore RAID variations and combinations in order to improve certain security aspects and performance features from different levels individually or to use them in combination. These are not designated as separate RAID levels - this is reserved for levels 0 to 6 - but are nevertheless given their own names. This helps to identify the technologies used and to draw conclusions about the advantages and disadvantages. There are among others: RAID 01, RAID 03, RAID 05, RAID 10, RAID 1.5, RAID 1E, RAID 30, RAID 45, RAID 50, RAID 55, RAID 5E, RAID 60, Matrix RAID, RAID TP (Triple Parity = triple parity), and more.

Summary of the Redundant Array of Independent Disks

In the field of data storage, RAID has been established as an important technology to improve data integrity and, in many cases, performance. The different RAID levels offer a wide range of options to meet the needs of different users. While some levels focus more on performance, others focus on redundancy and data security. In order to prioritize or optimize individual aspects, there are also certain adjustments and combinations, each with its own advantages and disadvantages. In addition to the "how?" one should also include the "why?" when implementing the RAID system.

Links to sources with more details

In addition to the individual levels into which RAID technology can be divided, there is a lot more information. Going into all combinations and adjustments, storage formulas, alternative hard disk networks, certain technical terms and the like would go beyond the scope of this article. That's why I prefer to give you two links that supplement the explanations here:

• German Wikipedia article on this topic: Read here
• History and information from TecChannel / Computerwoche: Have a look here