What is RAID (redundant array of independent disks)?

How RAID works

RAID works by placing data on multiple drives and allowing input/output (I/O) operations to overlap in a balanced way, improving performance. Because using multiple drives increases the mean time between failures, storing data redundantly also increases fault tolerance.

RAID arrays appear to the operating system (OS) as a single logical drive.

RAID employs the techniques of disk mirroring or disk striping. Mirroring copies identical data onto more than one drive. Striping partitions help spread data over multiple drives. Each drive's storage capacity is divided into units ranging from a sector of 512 bytes up to several megabytes. The stripes of all the drives are interleaved and addressed in order. Disk mirroring and disk striping can also be combined in a RAID array.

In a single-user storage system where large records are stored, the stripes are typically set up to be small -- 512 bytes, for example -- so that a single record spans all the drives and can be accessed quickly by reading all the drives at the same time. In a multiuser system, better performance requires a stripe wide enough to hold the typical or maximum size record, enabling overlapped disk I/O across drives.

RAID controller

A RAID controller is a device used to manage drives in a storage array. It can be used as a level of abstraction between the OS and the physical drives in the data center, presenting groups of drives as logical units. Using a RAID controller can improve performance and help protect data in case of a crash.

A RAID controller may be hardware- or software-based. In a hardware-based RAID product, a physical controller manages the entire array. The controller can also be designed to support drive formats such as Serial Advanced Technology Attachment and Small Computer System Interface. A physical RAID controller can also be built into a server's motherboard.

With software-based RAID, the controller uses the resources of the hardware system, such as the central processor unit (CPU) and memory. While it performs the same functions as a hardware-based RAID controller, software-based RAID controllers may not enable as much of a performance boost and can affect the performance of other applications on the server.

If a software-based RAID implementation is not compatible with a system's boot-up process and hardware-based RAID controllers are too costly, firmware, or driver-based RAID, is a potential option.

Firmware-based RAID controller chips are located on the motherboard, and all operations are performed by the CPU, similar to software-based RAID. However, with firmware, the RAID system is only implemented at the beginning of the boot process. Once the OS has loaded, the controller driver takes over RAID functionality. A firmware RAID controller is not as pricey as a hardware option, but it puts more strain on the computer's CPU. Firmware-based RAID is also called hardware-assisted software RAID, hybrid model RAID and fake RAID.

RAID levels

RAID devices use different versions called levels. The original paper that coined the term and developed the RAID setup concept defined six levels of RAID -- 0 through 5. This numbered system differentiated the capabilities of the RAID versions. The number of levels has since expanded and has been broken into three categories: standard, nested and nonstandard RAID levels.

Standard RAID levels

RAID 0. This configuration has striping but no data redundancy. It offers the best performance, but it does not provide fault tolerance.

RAID 1. Also known as disk mirroring, this configuration consists of at least two drives that duplicate the storage of data. There is no striping. Read performance is improved since either disk can be read at the same time. Write performance is the same as that of single-disk storage. RAID 1 is an effective tool for disaster recovery.

RAID 2. This configuration uses striping across drives, with some drives storing error-correction code (ECC) information. RAID 2 also uses dedicated Hamming code parity, a linear form of ECC. RAID 2 has no advantage over RAID 3 and is no longer used.

RAID 3. RAID 3 uses striping and dedicates one drive to store parity information. The embedded ECC information is used to detect errors. Data recovery is accomplished by calculating the exclusive information recorded on the other drives. Because an I/O operation addresses all the drives at the same time, RAID 3 cannot overlap I/O. For this reason, RAID 3 is best for single-user systems with long-record applications.

RAID 4. RAID 4 uses large stripes, which means a user can read records from any single drive. Overlapped I/O can then be used for read operations. Because all write operations are required to update the parity disk drive, no I/O overlapping is possible.

RAID 5. RAID 5 is based on parity block-level striping. The parity information is striped across each drive, enabling the array to function, even if one drive were to fail. The array's architecture enables read/write operations to span multiple drives. This results in performance better than that of a single drive but not as high as a RAID 0 array. RAID 5 requires at least three drives, but it is often recommended to use at least five for performance reasons.

RAID 5 arrays are generally considered to be a poor choice for use on write-intensive systems because of the performance impact associated with writing parity data. When a disk fails, it can take a long time to rebuild a RAID 5 array.

RAID 6. RAID 6 is similar to RAID 5, but it includes a second parity scheme distributed across the drives in the array. The use of additional parity enables the array to continue functioning, even if two drives fail simultaneously. However, this extra protection comes at a cost. RAID 6 arrays often have slower write performance than RAID 5 arrays.

Nested RAID levels

Some RAID levels that are based on a combination of RAID levels are referred to as nested RAID. Here are some examples of nested RAID levels.

RAID 10 (RAID 1+0). RAID 10, which combines RAID 1 and RAID 0, offers higher performance than RAID 1 but at a much higher cost. In RAID 10, the data is mirrored, and the mirrors are striped.

RAID 01 (RAID 0+1). RAID 01 is similar to RAID 10, except the data organization method is slightly different. Rather than creating a mirror and then striping it, RAID 01 creates a stripe set and then mirrors the stripe set.

RAID 03 (RAID 0+3, also known as RAID 53 or RAID 5+3). This level uses striping in RAID 0 style for RAID 3's virtual disk blocks. RAID 03 offers higher performance than RAID 3 but at a higher cost.

RAID 50 (RAID 5+0). RAID 50 combines RAID 5 distributed parity with RAID 0 striping to improve RAID 5 performance without reducing data protection.

Nonstandard RAID levels

Nonstandard RAID levels have features that vary from standard RAID levels and are usually developed by companies or organizations primarily for proprietary use.

RAID 7. This is a nonstandard RAID level based on RAID 3 and RAID 4 that adds caching. It includes a real-time embedded OS as a controller, caching via a high-speed bus and other characteristics of a standalone computer.

Adaptive RAID. This variation enables the RAID controller to decide how to store the parity on disks. It chooses between RAID 3 and RAID 5. The choice depends on what RAID set type performs better with the type of data being written to the disks.

Linux MD RAID 10. This level, provided by the Linux kernel, supports the creation of nested and nonstandard RAID arrays. Linux software RAID can also support the creation of standard RAID 0, RAID 1, RAID 4, RAID 5 and RAID 6 configurations.

Hardware RAID vs. software RAID

As with RAID controllers, RAID is implemented through either hardware or software. Hardware-based RAID supports different RAID configurations and is especially well suited for RAID 5 and 6. Configuration for hardware RAID 1 is good for supporting the boot and application drive process, while hardware RAID 5 is appropriate for large storage arrays. Both hardware RAID 5 and 6 are well suited for performance.

Hardware-based RAID requires a dedicated controller to be installed in the server. RAID controllers in hardware are configured through a card BIOS or option read-only memory either before or after the OS is booted. RAID controller manufacturers also typically provide proprietary software for their supported OSes.

Software-based RAID is provided by several modern OSes. It is implemented in a number of ways:

• As a component of the file system.
• As a layer that abstracts devices as a single virtual device.
• As a layer that sits above any file system.

This method of RAID uses some of the system's computing power to manage a software-based RAID configuration. As an example, Windows supports software RAID 0, 1 and 5, while Apple's macOS supports RAID 0, 1 and 10.

Benefits of RAID

Advantages of RAID include the following:

• There is improved cost-effectiveness because lower-priced drives are used in large numbers.
• Using multiple drives enables RAID to improve the performance of a single drive.
• There can be increased computer speed and reliability after a crash, depending on the configuration.
• Reads/writes can be performed faster than with a single drive with RAID 0 because a file system is distributed across drives that work together on the same file.
• There is increased availability and resiliency with RAID 5 because, with mirroring, two drives can contain the same data, ensuring one continues to work if the other fails.

Downsides of using RAID

RAID does have its limitations, however, including the following:

• Nested RAID levels are more expensive to implement than traditional RAID levels because they require more drives.
• The cost per gigabyte for storage devices is higher for nested RAID because many of the drives are used for redundancy.
• When a drive fails, the probability that another drive in the array will also soon fail rises, an event that would likely result in data loss. This is because all the drives in a RAID array typically are installed at the same time, so all the drives are subject to the same amount of wear.
• Some RAID levels, such as RAID 1 and 5, can only sustain a single drive failure.
• RAID arrays and the data they hold are vulnerable until a failed drive is replaced and the new disk is populated with data.
• Because drives have much greater capacity now than when RAID was first implemented, it takes a lot longer to rebuild failed drives.
• If a drive failure occurs, there is a chance the remaining drives may contain bad sectors or unreadable data, which may make it impossible to rebuild the array fully.

However, nested RAID levels address these problems by providing a greater degree of redundancy, significantly decreasing the chances of an array-level failure due to simultaneous disk failures.

What are RAID use cases?

Instances where it is useful to have a RAID setup include these requirements:

• A large amount of data needs to be restored. If a drive fails and data is lost, that data can be restored relatively quickly because it is also stored on other drives.
• Uptime and availability are important business factors. If data needs to be restored, it can be done quickly with little downtime or disruption to business operations.
• Working with large files. RAID provides speed and reliability when working with large files.
• Need to reduce strain on physical hardware and increase overall performance. As an example, a hardware RAID card can include additional memory to be used as a cache rather than relying on server resources.
• Having I/O disk issues. RAID provides additional throughput by reading/writing data from multiple drives instead of needing to wait for one drive to perform tasks.
• Cost as a factor. The cost of a RAID array is lower than it was in the past, and lower-priced drives are used in large numbers, providing overall cost savings.

History of RAID

The term RAID was coined in 1987 by David A. Patterson, Randy H. Katz and Garth Gibson. In their 1988 technical report, "A Case for Redundant Arrays of Inexpensive Disks (RAID)," the three argued that an array of inexpensive drives could beat the performance of the top expensive disk drives available at that time. By using redundancy, a RAID array could be more reliable than any one disk drive.

While this report was the first to put a name to the concept, others were already discussing the use of redundant disks. IBM's Norman Ken Ouchi filed a patent in 1977 for the technology, which was later named RAID 4. In 1983, Digital Equipment Corp. shipped the drives that would become RAID 1, and in 1986, another IBM patent was filed for what would become RAID 5. Patterson, Katz and Gibson also looked at what was being done by companies such as Tandem Computers, Thinking Machines and Maxtor to define their RAID taxonomies.

While the levels of RAID listed in the 1988 report essentially put names to technologies that were already in use, creating common terminology for the concept helped stimulate the data storage market to develop more RAID array products.

The term inexpensive in the acronym was soon replaced with independent by industry vendors due to the implication of low cost.

RAID manufacturers

Virtually all storage system vendors provide one or more RAID levels in their array products. Also, a number of vendors manufacture or resell RAID technology and systems, including the following:

• Asus.
• Broadcom.
• Dell EMC.
• EUROstor.
• Gigabyte Technology.
• JWIPC Technology Co., Ltd.
• Supermicro.
• Synology.

The future of RAID

RAID is not quite dead, but many analysts say the technology has become increasingly obsolete in recent years. Alternatives such as erasure coding offer better data protection -- albeit, at a higher price -- and have been developed with the intention of addressing the weaknesses of RAID technology. As drive capacity increases, the chances for error correction with a RAID array also increase. And drive capacities are consistently growing, putting a greater strain on RAID technologies and their abilities to recover in a timely manner.

The rise of SSDs is also seen as alleviating the need for RAID. SSDs have no moving parts and do not fail as often as hard disk drives. SSD arrays often use techniques such as wear leveling instead of relying on RAID for data protection. Modern SSDs are fast enough that modern servers may not need the slight performance boost that RAID offers. However, RAID still may be currently used to prevent data loss.

Hyperconverged infrastructure may also remove the need for RAID by using redundant servers instead of redundant drives.

Despite the above, the attributes of future RAID implementations ensure they will be an important choice for secure data storage:

• Support for higher-capacity disks. RAID technology will be able to support increasingly larger-capacity disk drives, both HDDs and SSDs.
• Better erasure coding. Systems will have more efficient erasure coding capabilities to improve data protection and fault tolerance.
• Advanced RAID controllers. New controllers will support technologies like non-volatile memory express and newer interfaces to improve performance and redundancy.
• Artificial intelligence. AI will increasingly be part of RAID technology to help improve performance, automate various functions, and improve security and scalability.
• Energy management. Using AI technology, RAID systems will be more energy-efficient and sustainable.

RAID safeguards data while enhancing storage performance and availability, though it can be perplexing. Learn about the various RAID levels, their advantages and disadvantages, and the scenarios in which they are most effective.