To get more from a drive, it's possible to use some form of RAID to group it with others and create larger storage volumes to save data. It's widely used on servers such as NAS and even inside data centers. You may have come across the term RAID on your travels within the computing sphere, but what exactly does this acronym mean, and why is it important to consider using it for larger storage or even your PC?

What is RAID?

Hard drives aren't the quickest forms of storage available, with SSDs offering speeds of a magnitude of 10, if not higher. To overcome this and to make using a group of drives easier for storing lots of data, a redundant array of inexpensive disks (RAID) can be used. This process essentially groups the included drives to act as a single storage medium. Think of it as combining four 4TB hard drives into one single 16TB hard drive. This can then be read and used by the OS.

As well as acting as a single drive for easier data management, RAID can also help with data protection by offering a way to shield the storage pool against loss through drive failures. This usually requires at least one drive to be reserved for storing all the necessary recovery data but does allow the RAID to be rebuilt and data kept with the replacement of a dead drive. In the case of using RAID, our four 4TB drives may only provide 12TB of usable space with one 4TB reserved by the system for recovery.

Using a RAID is a great way to save money as more inexpensive drives can be used in larger numbers. Using more than one drive can also improve overall system performance when storing or reading data, at least compared to a single drive. RAID is an important part of any deployment where data integrity and availability are a priority.

What are the different types of RAID?

There are many types of RAID for use, though the most popular seem to be RAID 0, RAID 1, RAID 5, and RAID 10. As well as the standard RAID types, there may be some custom options depending on what's being stored and which platform is being used. For instance, Unraid is an OS that stores data on a single drive alone, yet provides some protection against drive failure. Then there are NAS vendors with their hybrid RAID solutions such as Synology's SHR.

RAID 0

RAID 0 essentially splits (referred to as "stripes") data into smaller chunks and stores them on two or more disks. RAID 0 will provide among the best read and write speeds for the disk array, but should a drive fail inside a RAID 0, you will lose all data and need to replace the drive and apply a backup. You will require at least two drives for this RAID type, though you will see faster speeds compared to a single drive.

RAID 1

RAID 1 duplicates data and stores it on two separate drives. The total capacity is reduced by half to accommodate this, but that does mean you'll have full recovery available should a drive fail. Write speeds will be slower than other RAIDs due to the duplication of data. This RAID type does provide the best data redundancy protection, however.

RAID 3

RAID 3 reserves one drive to store parity information for the array. It takes a significant performance hit as a single operation addresses all included drives, making it the better choice for single-user deployments.

RAID 5

If RAID 0 and RAID 1 don't fully satisfy your needs, RAID 5 (or "disk striping with parity") would be the way to go. This type splits data across multiple drives, as well as making local backups through parity data on all drives. This is the better RAID type with more than three drives available. Should a drive fail, you can quickly swap it out and have the RAID rebuild itself using the parity data stored on other drives.

RAID 10

RAID 10 combines RAID 1 and 0 for maximum failure tolerance without sacrificing transfer speeds. This RAID type groups at least four drives into sets of two, dividing data across these drive groups. The data is then duplicated onto the second drive within each group. This will be the most expensive RAID type requiring more disks to be purchased for duplication to take place within each set.

How to create a RAID

The process of creating an array of redundant disks depends on the hardware and OS you're using. Many systems provide the means to run a RAID through software, though RAID controllers are available that can manage an array of disks. Linux, macOS, and Windows all support certain RAID types, which can be managed through processor power and dedicated RAM. The OS will allow the creation of a storage pool using multiple drives, effectively creating a RAID.

It's only with server deployments that I recommend purchasing a RAID controller. To set up a RAID within Windows 11, you'll need to use the Settings app to create a storage pool. Ubuntu (and Debian) users can use the mdadm package to create a storage pool, and Disk Utility on macOS can be used for all things relating to RAID and storage.

FAQ

Q: When should you use a RAID?

I would recommend considering the use of a RAID when a larger amount of data needs to be stored on a system. As well as external backups, a RAID would be able to restore the data. This is also the case for larger files. Using a RAID can help lower the overall cost of deployment by splitting it into more than one drive, which can also improve performance due to using multiple drives for storing and reading data. Using a RAID inside your desktop PC to store games, the OS, and a few files is not going to beat a full system backup. RAID makes more sense for a NAS or server.

Q: Do you need a RAID for SSDs?

In short: no. I would advise against using a RAID with the deployment of solid-state drives (SSD). SSDs are considerably faster and more reliable, having no moving parts, and should last considerably longer than mechanical drives, especially if more data is read than written. Modern SSDs also have features to help combat wear and tear that make them a better option for long-term storage. However, if you are using more than one SSD and wish to have some degree of data redundancy, RAID can be used.