You’ve probably come across the term RAID if you’ve been looking into PC storage. This article will explain what RAID is, and how different RAID levels affect overall performance.
What is RAID?
RAID stands for either ‘Redundant Array of Inexpensive Disks’ or ‘Redundant Array of Independent Drives.’ A RAID system comprises two or more drives that work together to offer you multiple benefits such as better performance, increased redundancy for data, and improved data parity. RAID is typically used with hard drives, although it can be used with SSDs too.
The software that performs the RAID functionality itself is usually just a simple driver, but it can also sometimes be located on a separate controller card, called a hardware RAID controller. Some versions of Windows and the Mac OS allow you to use either option. Hardware RAID controllers are more expensive but offer better performance as well.
RAID Levels
The number of drives used depends on the RAID level, discussed next. Each level is meant for a specific situation, but the levels themselves aren’t standardized as such, so companies can sometimes come up with their own unique implementations. Talking about these levels in more generic terms, the pros and cons of each are as follows.
What is RAID 0?
RAID level 0 is the most basic RAID system available. It uses at least 2 disks at the same time (this is called striping) to offer superior performance, which can further be enhanced by using multiple controllers (ideally one for each disk).
Pros and Cons
RAID 0 offers efficient performance because there is no overhead caused by parity controls, and also because all storage capacity is utilized. The technology is simple and easy to implement with different types of storage interfaces, and introducing new drives is a good way to increase performance without diminishing returns.
That being said, RAID 0 isn’t fault-tolerant. This means that if one drive fails, all the data in the RAID array will be lost. The more drives you have, the greater is the risk of data loss. For this reason, RAID 0 should not be used for OS and such. It is best for non-critical data that just needs to be read/written at high speeds.
What is RAID 1?
RAID level 1 uses mirroring, which means that data is stored twice; once by being written to the set of data drives in the system, and then by being stored to a mirror drive for recovery by the controller. A RAID 1 array also requires a minimum of two drives.
Pros and Cons
RAID 1 offers excellent read speed, even better than RAID 0 sometimes, but it lacks severely in write speed as compared to RAID 0. Thanks to mirroring, there is less data loss risk because if a drive fails, its data can be recovered from the replacement drive.
Effective storage capacity is reduced to half of the total device capacity though because all data has to be written twice. RAID 1 is ideal for mission-critical storage on small servers, such as accounting systems.
What is RAID 5?
The most commonly used RAID level is RAID 5. It requires a minimum of 3 drives but can support a maximum of 16 drives. It uses striping parity, which means that data is first striped across all the drives, and then a parity of all the block data is written.
This parity is not written on a single drive; rather it is spread across all the drives in the system, which is why RAID 5 can withstand a single failing drive without losing any data.
Pros and Cons
Read speed is higher while write speed is reduced due to the parity that needs to be calculated. But because of this superior read speed, extra cache memory is often used with RAID 5 systems. Single drive failure will never be a problem with RAID 5 because the storage controller immediately rebuilds the lost data onto a new drive.
Drive failures still affect throughput, although not too much. RAID 5 can be too complex for most users though. The reason for RAID 5’s popularity is that, overall, it offers efficient storage with good security and sufficient performance for users with limited data drives.
What is RAID 6?
RAID level 6 improves on RAID 5 in that it uses two drives for writing parity data. This means that it can withstand two drives failing simultaneously. In RAID 5, the first failing drive can be dealt with, but if a second one fails you will lose all your data. RAID 6 requires a minimum of four drives.
Pros and Cons
While RAID 6 is more secure than RAID 5, the chances of two drives failing at the exact same time are low, so this advantage may not work out practically. RAID 6 is much more complex than RAID 5 though, so rebuilding an array after a drive fails can take a long time.
What is RAID 10?
RAID level 10 is sometimes written as RAID 1+0 because it solves RAID 0 vs RAID 1 problem by combining the advantages of both into a single system. Hence, RAID 10 is a hybrid configuration. It mirrors data onto secondary drives and uses striping across each set of drives. This speeds up data transfers and provides greater security.
Pros and Cons
Rebuilding after drive failure is fastest in a RAID 10 configuration. It can be as quick as 30 minutes per TB of data. But since half the storage capacity goes to mirroring, the redundancy offered by RAID 10 becomes quite expensive, especially when compared to RAID 5 or RAID 6 arrays.
Conclusion
Other RAID levels such as 2, 3, 4, and 7 do exist, but these are not very common and not very efficient. Questions like RAID 5 vs RAID 10 and RAID 6 vs RAID 10 don’t matter because the RAID level that suits you depends on your data needs. For a limited budget, RAID 5 and RAID 6 are ideal for getting the most out of your disks. RAID 10 is more suitable for big data operations.
But keep in mind that a RAID system is not a substitute for backing up data. Keep a back-up of your RAID system data always, because situations like a (hypothetical) fire in your workplace, or, less drastically, a power spike that causes all your drives to fail simultaneously, could render your RAID system useless. Even user error could result in permanently lost data, Backed up data can be stored off-site and will require minimum effort.