When attempting to ensure the security and safety of your data, you may run into the term RAID. RAID is a technology employed to store data that you will often see when researching servers or backup options.
What is RAID?
RAID, or Redundant Array of Independent Disks, is used to build redundancy for your data. Essentially, the purpose of RAID is to distribute your data across locations in a duplicated or distributed fashion so that if one location experiences a failure or loss, your data can be recovered. RAID utilizes multiple physical drives for an array that the computer uses as a single logical drive.
Why do I need RAID?
RAID is essential for computers, but particularly for servers. Servers host an intense amount of data and are expected to provide as close to 100% uptime as possible. This can be an arduous task without redundancy, especially for traditional dedicated servers. Redundancy ensures if there is an issue with one drive on your server, that your server can still have access to data and continued uptime despite this issue.
RAID works in three ways: stripping, mirroring, and parity. Stripping is when data is split up among various disks within the RAID. This ensures data isn’t consolidated into one single location. Mirroring places the same data on multiple disks. This ensures that if one disk is compromised, the exact same data still exists somewhere else on another disk in the raid. Finally, parity is a performance of calculations to determine what data is missing and places composites of it on another disk in the RAID. This ensures that calculations can be quickly made to reconstruct the missing data from other pieces of data within the RAID if a disk fails.
What are the different types of RAID?
RAID 0
RAID 0 is the simplest but least ideal type of RAID. RAID 0 uses striping, splitting data up between multiple disks. The benefit of this is that your data isn’t condensed into one location on one drive, so it is better than using no RAID. It is also very fast to access these files since it uses less storage than other options and isn’t as complex. However, this option doesn’t truly offer redundancy, as there are no copies of data; they are only striped amongst disks.
With RAID 0, if one disk fails, there is still a loss of data that can’t be recovered within the RAID; however, not all data is lost when one disk fails, making it a better option than not utilizing RAID.
RAID 1
RAID 1 is another common option. RAID 1 uses mirroring, identically mirroring or replicating data between multiple disks. If disk A has files XYZ on it, then disk b, c, and so on will also have files XYZ. This is beneficial because any drive can fail within the array, and no data will be lost because every disk holds an identical copy of every other disk in the system.
RAID 1 makes a recovery easy because you only need access to one disk for recovery. However, RAID 1 has some disadvantages. This system requires a lot of storage. It at least doubles the amount of storage you need even though you are still only storing the original amount of data. If you have 1 TB of original data and want to use a RAID 1 system with two disks, you need 2 TB of storage. Even though there’s only 1 TB of usable data, the other 1 TB exists only for redundancy. This can make RAID 1 slow as well because it is so storage intensive.
RAID 5
RAID 5 is the most complex RAID option of these 3. RAID 5 utilizes parity calculations to determine extra data that is stored with the original data to build redundancy. RAID 5 requires at least three disks because it allows one disk worth of data to the data calculated with parity. With this system, the original data is split amongst the disks in the RAID. Then parity data is calculated and placed within the disks in the RAID alongside the original data.
RAID 5 is beneficial because it provides redundancy and does not take up as much unused storage as RAID 1. Additionally, it can provide extra security because parity files can be used to verify files. However, because it relies on calculations, it can occasionally slow things down. Additionally, the parity data does take up storage, and so it is somewhat storage intensive.
With RAID 5, if one drive fails, it can always be restored by using parity to determine what is missing and verifying that the data stored is correct.
Other options
These are the most common types of RAID; however, there are some other popular variations.
RAID options can be combined to form hybrid RAID arrangements, such as RAID 10, which is RAID 1 combined with RAID 0. RAID 10 stripes data, like in RAID 0, and splits pieces of data up amongst the disks in the array. It then mirrors these striped disks. This requires more disks at a minimum but takes up less storage than RAID 1 because the entire data set is not mirrored, only portions. This also offers redundancy, where RAID 0 does not. However, there is still no parity within the system, and it only allows for the failure of one drive for restoration.
Another popular RAID option is RAID 6. RAID 6 is very similar to RAID 5 in that it uses the same parity calculations to build redundancy across disks. However, it also makes use of an additional, more complex parity calculation. The primary advantage of this over RAID 5 is that it offers higher levels of redundancy. With RAID 6, parity data makes up at least two disks worth of data, rather than one in RAID 5. Because of this, two disks may fail simultaneously and still be restored with RAID 6, whereas with RAID 5, only one drive can fail at a time for restoration.
