What Types of RAIDs Are There?
When it comes to computer systems and servers, you have a lot of data to manage. It’s normal to want robust backups so that you never lose important information. At the same time, backups are costly, and they can slow down server systems.
In order to find the best options for various goals, you can manage storage disks in a number of different ways, and one of the most trusted and popular options is to implement RAID.
What Is a RAID?
RAID stands for Redundant Array of Independent Disks, and it’s a technique that allows multiple storage disks to work together for a specific aim. In a RAID configuration, each individual disk is called a member, and there are three primary goals of using RAID configurations:
- Fault tolerance is the concept of creating ways to recover or maintain a system when part of it fails.
- Performance is a measure of speed and efficiency across a system.
- Capacity is the measure of how much data the configuration can store.
Types of RAIDs
Below, you will see five different types of RAID configurations. Each is known as a RAID level. Each design has different goals in mind in terms of fault tolerance, performance, and capacity.
RAID 0
RAID 0 is also called disk striping. This is a technique where any bit of data is written across multiple disks. The disks work to split the task of reading and writing, meaning that the data shared between them is not redundant.
As an example, the data for a single document file might be split between two disks. This is done so that both disks are engaged to read or write that file, and in doing this, they split the task and get it done faster. After all, each disk is only reading and writing half of the data involved in the process.
This technique is popular with server systems because it increases overall server speed and performance. The thing to remember is that it comes at the cost of fault tolerance. When drives split data in this way, losing a single drive means that all of the data in the system is compromised.
RAID 1
RAID 1 is more or less the opposite of RAID 0. With level 1, data is copied exactly from one disk to another. This process creates a perfect replica of all data on a disk — often called a “mirror.”
RAID 1 has reasonably high levels of fault tolerance. You can lose a disk, and the system is still fine because the mirrored disk is ready to go. As long as any faulty disk is replaced and mirrored in a timely manner, a system will very rarely lose data (which happens in the case that both disks fail simultaneously).
RAID 5
RAID 5 is probably the most common level, and it offers very high levels of fault tolerance. With this technique, any data is split across three or more disks. This is done in a way that each disk can be completely rewritten by using data from the others. The technique is called parity.
When RAID 5 is used, a system can lose a single disk and continue to function normally. On top of that, the faulty disk can be hot-swapped, allowing for replacement without taking the system offline. The new disk will automatically replace the faulty disk, and the system continues without a problem.
RAID 6
RAID 6 takes the idea of RAID 5 and pushes it even further. A RAID 5 system can survive the loss of any one disk. A RAID 6 system can survive the loss of any two. This is because the parity technique is more robust and can rebuild lost data more effectively.
Currently, RAID 6 offers the best fault tolerance of any RAID level.
RAID 10
Lastly, RAID 10 combines the ideas of 1 and 0. RAID 10 uses the striping technique of 0, making systems faster by engaging multiple drives for any reading or writing activities.
At the same time, it also mirrors every drive using the same techniques as RAID 1. That means that RAID 10 improves server performance while providing reasonable fault tolerance. After all, there are two copies of all information on the system.
The drawback to this method is that it requires twice as many disks as any other system, making it rather costly to implement.
Additional Learning Center Resources