What is RAID?

Share

Today, we’re going to tell you all about RAID. We’ll explain exactly what it is, take a look at the different types available and also touch on the benefits of each one.

RAID stands for Redundant Array of Independent Disks and is essentially a way of joining several disks together to achieve better performance or redundancy.

It was originally conceived as a way to create a more powerful system for less money. With older technologies costing enormous sums of money back in the day, people saw RAID as a much more cost-effective way of achieving the desired disk performance.

These days, there are many different types of RAID, called levels, and each one performs differently and has different advantages and drawbacks.

The main advantages of using arrays of disks instead of independent disks are:

  • Greater capacity.
  • Higher failure tolerance.
  • Faster read and write capacity.

Even though there are many different types of RAID, they all work according to the same basic principle; they all either replicate or distribute data to provide better performance (read or write speeds) or greater redundancy to protect against disk failures.

Furthermore, the operating system and its applications see this configuration of multiple disks as a single logical volume, making it much simpler to work with.

 

What Types of RAID Are There?

As we just mentioned, there are multiple RAID levels out there, and each one has a specific disk configuration. However, we can also divide these RAID levels into three distinct categories, which are:

  • Type 1: Standard RAID
  • Type 2: Nested or Hybrid RAID
  • Type 3: Non-standard RAID

In this article, we’re going to take a look at each of these categories and explain some of the levels that belong to each one. We’re not going to look at every RAID level available because, frankly, there are just too many to cover in a single article!

 

Standard RAID

There are lots of standard RAID configurations, such as RAID 1, 2, 3, 4, 5 and 6, as well as several variations of each one. Most businesses these days use RAID 0, RAID 1 or RAID 5. We’ll also take a look at RAID 3 as you’ll need to understand it for when we touch on RAID 30. But the other levels aren’t so popular, so we’re going to skip over them today.

RAID 0

RAID 0 is the most basic level. It doesn’t offer any redundancy but it does share data across different disks. As a result, this configuration is often called a Spanned Volume or a Striped Volume.

With this level, the data is written evenly across the available disks, writing first on one disk, then on the next and so on. What does this achieve? Well, clearly not redundancy. If one disk fails, there’s a good chance that the data will be lost. However, what it does do is achieve faster read and write speeds as it can access multiple disks simultaneously in parallel.

Diagram of a RAID 0 configuration
Diagram of a RAID 0 configuration

With RAID 0, the total available disk space is equal to the sum of the two disks. Well… sort of. If you have two 1 TB disks, the total size of the RAID will be 2 TB. However, the disks need to be the same size. So, if you have a 1 TB disk and a 2 TB disk, the total size of the RAID will still be 2TB, leaving 1TB on the larger disk (33% of the total space) completely unused.

For RAID 0, you need at least two disks so that data can be evenly shared between them.

 

RAID 1

The next simplest level is RAID 1, which includes a certain amount of redundancy. It consists of at least two disks which all receive the same data. Basically, whatever is written to the first disk is copied onto the other disks. This is also known as Disk Mirroring.

With this RAID level, it’s important to remember two key drawbacks:

  • It will waste disk space because the same data will be written at least twice for redundancy.
  • It will also waste disk space if you have disks that are different sizes. As this configuration relies on making identical copies of the original data, if you have disks of a different size, the space used will only be as much as the space available on the smallest disk. So, if you have a 1 TB disk and a 2 TB disk, the total size of the RAID will be just 1 TB.
Diagram of a RAID 1 configuration
Diagram of a RAID 1 configuration

For RAID 1, you need at least two disks since you need to be able to copy data on at least one other disk for redundancy.

NOTE: If a disk failure is detected and the disk needs to be changed, the RAID will be vulnerable while the data is being rebuilt and recovered. It’s very important to keep this period of time as short as possible, and the time to do this will grow proportionally with the size of the disk because more data will need to be rebuilt.

 

RAID 3

RAID 3 is not as popular as RAID 0, 1 or 5, but we’re going to take a look at it anyway because it is used to create RAID 3+0, which does enjoy a certain level of popularity.

This configuration shares data evenly across all available disks, with one disk dedicated to parity checks. It’s a bit like a souped-up version of RAID 0.

As you’ve probably already guessed, using multiple disks means much higher data transfer rates compared to a single disk.

For RAID 3, you need at least three disks as you need to share data across at least two disks and have another disk for parity checks.

Diagram of a RAID 3 configuration
Diagram of a RAID 3 configuration

 

RAID 5

The next most commonly used configuration is RAID 5, which is also often called “Striping with distributed parity”.

With this configuration, data is evenly written across the available disks and there is a block dedicated to parity to allow data to be rebuilt in the event of a disaster.

At the start of this article, we mentioned that one of the reasons for RAID’s popularity was the fact that it can provide redundancy at a very low cost. And this could certainly be said about this particular configuration.

Diagram of a RAID 5 configuration
Diagram of a RAID 5 configuration

For RAID 5, you will need at least three disks because you need to be able to share data across multiple disks whilst maintaining a parity disk as well.

However, one of the disks will always be dedicated to parity, so you actually have one less disk available. That’s why this is sometimes referred to as an “n-1” configuration. In other words, if you have four disks, as in the diagram above, one will be the parity disk and won’t be available for writing useful data.

This means that, if you have four 1 TB disks, you actually only have 3 TB of space available because of the parity disk.

With this configuration, if one disk fails, we can be reassured of not losing data. But if a second disk fails before the data has been recovered, it will be lost.

 

RAID 6

RAID 6 is similar to RAID 5 but it’s able to handle two disk failures. While RAID 5 has n-1 useful disks, RAID 6 has n-2. This is because, when one disk fails, another one takes its place. Basically, it’s kind of a RAID 5 + Spare configuration.

As you might have guessed, this means that you need at least four disks for RAID 6 since you need to have two disks for failure tolerance.

While all this redundancy might sound marvellous, the costs of having so many disks are obviously an issue here. Furthermore, with all the necessary parity checks, etc., the write speed for RAID 6 will be slower than RAID 5. However, for some people, this may well be a price worth paying for greater protection.

Diagram of a RAID 6 configuration
Diagram of a RAID 6 configuration

Those are the most important ‘Standard RAID’ configurations. Now, it’s time to look at some Nested or ‘Hybrid’ RAID levels.

 

Nested RAID

 

RAID 0+1

This RAID level is fairly straightforward. It consists of creating a RAID 0 configuration and then mirroring it with another RAID 0 to create a RAID 1. In other words, the characteristics of the two levels have been combined.

For RAID 0+1, you need at least four disks, as you need to have two RAID 0 configurations.

This configuration can handle one disk failure but not a second.

Diagram of a RAID 0+1 configuration
Diagram of a RAID 0+1 configuration

 

RAID 1+0

As with RAID 0+1, RAID 1+0 (also known as RAID 10) consists of combining the two levels together. In this case, data is written evenly across each pair of disks, and each pair consists of two mirrored volumes.

As a result, RAID 1+0 requires at least four disks. So, it costs the same as RAID 6 but can tolerate up to two disk failures.

This RAID level offers the transfer rates of RAID 0 with the failure tolerance of RAID 1. This makes it a good choice for speed and recovery capacity in the event of a disaster. It’s often used for database servers.

Diagram of a RAID 1+0 configuration
Diagram of a RAID 1+0 configuration

 

RAID 30:

RAID 30, also known as RAID 3+0, is a combination of RAID 3 and RAID 0. It has the advantage of higher transfer rates as well as greater resilience and reliability. However, the installation costs are quite high since it needs a large number of disks.

With RAID 30, the system can handle a disk failure in each RAID 3 set. As always, while data is being recovered, the system will be vulnerable, but the chances of a disaster that would compromise the entire configuration are fairly low.

Diagram of a RAID 30 configuration
Diagram of a RAID 30 configuration

RAID 30 requires at least six disks, as you need to have at least two sets of three disks, consisting of two write disks and one parity disk.

 

RAID 50

This configuration follows a similar pattern to RAID 30 in that it consists of a RAID 0 configuration of multiple RAID 5 arrangements.

This system is particularly robust because it can handle a disk failure in each RAID 5 set without affecting system performance. RAID 50 also performs much better than RAID 5, particularly with regard to write speeds. As a result, it can be very effective for applications that intensely use read data in non-sequential positions.

Diagram of a RAID 50 configuration
Diagram of a RAID 50 configuration

RAID 50 requires 12 disks as you need to create at least three RAID 5 sets to make up the overarching RAID 0 structure.

 

Non-standard RAID

The RAID levels that we’ve looked at so far are standard public configurations that you can use on pretty much any system with no issues.

Besides this, there are also a number of Non-standard RAIDs that have been privately developed by one or more companies.

We’re not going into more detail on these today, but some of the following are common:

  • RAID Z
  • RAID 50EE
  • RAID Matrix
  • RAID S
  • RAID 7

And that concludes our introduction to RAID!

 

Conclusion

Today, we’ve talked about what RAID is, looked at some of the most common configurations (levels) and also discussed what they are useful for.

One thing we haven’t talked about is how to set up a RAID on your Windows Server. That’s because we already have some tutorials on our blog dedicated to this very topic, such as:

So, if you’re looking for further information or need any help with your sysadmin, cloud, etc., we recommend that you check out our blog for more useful tutorials and articles.

Thanks for choosing Jotelulu!

Category:Cloud and Systems

Other posts that may interest you

20 de May de 2024
If you’ve ever wondered how to respond to an IT security incident, this article will hopefully provide all the information
13 de May de 2024
In this short article, we’ll try to uncover the real cost of not investing in IT security. Each year,
5 de May de 2024
Read on to learn about the most common issues encountered when working with backups. These days, backups are a

Fill out the form and one of our Sales team will contact you soon.

growth@jotelulu.com  |  jotelulu.com 

You can unsubscribe from these communications at any time. For more information,  check our Privacy Policy.