What is RAID and why does it matter?

Many clients often wonder why they can only use just half of the available storage (SSD or HDD). The answer is the RAID array. It is a method connecting storage disks, usually SSDs or HDDs, with the purpose of redundancy and/or speed and is a useful tool. Would you like to know more about it?

What does RAID mean?

RAID means Redundant Array of Independent Disks. It is a replacement for the standard option of just one storage device and replacing it with an array of more storage devices – multiple SSDs or HDDs. This way you can get a better performance and/or hardware data protection with a system of more than one drive. 

It is a technology that has been here for a while, more precisely from the 70s. The first idea was to combine simple drives and make them work together. The result is higher capacity and faster configuration. A valuable extra is the failover protection. Even if one of the drives fails, you can continue without losing data (depending on the RAID configuration). Since then, there have been many applications of the RAID. RAID 0, RAID 1, RAID 5, RAID 6, RAID 10 and more.

Even 100% software RAID exists, but as you can imagine, the hardware RAID is the preferred method for data protection. 

Why should you consider a RAID array?

You should consider a RAID array because every hardware, no matter if it is an SSD (solid-state drive), HDD (hard disk drive), flash memory of a different sort, or another will eventually break down. It is inevitable, after a particular time. In the case of an HDD, it is hardware with mechanical parts that have a bigger risk of failure. If it is an SSD, it has read and write limits, which shows for how long it could be used. 

When should I use RAID?

You should use a RAID array when you need to guarantee uptime. Also, most RAID arrays can keep you safe in case of hardware failure of an HDD or SSD, since there are data copies between the drives. In many cases, the change of a faulty drive can be performed by hot-swapping the problem.

RAID 0 (Striping)

RAID 0 merges all the available drives (SSD or HDD) into one. RAID 0 has a higher read and write speed thanks to the distribution of the data across the different drives. It has far better performance than others, and you can use all the space, but it lacks redundancy. It doesn’t make a copy of your data across the drives and if it fails you will probably lose data. A simple RAID 0 could be very dangerous for your data. RAID 0 was used more often in the past when the HDDs were slow, but now with the increased popularity of the SSD, it is no longer the preferred RAID solution. 

Taking into consideration the average HDD failure rate of 2.5% per year, if you have three disks this number will jump to 15.625%.

This setup is all about speed.

RAID 1 (Mirroring)

RAID 1 is mirrors the data between the drives. Usually, in the RAID 1 array, users have identical drives for best results. This is good failover protection, if one of the drives gets damaged, the rest will still have a complete copy of your data and won’t affect you negatively. You can change a damaged drive without experiencing downtime and without data loss, which is very beneficial. 

It has a faster reading speed but the same writing one. Also, the duplication of data will take extra space. That is why you will be able to use only half of the storage capacity. This is the case with many of our Cloud servers and Dedicated servers. We value your data integrity, and we protect it with a RAID 1 array. 

RAID 5 (Striping disks with distributed parity)

In the case of RAID 5, we have striping drives with distributed parity. It has better speed and protection. The best from both previous RAIDs. The data is spread across multiple drives, and there is a trick; parity data. This one is spread across the drives and has the information to recuperate the data of any of the failed drives. But watch out, you can only afford to lose just one of the multiple drives. If you lose two or more, you won’t be able to recuperate the data and it will be permanently lost. 

The RAID 5 configuration requires at least three disks, where one of the drives is reserved to rebuild the data on the array if it dies. If you have five drives, you will have one to rebuild the data and have the capacity of four drives for the array. You can read data from it very quickly. For such a RAID array you would need an array controller, to be able to write on the array with a decent speed. In case of a drive failure, it is time-consuming to rebuild the data. 

RAID 6 (Double parity)

Similar to RAID 5, however RAID 6 can survive a failure of up to two drives at the same time. RAID 6 uses double parity, for twice better fault tolerance in comparison with RAID 5. You can suffer two drives fail at the same time, something that is far less common than a single drive failure and still experience no data loss. 

The disadvantage of the RAID 6 array is that you will use half of the storage space for parity. Also restoring the data, when you put a new drive in the place of a broken drive will take a long time. Just like the previous configuration, it is not bullet-proof, and losing three or more drives will lead to complete data loss. The minimum disk configuration of RAID 6 is four drives, and you will need an array controller for it. 

RAID 10 (Mirroring and Striping) or RAID 1 + RAID 0

RAID 10 combines the RAID 0 and RAID 1 arrays. It is an array from two RAIDs which makes it superior when it comes to performance. It inherits the benefits of both RAID 0 (Mirroring) and RAID 1 (Striping). Increase speed, both read and write, will depend on the number of drives that the array contains. In many cases, you can use multiple economical disks, connect them to work together, and get better performance than a single more expensive drive. 

And data mirroring will help keep your data safe. For implementing RAID 10, you will need at least four drives and a RAID disk controller, which makes it a bit more complicated to implement. What we see as a downside of this configuration is that you can’t use the full storage capacity because of the mirroring of data, and you’re only able to use half of the capacity.

What type of RAID should I use?

Based on your needs and priority you have several choices. Find the best RAID for your server by thinking about one of these key aspects. 

For speed and simplicity

Go for RAID 0. It is a high-risk scenario, but if you are using only HDDs, you can get decent performance. 

Economic and simple

RAID 1 is your choice. It is easy to configure, and it will provide sufficient failure tolerance for many businesses. 

For extra data protection

You can choose one of the following – RAID 5, RAID 6, or RAID 10. All of them are more expensive, but you can enjoy a higher level of data security and in different cases higher performance. 

What does RAID not do?

• RAID does not automatically guarantee 100% uptime. Depending on the problem, it could take time to restore the data, or it could even be impossible. 
• RAID is not a backup. When you delete a file from a drive, part of a RAID array, this file is gone forever. If you want to learn more about backups see “7 backup mistakes to avoid”. 
• RAID does not protect you from human errors. Bad configurations are always possible, and people can often make mistakes with big consequences. 
• RAID might limit the ability to expand the storage. In some cases, you won’t be able to increase the capacity with a few extra drives. You might need to configure the whole system from the beginning. 
• RAID does not protect from data corruption. The RAID configuration help in case of hardware failure, but if the data gets corrupted, there is no protection. The RAID will continue to replicate the data even corrupted. 
• RAID is not protected from catastrophic events like earthquakes, floods, etc. If the server is physically damaged, the data will be gone too. 

Hardware vs Software RAID

Hardware RAID

The Hardware RAID is created with multiple physical drives – SSDs or HDDs. Depending on the configuration it could require a RAID controller. The controller is managing the data, which makes it easier on the CPU of the server. 

Software RAID

The Software RAID is just software, that runs on a server, that organizes the connected SSDs and HDDs. The storage devices are connected directly to the server and are managed by the software RAID. This makes it harder on the CPU of the server. It is an economic option, but the OS could have different limits. 

RAID vs AHCI

AHCI is Advanced Host Controller Interface, and it is an operation mode for the SATA interface. The tech allows multiple drives to work on a single system, and hot-swapping of drives too. But it does not provide redundancy like the RAID array.  

RAID vs NAS

NAS is network-attached storage. It is basically a small data server, that is connected to a network and is available on that network. It can contain multiple disks, just like RAID. It can be used for backup too. The difference is that it is a separate device, while RAID is a disk configuration. A NAS can use a RAID array if it has multiple drives inside. 

RAID vs ZFS

While both RAID and ZFS uses distributed data across drives, they have one big difference. The ZFS is not only a disk array, but it also includes a file system layer. It can provide advanced features like compression, and a journal system, where you can see the journal entries, which the RAID can’t do. 

Conclusion

Now that you know what RAID is, what it’s not, and why is it so important to have storage devices in a RAID configuration – always check that your service provider uses it! Find the best setup for you and increase the speed and redundancy – adding one extra layer of protection is always worth it!

More of the same author:

How to avoid Internet scams