We explain what RAID 0, 1, 5, 6, 01, 10, 50, 100 and more are

There are different possibilities of combining hard drives according to the needs we have. This combination of hard drives can be a great performance or security improvement for stored data. Said combinations of hard drives is called RAID and there are many types and here we will explain them to you.

No products found.

What is a RAID

It is a typology of configuration of multiple hard drives that can work in a coordinated manner. The term RAID comes from the English acronym for: Redundant Array of Independent Disks.

Within RAID systems we find two large groups:

• Disk mirroring: RAID systems looking for redundancy in the event of a possible failure of one of the hard drives
• Disk stripping: RAID system that does not seek redundancy, but rather high data transfer speed

What is a RAID for

The Uses of this type of groupings are usually very varied. It depends a lot on the needs of each user, since we have configurations that improve data transfer speed and other data redundancy. Some RAID configurations combine these two elements.

Currently there are settings that are no longer used as much, such as RAID 0s that are intended for transfer speed. This type of configuration had its heyday with mechanical hard drives. Such drives had the disadvantage of offering low read and write speeds. What was done was to combine several of these units to improve performance.

Especially the RAID configurations serve to "protect" data on hard drives. Some configurations offer simple data copy and other configurations with parity. Depending on the needs, we can choose a type of configuration with data redundancy.

Types of RAID levels

There are different types of configurations depending on the needs you have. Here we will explain each of these configurations. One thing must be taken into account in all of them and that is that the hard drives must always be of the same capacity. If the capacity is different, the capacity of the smallest will always be taken, leaving the remaining capacity of the largest without use.

RAID level 0

We could say that it is the easier setup. The configuration RAID 0 does not offer data redundancy, but data distribution. All units under this configuration will function as a single unit. The number of units that can be integrated under this configuration is infinite, limited only by the controller.

[box type=»info» align=»» class=»» width=»»]It is also known as RAID Gaming, since it was very popular when only mechanical hard drives existed, allowing high read and write speeds.[/box]

An RAID 0 configuration not looking for data redundancy, but high data access speeds. When we write, the first bit is written to 'Disk 1', the second bit is written to 'Disk 2', the third bit is written to 'Disk 1' and so on. The reading process is the same, always starting with 'Disc 1'.

There is no parity or data redundancy in RAID 0 configurations.

The total speeds is not the sum of the speeds of the two discs, it is usually between 10-25% less than the sum of both speeds. If the two disks offer 100MB / s read, the sum will not be 200MB / s read, they will be between 180-150MB / s.

If any of the units that make up the set is damaged, all data will be irretrievably lost.

RAID level 1

It is commonly called mirror or "mirroring"«. Its use is quite widespread among home users looking for a simple and inexpensive setup. It is characterized by the information is written simultaneously to both disks. The operating system only sees one drive, even if we have two.

It is wrongly considered that the information is recorded on one disc and then copied to the other. What is done is to write simultaneously on both disks. This supposes a write speed reduction of approximately 25%. As well read on both discs simultaneously, but here the opposite phenomenon occurs, reading speed can be improved by up to 25%.

If a unit is damaged, we can replace it and request the copy of the data in the new unit. The problem with these systems is that data can be corrupted by not having parity. If both disks fail, the information will be irretrievably lost.

RAID level 2

Let's go now with a little used and quite complex configuration. What this configuration does is a distributed storage across multiple disks at bit level. To correct errors, it uses the Hamming code.

The controller for the RAID 2 syncs disks to rotate in the same angular orientation. This prevents them from generally being able to service multiple requests simultaneously.

Depending on the Hamming code, many disks could run in parallel in transferring data simultaneously. This allows the transfer rates to be very high.

As the data is divided at the bit level and not at the block level, the complexity increases. Actually A minimum of 39 disks would be required, of which 32 would be for storing individual bits and 7 for error correction.

RAID level 3

This configuration is currently little used due to complexity. Distribute data at the byte level instead of the block level, with a disc exclusively dedicated to parity.

For this system at least three disks are required, two for storage and one for parity. The data is stored by bytes on the first two disks that would act as a RAID 0. In the third disk a parity byte is created. So in case of losing a byte due to an 'x' error, we can recover it from the parity disk.

The advantage is that first two disks operate in RAID 0, greatly improving read and write speeds. As a drawback we have that it requires the activation of all disks simultaneously. This makes it difficult to make simultaneous requests.

RAID level 4

This configuration is quite similar to RAID 3, with the exception that here the disk is divided into blocks, not bytes. This allows each disk in the set to function independently of the rest.

For this type of RAID 4 configuration requires at least 3 disks, one of them will be the parity. The interesting thing about this system is that if the controller allows it, it can serve several read requests simultaneously.

Theoretically, this configuration also supports several write requests simultaneously. The problem is that single disk parity would create a bottleneck.

RAID level 5

Distributed disk system with parity widely used today, especially on NAS devices. Information is stored on hard drives in blocks. Each of the system disks has a parity block to ensure redundancy.

Redundancy in this system allows reconstruction of stored data in the event of a hard drive problem. The parity block is stored in a different unit than the different data blocks involved. It simply means that the parity block is stored on a different disk to ensure that the information can be restored.

For these systems require at least three hard drives to ensure data redundancy with parity. This system only allows one disk failure at a time. If you dadd two units simultaneously, we will lose the stored information. In the event of an error during the restore process, the information will also be corrupted and lost.

RAID level 6

This system is a advanced version of RAID 5 configuration. A additional parity block, thus counting two parity blocks instead of one.

Blocks parity are distributed in two different units. This allows the system has higher fault tolerance, supporting the failure of up to two units. Even so, the information can be lost due to multiple failure in the units.

For this type of configuration a minimum of four units are required. The greater the number of hard drives in the system, the greater the probability of failure.

👉 Advantages of RAID level 6 versus RAID level 5 👈

Types of nested RAID levels

So far we have seen the basic RAID levels, but there are more advanced systems, such as nested RAID levels. These configurations are combinations of the different RAID systems seen. For this type of configuration we will have a main RAID level and within this several sublevels with another configuration. This configuration is much more common in business-class NAS, since they are where it pays off to make a more complex configuration in order to secure the data on several units and have a high speed of response and data reading.

RAID level 01

Partition mirroring systems. We have a main RAID 1 level and within this at least two sets with a RAID 0. This level is also called as RAID 0 + 1.

The data is will write in the first set level 0. These data will be replicated in set level 0 secondary because both levels are linked by a level 1.

One of the drawbacks of this RAID system is scalability. If we add a hard disk in primary level 0, we must add a second disk in second level 0. Logically the disks must have the same capacity as the disks already installed.

La fault tolerance of this RAID configuration is limited. The breaking of a disk in one of the levels 0 and even of the disks of the same level is supported. Any other combination of failures would cause data loss.

RAID level 10

Mirror division system. Here the main level is a RAID 0, having at least two RAID 1 sublevels. This type of configuration is also called RAID 1 + 0.

The data will be written as if we had a level 0, but there is a copy on a secondary disk, as in a level 1. At least one level 1 disk must be kept in order to recover data. If all the disks failed or an error occurred, all the information contained in all the levels 1 that we have would be lost.

RAID level 30

Division with dedicated parity set. This type of system is the combining a RAID 3 with a RAID 0. It can also be found as RAID 3 + 0.

This type of disc mounting offers a high transfer rate with uncompromising reliability. One of the downsides is the high cost of implementation.

Its construction is based on combining two RAID 3 sets with the data divided between both sets. The data is divided into smaller blocks and divided among the level 3 sets. These in turn are divided into smaller parts and parity is calculated by applying an XOR to each one. Data is written to all hard drives except one, where we store parity information.

Allow one disk in each RAID 3 set to fail. This means that yes multiple of the RAID 3 disks, the array data will be lost. The period of Recovery time necessary, ranging from detection to reconstruction of the whole, it is a period of vulnerability for the RAID set.

RAID level 50

High speed parity assembly. This system combines a RAID 0 level from which multiple RAID 5 systems hang. It is characterized by offering a good redundancy and good transfer speeds. Are required at least six hard drives for a RAID 50.

This system allows us one disk in each level 5 set may have a failure. In case of failure of more than one drive, data in the set is lost. From the detection of the failure to the reconstruction of the data on the new disk, there is a vulnerability of the RAID set.

The configuration of the RAID sets affects the fault tolerance. If we have three groups of level five with seven units each, we get higher storage capacity and efficiency. This configuration only supports the failure of three storage units.

As the reliability is based on quick replacement of damaged drives for rebuilding, Typically six-disk RAID 5 systems are built with one online spare. This allows a immediate replacement and rapid rebuilding in case of failure. We must bear in mind that the reconstruction puts a stress on the system, since every bit is required to be read, making it very vulnerable.

This system RAID 50 offers you a great writing speed and a higher fault tolerance than a RAID 5. This configuration is recommended for applications that require high fault tolerance, capacity, and performance.

RAID level 60

High speed additional parity set. Is based on the combining a RAID level 0 with RAID level 6 double parity sets. In the RAID 60 requires at least eight hard drives.

Offers the advantage of high speeds thanks to level 0 and improves security thanks to double parity of level 6. It also offers us greater data protection than a simple level 6.

This setting is recommended for when high fault tolerance, high capacity, and high write speeds are required.

RAID level 100

A system of divided sets that are in turn divided together again. It is a division of RAID 10 sets within a RAID 0 set, or gridded RAID. This system is often referred to as RAID 10 + 0.

Damage to all level 1 discs except one is supported without losing stored information. At the same time the remaining disk of level 1 would become a single point of failure of the damaged system. Usually the upper level of division is done through software.

A gridded RAID offers benefits over a single RAID level. Above all it offers random read performance improvements and mitigation of risk hot spots in the set.

This system normally mounts on very large database systems, where the underlying software limits the number of physical disks that can be attached to the array. Creating these animated levels allows you to virtually eliminate the limit of physical units.

What happens if a disk fails in RAID 0?

When a RAID 0 volume fails or is offline, the data on the drive is no longer accessible or recoverable normally. This is why it is always recommended to have at least RAID 1 on small servers and workstations.