▷ Raid 0, 1, 5, 10, 01, 100, 50: explanation of all types

Today we are going to see what RAID technology is, which in addition to being a very good anti-mosquito spray brand, also has to do with technology from the computer world. We will see what its operation consists of and what we can do with it and its different configurations. In it, our mechanical hard drives or SSDs will take center stage, whatever they are, which allow us to store enormous amounts of information thanks to the drives of more than 10 TB that we can currently find.

You may have also heard of cloud storage and its advantages over storage on our own team, but the truth is that it is more business-oriented. These pay a price to have this type of service that is provided through the internet and remote servers that have advanced security systems and proprietary RAID configurations with great data redundancy.

What is RAID technology?

The term RAID comes from "Redundant Array of Independent Disks" or said in Spanish, redundant array of independent disks. By its name we already have a good idea of what this technology intends to do. Which is nothing more than creating a system for data storage using multiple storage units among which the data is distributed or replicated. These storage units can be either mechanical or HDD hard drives, SSD or solid state drives.

RAID technology is divided into configurations called levels, through which we can obtain different results in terms of information storage possibilities. For practical purposes, we are going to see a RAID as a single data store, as if it were a single logical drive, even though there are several physically independent hard drives within it.

The ultimate goal of RAID is to offer the user a greater storage capacity, data redundancy to avoid data loss and to provide faster data reading and writing speeds than if we only had a hard disk. Obviously these features will be independently enhanced depending on what level of RAID we want to implement.

Another advantage of using a RAID is that we can use old hard drives that we have at home and that we can connect via SATA interface to our motherboard. In this way, with low-cost units, we will be able to mount a storage system where our data will be safe against failures.

Where RAIDs are used

In general, RAIDs have been used for many years by companies, due to the special importance of their data and the need to preserve it and ensure its redundancy. These have one or more servers that are specifically dedicated to managing this information store, with hardware specifically designed for this use and with a protection shield against external threats that will prevent undue access to them. Typically, these warehouses use identical hard drives in performance and manufacturing technology, for optimal scalability.

But today, almost all of us will be able to use a RAID system if we have a relatively new motherboard and with a chipset that implements this type of internal instructions. We will only need several disks connected to our base bale to start configuring a RAID from Linux, Mac or Windows.

In case our team does not implement this technology, we will need a RAID controller to manage the warehouse directly from hardware, although in this case the system will be susceptible to failures of this controller, something that for example does not happen if we manage it through software.

What a RAID can and cannot do

We already know what a RAID is and where it is possible to use it, but now we must know what advantages we are going to obtain by implementing such a system and what other things we will not be able to do with it. In this way we will not fall into the error of assuming things when they really are not.

Advantages of a RAID

High fault tolerance: With a RAID we can obtain a much better fault tolerance than if we only have a hard disk. This will be conditioned by the RAID configurations that we adopt, since some are oriented to provide redundancy and another simply to achieve access speed. Read and write performance improvements: As in the previous case, there are systems aimed at improving performance, by dividing data blocks into several units, to make them work in parallel. Possibility of combining the two previous properties: RAID levels can be combined, as we will see below. In this way we can take advantage of the access speed of some and the redundancy of data of another. Good scalability and storage capacity: another of its advantages is that they are generally easily scalable systems, depending on the configuration we adopt. In addition, we can use discs of different nature, architecture, capacity and age.

What a RAID cannot do

A RAID is not a means of data protection: RAID will replicate data, not protect it, they are two very different concepts. The same damage will be done by a virus on a separate hard drive, as if it entered a RAID. If we do not have a security system that protects it, the data will be equally exposed. Better access speed is not guaranteed: there are configurations that we can make ourselves, but not all applications or games are capable of working well on a RAID. Many times we are not going to make a profit by using two hard drives instead of one to store data in a divided way.

Disadvantages of a RAID

A RAID does not ensure recovery from disaster: as we know, there are applications that can recover files from a damaged hard disk. For RAIDs, you need different and more specific drivers that are not necessarily compatible with these applications. So in the event of a chain or multiple disk failure, we could have unrecoverable data. Data migration is more complicated: cloning a disk with one operating system is quite simple, but doing it with a complete RAID to another is much more complicated if we don't have the correct tools. This is why migrating files from one system to another to update it, is sometimes an insurmountable task. High initial cost: implementing a RAID with two disks is simple, but if we want more complex and redundant sets, things get complicated. The more disks, the higher the cost, and the more complex the system, the more we will need.

What RAID levels are there

Well we can find quite a few RAID types today, although these will be divided into standard RAID, nested levels and proprietary levels. The most frequently used for private users and small businesses, are of course the standard and nested levels, since most high-end equipment have the possibility of doing it without installing anything extra.

On the contrary, the proprietary levels are only used by the creators themselves or who sell this service. They are variants of those considered basic, and we do not believe their explanation is necessary.

Let's see what each of them consists of.

RAID 0

The first RAID we have is called Level 0 or divided set. In this case, we do not have data redundancy, since the function of this level is to distribute the data that is stored among the different hard drives that are connected to the computer.

The objective of implementing a RAID 0 is to provide good access speeds to the data stored on the hard drives, since the information is equally distributed on them to have simultaneous access to more data with their drives running in parallel.

RAID 0 does not have parity information or data redundancy, so if one of the storage drives breaks, we will lose all the data that was inside it, unless we have made external backups to this configuration.

To perform a RAID 0 we must pay attention to the size of the hard drives that make it up. In this case it will be the smallest hard disk that determines the added space in the RAID. If we have a 1 TB hard drive and another 500 GB in the configuration, the size of the functional set will be 1 TB, taking the 500 GB hard drive and another 500 GB from the 1 TB disk. This is why the ideal would be to use hard drives of the same size to be able to use all the available space in the designed set.

RAID 1

This configuration is also called mirroring or “ mirroring ” and is one of the most commonly used to provide data redundancy and good fault tolerance. In this case, what we are doing is creating a store with duplicate information on two hard drives, or two sets of hard drives. When we store a data, it is immediately replicated in its mirror unit to have twice the same data stored.

In the eyes of the operating system, we only have one storage unit, which we access to read the data inside. But in case this fails, the data will be automatically searched in the replicated drive. It is also interesting to increase the speed of reading data, since we can read the information simultaneously from the two mirror units.

RAID 2

This level of RAID is little used, since it is basically based on making distributed storage on several disks at the bit level. In turn, an error code is created from this data distribution and stored in units exclusively intended for this purpose. In this way, all the disks in the warehouse can be monitored and synchronized to read and write data. Because the disks currently already carry an error detection system, this configuration is counterproductive and the parity system is used.

RAID 3

This setting is also not currently used. It consists of dividing the data at the byte level into the different units that make up the RAID, except one, where parity information is stored to be able to join this data when it is read. In this way, each stored byte has an extra parity bit to identify errors and to recover data in the event of loss of a drive.

The advantage of this configuration is that the data is divided into several disks and access to information is very fast, as much as there are parallel disks. To configure this type of RAID you need at least 3 hard drives.

RAID 4

It is also about storing the data in blocks divided among the disks in the store, leaving one of them to store the parity bits. The fundamental difference from RAID 3 is that if we lose a drive, the data can be reconstructed in real time thanks to the calculated parity bits. It is aimed at storing large files without redundancy, but data recording is slower precisely due to the need to do this parity calculation every time something is recorded.

RAID 5

Also called a parity distributed system. This one is used more frequently today than levels 2, 3 and 4, specifically on NAS devices. In this case, the information is stored divided into blocks that are distributed among the hard drives that make up the RAID. But also a parity block is generated to ensure redundancy and to be able to reconstruct the information in the event that a hard disk becomes corrupted. This parity block will be stored in a different unit from the data blocks that are involved in the calculated block, in this way the parity information will be stored in a different disk from where the data blocks are involved.

In this case, we will also need at least three storage units to ensure data redundancy with parity, and failure will only be tolerated on one unit at a time. In case of breaking two simultaneously, we will lose the parity information, and at least one of the data blocks involved. There is a RAID 5E variant where a spare hard drive is inserted to minimize data rebuild time if one of the major fails.

RAID 6

RAID is basically an extension of RAID 5, in which another parity block is added to make a total of two. The information blocks will be divided again into different units and in the same way the parity blocks are also stored in two different units. In this way the system will be tolerant to the failure of up to two storage units, but, consequently, we will need up to four drives to be able to form a RAID 6E. In this case there is also a variant RAID 6e with the same objective as that of RAID 5E.

Nested RAID levels

We left behind the 6 basic levels of RAID to enter the nested levels. As we can assume, these levels are basically systems that have a main level of RAID, but that in turn contain other sublevels that work in a different configuration.

In this way, there are different RAID layers that are capable of simultaneously performing the functions of the basic levels, and thus be able to combine, for example, the ability to read faster with RAID 0 and the redundancy of RAID 1.

Let's see then which are the most used today.

RAID 0 + 1

It can also be found under the name RAID 01 or partition mirror. It basically consists of a main level of type RAID 1 that performs the functions of replicating the data found in a first sublevel in a second. In turn, there will be a sub-level RAID 0 that will perform its own functions, that is, store the data in a distributed way among the units that are in it.

In this way we have a main level that does the mirror function and sublevels that do the data division function. This way when a hard drive fails, the data will be perfectly stored in the other mirror RAID 0.

The disadvantage of this system is scalability, when we add an additional disk on one sublevel, we will also have to do the same on the other. In addition, fault tolerance will allow us to break a different disk at each sublevel, or break two at the same sublevel, but not other combinations, because we would be losing data.

RAID 1 + 0

Well now we would be in the opposite case, it is also called RAID 10 or mirror division. Now we will have a main level of type 0 that divides the stored data between the different sublevels. At the same time we will have several type 1 sublevels that will be in charge of replicating the data on the hard drives that they have inside.

In this case, fault tolerance will allow us to break all the disks in one sublevel except for one, and it will be necessary for at least one healthy disk to remain in each of the sublevels so as not to lose information.

RAID 50

Of course, this way we can spend some time making possible combinations of RAID to which is more convoluted to achieve maximum redundancy, reliability and speed. We will also see RAID 50, which is a main level in RAID 0 that divides the data from the sublevels configured as RAID 5, with their respective three hard drives.

In each RAID 5 block we will have a series of data with its corresponding parity. In this case, a hard disk can fail in each RAID 5, and it will ensure the integrity of the data, but if they fail more, we will lose the data stored there.

RAID 100 and RAID 101

But not only can we have a two-level tree, but three, and this is the case of RAID 100 or 1 + 0 + 0. It consists of two sub-levels of RAID 1 + 0 divided in turn by a main level also in RAID 0.

In the same way we can have a RAID 1 + 0 + 1, made up of several RAID 1 + 0 sublevels reflected by a RAID 1 as the main one. Its access speed and redundancy are very good, and they offer good fault tolerance, although the amount of disk to use is considerable compared to the availability of space.

Well this is all about RAID technology and its applications and features. Now we leave you with a few tutorials that will also be useful to you

We hope this information has been useful for you to better understand what a RAID storage system is. If you have any questions or suggestions, please leave them in the comment box.