You may already be aware of the word ‘RAID’ and you are here to know more about the RAID levels. An introduction to RAID is given below for the benefit of newbies.
The word RAID actually stands for Redundant Array Independent Data. Usually, hard disks of our computers are directly connected to the motherboard. But that should not be the ideal case always. We hardly care for the read/write speed of data in our home environment, but it matters when we consider an Enterprise Storage Solution. For the redundancy of data and to improve the reliability & performance, we use an additional hardware called RAID, Controller. Unlike in ordinary system, here we connect the Array of drives into the RAID card controller and hence RAID card is fully responsible for managing the data read/write to each HDDs. Even though, there is a dedicated storage controller, we have to setup the disk controller according to our requirements.
There must be at least two drives in RAID0 and the data will be divided into two blocks and it writes to all drives equally. None of the drive stores the whole data, instead it is distributed to all drives as blocks.
It has good IO performance, as it divides the data into blocks and distributes it across all drives.
The best option for high read/write applications.
Easy setup process.
- Cannot retrieve the data from the storage without all drives.
Unlike RAID0, RAID1 is safer to store the data since RAID1 is using mirroring mechanism. Here, instead of distributed writing of data blocks to different drives, RAID1 keeps same data in both drives.
1. Chances of data loss because of single drive failure are very less.
2. Can easily retrieve the data because the same data is available in another drive too.
3. Even though it is expensive, performance of RAID1 is good.
1. You can only use half of the total disk space (another half for mirroring)
RAID5 uses a striping mechanism like RAID0, but here an additional feature called ‘parity check’ is used to solve the main disadvantage of RAID0. RAID5 gives the solution to single disk failure data loss, the parity information along with the Data will help to reconstruct the corrupted data in the failed disk.
- Data is more reliable and redundant in RAID 5
- Gives best IO performance
- Good option for high IO operation servers such as a Database server.
- The time required to reconstruct the data in failed disk depends on the size of the drive. If it is a heavy size disk then it may take two or three days to complete.
- Chances of another disk failure during the reconstruction of data in the first disk.
RAID6 exactly uses the mechanism of RAID5 (data stripping and writing). In order to solve the chances of another disk failure during the reconstruction of data in the first disk, RAID5 uses two parity information distributed across the drives.
- More safe than RAID5 because of the two parity blocks.
- High fault tolerance protection, even after two concurrent disk failures.
- Read performance is good.
- Sometimes write performance is poor because of the time required to calculate parity information.
RAID 10 is the combination of RAID0 and RAID1, hence the advantages of both raid levels are available in RAID10. This RAID level gives the best protection for data because of its mirroring and stripping mechanism. RAID10 requires at least four drives, but you can only make use of half of its total storage because of data mirroring.
1. RAID10 gives best IO performance and redundancy along with data protection.
2. Provides best fault tolerance
3. The time required to rebuild the data is very less.
1 Cannot use the entire disk space due to mirroring.
Even though, RAID provides best fault tolerance and protection to your data, it does not replace the need of backups. RAID is actually designed to improve the IO operations and fault tolerance such as data loss because of Hard disk failures. It is always recommended to use backups for your server to keep the data secure.