Raid level difference
Technical work guidance
Have a comprehensive understanding of the advantages and disadvantages of RAID-1 to RAID-6 to choose the RAID that best suits your specific storage needs.
Data is the lifeblood of many organizations, and fast and reliable access to data is the key to a company's survival in today's "Internet age" world. Similarly, many organizations will use a certain level of RAID (Redundant Array of Inexpensive Disks) to protect data integrity. Now you can take a look at how many people around you are adopting RAID-5 on servers because of its simplicity and the seemingly best-fitting needs. Although RAID 5 is a good choice in many cases, if you pay more attention to the initial write performance, then other levels of RAID may also have good performance. Now how many of you can explain the design purpose of RAID 10 and RAID 50? These relatively new RAID levels can overcome some of the shortcomings of RAID 5, while still ensuring that the storage system has sufficient redundancy. This article will evaluate the advantages and disadvantages of representative RAID levels one by one.
RAID 0 (striped disk)
- Minimum number of drives required: 1
- Maximum capacity: the number of disks x the capacity of a single disk
- Description: The data written to the disk is divided into blocks and written to each disk separately.
- Advantages: Because the data is read and written through multiple axes, the input and output loads are dispersed, so the read and write speed is very fast. In theory, the more disks, the better the system performance. Under normal circumstances, if you are pursuing performance improvements, it is best to use Iometer to test the specific storage performance, because the actual improvement may not be as obvious as expected.
- Defect: When a certain disk has a problem, the entire disk array will fail, because this level of RAID does not have any safety devices. The increase in the number of disks means an increase in the risk of failure.
- Minimum number of drives required: 2 or a multiple of 2
- Maximum capacity: total capacity of the disk array/2
- Description: All data will be duplicated and stored on two physical hard disks to provide a high level of redundancy.
- Advantages: High reliability. The failure will only affect one disk. The read performance of RAID 1 is relatively high (equal to or better than the performance of a single disk).
- Defect: Since each disk has a mirror disk, the operating cost has increased by 100%. Since data needs to be written on two disks, the write performance will be affected to a certain extent, but even so, its write performance is better than other levels of RAID.
RAID 2: Discontinued
RAID 3 (Parallel transmission disk with parity information)
- Minimum number of drives required: 3
- Maximum capacity: (number of disks-1) x capacity of each disk
- Description: The data is written to each disk evenly after being broken up in bytes. All verification information is additionally written on a separate dedicated disk.
- Advantages: Can tolerate the loss of data on a separate disk. Continuous write performance is acceptable. Continuous reading performance is better.
- Defects: It is not widely used, so there is less information to solve the problem. RAID 3 is generally considered to be more effective. Random write performance is poor. It is better to read at will.
RAID 4 (independent data disk with shared parity block)
- Maximum capacity: (number of disks-1) x capacity of each disk
- Description: A file is divided into blocks, and each block is written to multiple disks, but the average write is not required. Similar to RAID 3, RAID 4 also uses a separate physical disk to store parity information. It is the best choice under the environment of large data volume and high read rate.
- Minimum number of drives required: 3
- Advantage: High reading rate. Allow single disk data loss.
- Defect: Poor writing performance. Block read performance is good.
- Maximum capacity: (number of disks-1) x capacity of each disk
- Description: Similar to RAID 4, the data blocks are scattered and written on each disk (sometimes not evenly), but in this case, the parity information and the written information are scattered on each disk.
- Minimum number of drives required: 3
- Advantages: Good support. Allow single disk data loss.
- Defect: Poor performance during data reconstruction. Because the verification information needs to be constantly updated during the writing process, the writing performance is sometimes unsatisfactory.
RAID 6 (independent data disk group with two independent stripe parity information)
- Maximum capacity: (number of disks-2) x capacity of each disk
- Description: Similar to RAID 4, the data is divided into blocks and written to the entire disk group (sometimes not evenly written), but in this case, the parity data is also distributed to the entire disk group.
- Minimum number of drives required: 3
- Advantages: Data loss is allowed on up to two disks. Excellent reading performance. Suitable for all critical application environments.
- Defect: The writing performance is not very satisfactory. Due to the need to update multiple parity information, the write performance is not even as good as RAID 5. The performance drops significantly during the data reconstruction process.