The blog post does not talk about professional vocabulary, only writes my own vernacular understanding...
First of all, let me mention the definition of RAID, at least you have to know its full name, haha~
Redundant Array of Independent hard disk ( the RAID , Redundant the Array of work of the Independent Disks), formerly known as the Redundant Array of Inexpensive Disks (Redundant Array of Inexpensive Disks), referred to as a disk array .
As a data protection method, RAID technology is widely used in the field of enterprise data security.
There are mainly 7 standard RAID levels: RAID0, RAID1, RAID2, RAID3, RAID4, RAID5, RAID6 (Although they look similar in name, the mechanism of each level is quite different...)
In addition, there are some combination implementation methods: RAID01 (that is, RAID0+RAID1 mixed deployment), RAID100 (the reason is the same as before) and so on.
One, RAID0
Simply understand that RAID0 does not essentially provide any form of data redundancy. It just strips the data and stores it separately on different disks. The data stored on different disks can be read and written concurrently by I/O at the same time. To put it bluntly, the data is sliced and stored in multiple disks, and the stored data can be read from these multiple disks at the same time when reading, and the speed is greatly improved. Therefore, RAID0 improves the efficiency of reading and writing, and has high bandwidth utilization, and its final rate will be affected by bandwidth. In theory: If there are n disks, the data read and write efficiency after RAID0 can be increased by n times.
Two, RAID1
RAID1 is actually mirroring . Whenever a copy of data is generated, an identical backup data will be generated on the mirrored disk at the same time. When the disk is damaged, the system can use the data in the mirrored disk for recovery, so in essence, the space of RAID1 The utilization rate is 50%. In other words, after completing RAID1 with two 300GB hard drives, there will be less than 300GB of storage space left.
Three, RAID2
RAID2 is called Error Correction Hamming Code Disk Array. Hamming code verification refers to inserting a check bit at a specific position in the original data to perform data check and error correction. The Hamming code check bit is 2n, which is the first of the data. Bits 2, 4, 8... are parity bits, and the remaining bits are data bits. In RAID2, data is stored bit by bit, each disk stores one bit of data, and the data width can be set by the user.
Hamming code has its own error correction capability, so RAID2 can correct errors in the case of data errors and ensure data security. Its data transmission performance is quite high, and the design complexity is lower than the RAID3, RAID4 and RAID5 introduced later.
However , the data redundancy overhead of Hamming Code is too large, and the data output performance of RAID2 is limited by the slowest disk drive in the array. Furthermore, the Hamming code is a bitwise operation, and RAID2 data reconstruction is very time-consuming. Due to these significant defects and most of the disk drives themselves have error correction functions, RAID2 is rarely used in practice and has not formed a commercial product. At present, mainstream storage disk arrays do not provide RAID2 support.
Four, RAID3
RAID3 is a parallel access array using dedicated parity disks, that is, one disk is used as a parity disk, and the remaining disks are data disks. When storing, the data is cross-stored to the data disk by bit, and at the same time, the calculated check digit data is generated and stored in the check disk, which can realize the data simultaneous reading, and has high fault tolerance and high reading efficiency. When a piece of disk data is damaged, it can be restored with the help of the check disk data and other disk data.
Five, RAID4
RAID4 and RAID3 are the same, except that RAID4 stores data in units of blocks to ensure the integrity of data blocks. (This is not commonly used)
Six, RAID5
RAID5 is currently the most widely used RAID level. It strips data and parity data and stores them on each disk. In other words, each disk stores data and parity data. When a certain disk occurs When an error occurs, it can be recovered based on the data and check data of other disks.
Seven, RAID6
RAID6 is an extension of RAID5, which introduces double parity, which can support data recovery when two disks are damaged. RAID6 supports not only data recovery, but also verification data recovery, but at a higher cost.
The above is only a learning record and will continue to be improved...