Software disk level construction, this process is generally used to service with large servers, for the general small PC machine, is almost not used, his name is raid
1. RAID-0 (isometric mode)
When it is distributed, it adopts the principle of equal distribution, the same model and capacity of the disk to make up, when a file to be stored, the file will be cut according to the fast size, and then deposited to the disk, if there are two disks, then 100M of the file will be divided into two 50M, put into the two, At this point the capacity of each block is reduced, the capacity is larger, but if a disk is damaged, then all the data stored at this time will be lost and unreadable
2. RAID-1 (full backup)
This pattern also requires the same capacity, the best time to be identical to the disk, this mode is to keep the same data intact on two disks, while the capacity of the disk will be reduced by 50%, this mode is the largest bit of data backup, can be lost data, maintain the security of the data
3. RAID 0+ 1,raid 1 + 0
For the two ways described earlier, one data security is not good, the other is not the capacity of the disk to make an effective use of the two methods mentioned are a combination of the first two, there are two groups of RAID0, a set of data storage, a set of data for the backup, Merging is RAID1, as shown in the following figure
1. RAID 5 (Balance of performance and data backup considerations)
It takes at least 3 or more disks in this structure to make up this type of disk array, a bit like a RAID 0 structure, but each loop is written with a single check on each disk, which records backup data from other disks for rescue when a disk is damaged, Equivalent to two other pieces of data records, the backup data will be saved in the third block, the first two pieces of which damage, you can use the third and another (undamaged) calculation. And the same-bit check code is computed by CPU rather than a dedicated disk array card, RAID 5 can only support a disk damage by default