Common RAID levels: RAID0, RAID1, RAID5, RAID10, RAID50, JBOD
RAID0: ide controller is not required. The use of a plurality of disk controller chip (HBA host adapter) connected in parallel, and through other means in the computer motherboard link. Select a reasonable band on the N disks to create a stripe set. The data is simultaneously written into all hard dispersed, the read data is simultaneously read all hard, n times faster. While RAID0 can provide more space and better performance, but the system is not reliable, if one disk fails the entire system will collapse, unable to remedy. More suitable for temporary file storage. Disk size is equal to the best, the best models of the same disk.
RAID1: also known as mirrored disks. The principle of a mirror disk data to another disk. It has at least two disks, a disk when a problem occurs, the system will run properly, because it will use another piece of data backup disk to work. Have a good disk redundancy. Although the absolute security of data, but the cost will be significantly increased, disk utilization is 50%. Disk problems must be promptly replaced, otherwise the remaining mirror disk problems occur, then the system will collapse. The new disk will be more original data will take a long time synchronous mirroring, external data access will not be affected. Used for occasions to save important data.
RAID4: at least three disks, two disks store data, a disk-storage heterologous additional first block and the second block or a data disk, (they are equal to false, the two different True), when a broken disk the bad, the system can work. Efficacy a further code may be calculated based on the data pieces data disk bad disk, then the disk transfer data to the host. RAID4 which a spare disk can also be added as a substitute, substitute called hot standby disk, usually do not work hot standby. When a bad disk broken, broken pieces of the spare disk replacement work is called heat exchange. Allowing a disk error, read and write performance, utilization of (n -) / n.
RAID5: at least three disks, RAID4 efficacy code storage disk read easily lead to increased pressure frangible disk bad. In order to alleviate the pressure RAID5 each disk are stored codes and data, as each disk work. In order to improve read and write rates of the RAID, which added RAID cache memory, the write data is temporarily stored in the buffer memory, the buffer memory is then written to disk slowly. Faster data write capability. In order to prevent power outages cause loss of data buffer memory which can not be timely written to disk, RAID will generally add a spare battery, so that RAID will continue working data buffer memory is written to disk.
RAID01: at least four disks, data is first performed RAID0 RAID1 mode re-execution mode, the data transfer incoming data to be dispersed, and then enters the next one, dispersing mirror into two data disks. A disk corruption does not affect the system work, even if another group damaged a disk does not affect the system work, but not when the same set of two disk corruption, causing the system data loss. Read and write performance, memory utilization of 50%.
RAID10: data transfer in, RAID1 perform a mirroring mode, and then the image data are transmitted to the next stage, and then execute RAID0 mode, the data storage disk dispersion. Allows a group of a disk is damaged, the system can continue to work. Each group appears if a damaged system can not continue to work.
JBOD: n the form of a block of the same hard disk hardware used by the intelligent disk controller or the disk drive operating system software program in a manner together to create the play of the disk. In order to write data to the hard disk. Advantage aneuploidy increase disk capacity. Disadvantage of one disk fails, the entire system will be destroyed.
---------------------
---------------------