System information

Appendix
A-5
A.5 RAID 5
With RAID 5, the system calculates parity from data on three drives. If one of the drives fails, parity data can be used to
rebuild the lost data. Under RAID 5, parity data is stored across all disks in the array. This maximizes the amount of
storage capacity available from all drives in the array while still providing data redundancy. Data under RAID 5 is block-
interleaved.
The diagram below represents the writing of data on a RAID 5 array composed of four HDDs connected to the controller.
Parity blocks are represented by the letter P.
RAID 5: Independent data disks with distributed parity blocks
Characteristics
• Storage capacity = (number of disks -1) x (capacity of the smallest disk)
• A minimum of three disks are required.
• Fault tolerance: Good
• Each data block is written to a disk. The parity of blocks with the same rank is generated on writes,
recorded in a distributed location and checked on reads.
• Highest read data transfer rate, medium write data transfer rate
• Relatively low ration of (parity) disks to data disks results in high efficiency.
• Good aggregate transfer rate
• Most versatile RAID level
Recommended use
• File and application servers
• Database servers
• Internet, email and news servers
• Intranet servers
B
P
CD
E
H
PAB
D
F
P
GH
A
C
P
EF
G
A
B
C
G
F
D
E
CONTROLLER
Hot Spare
Figure A-4 RAID 5 disk array