A.3
RAID 5 (Striping + Parity)
In addition to striping (the process distributing data across multiple hard disks), RAID5 creates parity
data for redundancy. Parity data is derived via an algorithm from the original data. For example,
when writing six blocks of data (D0 - D5) with three hard disks, each data block and the calculated
parity data P (DX, DY) are distributed across the hard disks (see below figure). Because redundancy
is achieved by the parity data, a hard disk worth of capacity is used for the parity data.
56
D0
D2
P(D2,D3)
P(D4,D5)
Disk1
Disk2
D3
D4
D2
D5
D1
D0
Data
D1
P(D0,D1)
D3
D4
D5
Disk3