IBM N Series, Hardware Manual

Page: 173 / 368

Chapter 10. Data protection with RAID Double Parity
153
10.4.1 Internal structure of RAID-DP
With RAID-DP, the traditional RAID 4 horizontal parity structure is still employed and becomes
a subset of the RAID-DP construct; that is, how RAID 4 works on storage is not modified with
RAID-DP. Data is written out in horizontal rows with parity calculated for each row in RAID-DP,
which is considered the row component of double parity. If a single disk fails or a read error
from a bad block or bit error occurs, the row parity approach of RAID 4 is used to re-create the
data. RAID-DP is not engaged. In this case, the diagonal parity component of RAID-DP is a
protective envelope around the row parity component.
10.4.2 RAID 4 horizontal row parity
Figure 10-7 shows the horizontal row parity approach that is used in the traditional RAID 4
solution. It is the first step in establishing an understanding of RAID-DP and double parity.
Figure 10-7 RAID 4 horizontal parity
Figure 10-7 represents a traditional RAID 4 group that uses row parity. It consists of four data
disks (the first four columns, labeled D) and the single row parity disk (the last column,
labeled P). The rows represent the standard 4 KB blocks that are used by the traditional RAID
4 implementation. The second row is populated with sample data in each 4 KB block. Parity
that is calculated for data in the row is then stored in the corresponding block on the parity
disk.
In this case, the way parity is calculated is to add the values in each of the horizontal blocks.
That sum is stored as the parity value (3 + 1 + 2 + 3 = 9). In practice, parity is calculated by an
exclusive-OR (XOR) process, but addition is fairly similar and works as well for the purposes
of this example. If you must reconstruct data from a single failure, the process that is used to
generate parity is reversed. If the first disk fails, RAID 4 re-creates the data value 3 in the first
column. It subtracts the values on the remaining disks from what is stored in parity (9 - 3 - 2 -
1 = 3). This example of reconstruction with single-parity RAID shows why data is protected up
to, but not beyond, one disk failure event.