Data Striping - HP Surestore Disk Array 12h - And FC60 Service Manual

Data
+
0
1
1
If this bit is
0
now written
as 1...
0
Figure 17
Calculating Data Parity

Data Striping

Data striping, which is used on RAID 0, 0/1, 3 and 5 volume groups, is the performance-
enhancing technique of reading and writing data to uniformly sized segments on all disks in
a volume group simultaneously. Collectively, the segments comprise a stripe of data on the
volume group. Data striping enhances performance by allowing multiple sets of read/write
heads to execute the same I/O transaction simultaneously.
The amount of information simultaneously read from or written to each disk is the stripe
segment size. The stripe segment size is configurable to provide optimum performance
under varying sizes of I/O transactions. Stripe segment size is specified in 512-byte blocks
of data.
Stripe segment size can affect disk array performance. The smaller the stripe segment size,
the more efficient the distribution of data read or written across the stripes in the volume
group. However, if the stripe segment is too small for a single I/O operation, the operation
requires access to two stripes. Called a stripe boundary crossing, this action may negatively
impact performance.
The optimum stripe segment size is the smallest size that will rarely force I/Os to a second
stripe. For example, assume your application uses a typical I/O size of 64
Data Data Data
+ +
1
1
0
0
1
1
1
0
0
1
Parity
=
+
1 1
1 0
0 1
0 1
1 0
This bit will also
be changed to a 1
so the total still
equals 0.
Disk Array High Availability Features
0
0
0
0
0
B. If you are
K
47