| |
RAID
Parity
In
RAID systems one or more drives are used to store compressed parity
that describes the data being stored on the other drives in the
array. This is a far more sophisticated form of parity that allows
the reconstruction of missing or corrupted data.
Example:
There
are five people in a room. The first four are all assigned a unique
number from "1" to "4". This is the equivalent
of the data stripe across the hard drives performed by RAID controllers.
A fifth number is assigned to the fifth person. This number is "10".
Ten is the sum of all the other numbers: 1 + 2 + 3 + 4 = 10. This
is the equivalent to the parity chunk written to one of the RAID
drives.
If the person designated "2" left the room, you could
determine the number they had been assigned by adding together the
remaining three people's numbers and subtracting this from the number
held by the person designated as the parity number ("10"):
1 + 3 + 4 = 8. 10 - 8 = 2.
Therefore number "2" is missing and can be reassigned
to a new person who joins the group (a spare or new drive). In this
way data security is preserved with a data to storage capacity ratio
far below direct mirroring of data.
Related
topics:
|
