Contemporary disc systems of RAID
RAID-4.
Level 4: disks in the array level 4 is not synchronized, as has been done in the
disk level 3, and therefore they could work independently, providing a good
array of features when using it in multi-file, although the parity continues to
inscribe highlighted drive. This is the best solution for file servers, from
which users frequently read data, but rarely performed on the record information
on disks.
RAID-5.
The user perceives array as a single logical disk containing the required memory
capacity. Restoring data is also provided through Hog operations, but in this
case Hog-segmenty distributed to all disks. A segment is located on the first
drive, V-segment on the second disc, C-segment in the third drive. The first
segment of the fourth disc contains Xor (A, B, C). Segment Data 0 placed on the
first drive next to the A segment, a segment on the E-drive followed by the
second segment B. The second segment of the third disc, following the C segment
contains Xor (D, E, F), and F is the second segment in the fourth drive. For the
next three segments Hog (G, H, I) put on the second disc, and so on. The size
segment in RAID-5 is also closely related to the mechanism of redundancy. It is
expected that during the recording available to only the data needed to
calculate the Hog. Used algorithm: Read-Write-changed. A large segment size is
selected in relation to the size of the user query. In reading only access to
this area on the disk. When the record, the only record this area on the second
segment of the disk and related Xor area on the fourth drive. With this amount
of the request almost all requests fall inside one segment. When reading data
address specified in the request, is shown in a physical address segment on the
disk containing the data. Then read data from all disks simultaneously. When
recording also made a request to display the physical address, and then read the
old data and related Hog-dannye. Recordable data stack modulo 2 with previous
Hog-dannymi, the result is a checklist for the amount of new data. New data and
new Hog-kombinatsiya written to the appropriate disks.
Next part
|