What is RAID?

Toggle sidebar Toggle sidebar
Solution
P
Sort by date Sort by votes


RAID is an acryonym that stands for Redundant Array of Independent/Inexpensive Disks.

RAIDs allow for multiple disks to work together. The exact benefit of the raid, as well as any drawbacks, depend on the particular configuration.

RAID-0, also called striping, interleaves blocks of data across two or more disks. A RAID-0 configuration has the total capacity of all the physical volumes that are a part of the configuration. If three two-terabyte hard disk drives are placed in RAID-0, the total size of the RAID volume will be six terabytes.

RAID-0 takes advantage of spatial locality. When data is accessed, there is a high probability that data that is logically nearby will also be accessed. By ping-ponging logically sequential blocks of data across two or more identical disks, the storage controller can put all of them to work at the same time. This allows data to be read faster, and written faster, than it would be if a file were contained on only a single hard disk drive. The drawback of RAID-0 is that a failure in any of the disks in the RAID will cause the entire RAID volume to fail.

RAID-1, also called mirroring, writes the same data to two or more disks. A RAID-1 configuration has the total capacity of one drive, regardless of how many drives are a part of the configuration. If three two-terabyte hard disk drives are placed in RAID-1, the total size of the RAID volume will be two-terabytes.

RAID-1 is all about redundancy. If one or more of the disks fails, the RAID will continue to operate as long as one disk in the RAID is healthy. When two or more disks are healthy, the storage controller can also exploit the ability to read from two or more disks at once, resulting in tremendous read performance.

RAID-10 is a combination of striping and mirroring. RAID-10 requires a minimum of four hard disk drives, and provides the capacity of two. Data striped across two sets of two disks.

RAID-5, also called distributed block-level parity, requires three or more disks and provides the capacity of n-1 disks (where n is the number of disks in the RAID). RAID-5 divides the disks into fixed size blocks. For each n blocks, n-1 are for data and 1 is for parity. The parity blocks are distributed across each of the disks in a rotating fashion (RAID-4 is simply RAID-5 with a dedicated parity disk), so each disk contains the same number of data and parity blocks. Whenever a data block is written, the corresponding parity block is recalculated and overwritten. Thus, writing any data to a RAID-5 volume requires not only overwriting the data block, but also recomputing the parity and rewriting the parity. RAID-5 incurs a substantial computational and IO overhead, so write performance is usually below that of a single disk. The advantage of RAID-5 is that if a disk fails, the RAID can be rebuilt. For each row of n blocks, if the parity block fails it can be recomputed from the data blocks; if a data block fails, it can be recomputed from the parity block and the other data blocks. This allows a RAID-5 configuration to suffer a single disk failure.

RAID-6 is an extension of RAID-5 that uses two parity blocks, providing the size of n-2 disks, and allowing for up to two disk failures at once.
 
Upvote 0 Downvote
If you are not a friend of Google search engine.
Here:
http://en.wikipedia.org/wiki/RAID
Click to expand...
I can't understand in any online sites and so I'm asking in Tom's Forums.

RAID is an acryonym that stands for Redundant Array of Independent/Inexpensive Disks.
RAIDs allow for multiple disks to work together. The exact benefit of the raid, as well as any drawbacks, depend on the particular configuration.

RAID-0, also called striping, interleaves blocks of data across two or more disks. A RAID-0 configuration has the total capacity of all the physical volumes that are a part of the configuration. If three two-terabyte hard disk drives are placed in RAID-0, the total size of the RAID volume will be six terabytes.

RAID-0 takes advantage of spatial locality. When data is accessed, there is a high probability that data that is logically nearby will also be accessed. By ping-ponging logically sequential blocks of data across two or more identical disks, the storage controller can put all of them to work at the same time. This allows data to be read faster, and written faster, than it would be if a file were contained on only a single hard disk drive. The drawback of RAID-0 is that a failure in any of the disks in the RAID will cause the entire RAID volume to fail.

RAID-1, also called mirroring, writes the same data to two or more disks. A RAID-1 configuration has the total capacity of one drive, regardless of how many drives are a part of the configuration. If three two-terabyte hard disk drives are placed in RAID-1, the total size of the RAID volume will be two-terabytes.

RAID-1 is all about redundancy. If one or more of the disks fails, the RAID will continue to operate as long as one disk in the RAID is healthy. When two or more disks are healthy, the storage controller can also exploit the ability to read from two or more disks at once, resulting in tremendous read performance.

RAID-10 is a combination of striping and mirroring. RAID-10 requires a minimum of four hard disk drives, and provides the capacity of two. Data striped across two sets of two disks.

RAID-5, also called distributed block-level parity, requires three or more disks and provides the capacity of n-1 disks (where n is the number of disks in the RAID). RAID-5 divides the disks into fixed size blocks. For each n blocks, n-1 are for data and 1 is for parity. The parity blocks are distributed across each of the disks in a rotating fashion (RAID-4 is simply RAID-5 with a dedicated parity disk), so each disk contains the same number of data and parity blocks. Whenever a data block is written, the corresponding parity block is recalculated and overwritten. Thus, writing any data to a RAID-5 volume requires not only overwriting the data block, but also recomputing the parity and rewriting the parity. RAID-5 incurs a substantial computational and IO overhead, so write performance is usually below that of a single disk. The advantage of RAID-5 is that if a disk fails, the RAID can be rebuilt. For each row of n blocks, if the parity block fails it can be recomputed from the data blocks; if a data block fails, it can be recomputed from the parity block and the other data blocks. This allows a RAID-5 configuration to suffer a single disk failure.

RAID-6 is an extension of RAID-5 that uses two parity blocks, providing the size of n-2 disks, and allowing for up to two disk failures at once.
Click to expand...
Pinhedd, I understood clearly now. To keep it in my memory for a long time, can you say some real-world examples for all the RAIDs?
For example, a professional photo/audio/video editor uses RAID-1 so that if one of his drive fails he still has his files stored on all other drives connected in RAID-1.
 
Upvote 0 Downvote
RAID is not a backup. It will not save you from viruses or accidental file deletions. RAID is for redundancy so that if a drive fails, the system can continue to operate normally. Businesses that require 24/7 uptime usually use RAID. They also make backups in case of critical failures such as two or or more drives in a RAID fail.
 
Upvote 1 Downvote


That's a good example. RAID-1 is fault tolerant. As Hawkeye said, it is not quite the same thing as a backup because RAIDs operate at the block level. The RAID controller doesn't care if the data written to the RAID is intentional, it has no knowledge of the filesystem or the contents. If the professional accidentally deletes or overwrites one of his videos, the file will be erased or replaced on both drives. The purpose of RAID-1 is largely to ensure that the contents of the RAID will be preserved in the event of a single disk failure.

RAID-0 on the other hand is typically used by individuals who wish to obtain more drive performance and do not particularly care about potential data loss. Gamers are especially fond of RAID-0 because it reduces loading times in games.

RAID-5 is heavily used in business and enterprise where the cost of RAID-1 would be prohibitive. Given the disadvantages of RAID-5 that I listed above, many RAID-5 configurations will be deployed on dedicated hardware RAID controllers that can accelerate the recalculation of the parity blocks and use high speed enterprise hard disks.
 
Upvote 0 Downvote
Solution


No, not always.
With the advent of reliable consumer grade SSD, RAID 0 is becoming less and less beneficial.

As said, an SSD is faster than RAID 0 HHD, so why go through the hassle and increased fail potential?

2 x 2TB in RAID 0 is exactly the same drive size as 2 x 2TB individual drives.
 
Upvote 1 Downvote


That depends on the RAID controller. Some RAID controllers, such as Intel's firmware RAID, support matrix RAID which allows a disk to be a member of two or more different RAIDs.

RAIDs are bound by the smallest physical member volume. For example, a pair of 500GB HDDs and a 1TB HDD can be matrixed to form two RAID-0 volumes of 1TB each by splitting the 1TB HDD into two 500GB volumes. You could also form two 500GB RAID-1 volumes using the same process. The drawback of doing this though is that the IO performance may be limited by the shared disk, but if both RAID volumes aren't used at the same time this should not be noticeable.

As for RAIDs vs SSDs...

SSDs are excellent for storing games that stream assets poorly. These are games that load textures, sounds, and models into memory when that asset is needed rather than when the level is loaded. For games that handle streaming well, or preload as much as possible, RAID-0 is a good alternative as it provides a very good amount of storage space for the same price.
 
Upvote 0 Downvote
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom