Create New Article

Wiki Search

 
 
Google
Personal tools

Understanding RAID levels

From Proprofs

Understanding RAID levels

Contents


You are here: Home > Schools > Comptia > Comptia A+ > Wiki Home > A+ Certification Wiki

[edit section] Getting to Know your Raid levels

RAID 0
RAID 0 was not defined in the 1987 Berkeley paper. In fact, it is not RAID because it does not provide any redundancy. RAID 0 is just an array or group of disk drives used as a single disk. The data is written in chunks or stripes to all the disk drives in the array. This improves disk input and output performance because several chunks of data can be written or read simultaneously. If a disk drive in the RAID 0 array fails, all data in the RAID 0 array is lost. RAID level 0 is also often called disk striping without parity. Figure shows an illustration of RAID 0.


RAID 1
RAID 1 requires a minimum of two disk drives. All other RAID levels, except level 0, require at least three disk drives to implement. RAID 1 writes all data to two separate locations. To store 20 gigabytes (GB) of data using RAID 1, two 20-GB disk drives are required. This is a 50 percent loss of storage capacity.

There are two ways to implement RAID 1:

  • Disk mirroring
  • Disk duplexing


In disk mirroring, the two disk drives are connected to the same disk controller. The only problem with disk mirroring is that if the disk controller fails, there is no access to the mirrored data. Figure shows a diagram of disk mirroring. To eliminate this single point of failure, use disk duplexing rather than disk mirroring.


In disk duplexing, each disk drive in the mirrored set is connected to a different disk controller. This eliminates the single point of failure in pure disk mirroring. The only additional cost is the additional disk controller. Figure shows a diagram of disk duplexing.


RAID 2
RAID 2 uses a hamming code to create an error correcting code (ECC) for all data to be stored on the RAID 2 array. The ECC can detect and correct single-bit errors and detect double-bit errors. The ECC code has to be read and decoded each time data is read from the disk. RAID 2 is very difficult and expensive to implement and has a very high overhead. For example, there are three parity bits for each four data bits.


NOTE:
A hamming code is an error correction method that mixes three check bits at the end of each four data bits. When these check bits are received, they are used to detect and correct one-bit errors automatically.


RAID 2 has no commercial implementations because of the expense and difficulty of implementation. It requires a minimum of three disk drives to implement.


RAID 3
RAID 3 uses bit-level parity with a single-parity disk to provide fault tolerance of data stored on the RAID 3 array in the event of failure of a single disk drive in the array. RAID 3 requires that all the disk drives in the array be synchronized with each other. The bits of the data and the parity information calculated from the data are written to all the disk drives in the array simultaneously. RAID 3 requires a minimum of three disk drives to create the array.


RAID 4
RAID 4 uses block-level parity with a single-parity disk to provide fault tolerance to the RAID 4 array in the event of failure of a single disk drive in the array. On a RAID 4 array, data and the parity information calculated from the data is written to the disk drives in blocks. There is no need for the disk drives to be synchronized together, and the disk drives can be accessed independently. A minimum of three disk drives is required to create the array. The problem with RAID 4 is that the parity drive is accessed on every write operation to the RAID array. This will cause heavy utilization of the parity drive, which will probably fail before the other drives in the array.


RAID 5
RAID 5 uses block-level parity, but it spreads the parity information among all the disk drives in the disk array. This eliminates the parity drive failure common in RAID 4 systems. The loss of storage capacity in RAID 5 systems is equivalent to the storage capacity of one of the disk drives. If there are three 10-GB disk drives in a RAID 5 array, the storage capacity of the array will be 20 GB, which is a loss of one-third, or 33 percent. In another example, if there are seven 10-GB disk drives in a RAID 5 array, the total storage capacity of the array will be 60 GB, which is a loss of one-sixth, or 16.67 percent. Figure shows a diagram of RAID 5.


RAID 0/1
RAID 0/1 is also known as RAID 0+1, and it is sometimes called RAID 10. This combination of RAIDs provides the best of both worlds. It has the performance of RAID 0 and the redundancy of RAID 1. RAID 0/1 requires at least four disk drives to implement. In RAID 0/1, there are two RAID 0 stripe sets, which are used to provide high input/output performance, that are mirrored. This provides the fault tolerance. Figure shows a diagram of RAID 0/1.

[edit section] Acknowledgements

Created & shared by Eric

Top 5 Contributors to this article

UsersArticle Contributions
Jbrown 2 contribs

 
   
Home  |  Site Map  |  Contact
Copyright © 2005-2014 ProProfs.com - Privacy & Terms