Redundant Array of Inexpensive Disks- RAID

Definition of RAID

The RAID array is a system that stores information in a number of disks (n), so that the process speeds up disk access. By placing data on multiple disks, operations I / O (input / output, input and output) can overlap in a balanced way giving better performance and system performance(Gupta: p. 79).

The RAID employs the technique known as "striping" (banding or striping) including the partition of the storage space on each disk in units ranging from a sector (512 bytes) to several megabytes.

Due to the decrease in the price of hard drives and the increased availability of RAID options included in the chipsets of motherboards , RAIDs are also an option in personal computers more advanced(Donald: p. 125). This feature is obviously not available in software RAID systems, which tend to be therefore the problem of reconstructing the set of disks when the system is restarted after a failure to ensure data integrity. By contrast, software-based systems are much more flexible (allowing, for example, build RAID partitions instead of full albums and grouped into a single RAID disks connected to multiple controllers) based hardware and add a point of failure more to the system (RAID controller).

RAID Implementations RAID 0, RAID 1, RAID 5, RAID 1+0, RAID 0+1

RAID-0

This arrangement uses a technique called "striping", which distributes the information in blocks between the different disks. It is the only RAID level that no duplicate information, therefore no storage capacity is wasted, so that no data redundancy(Delmar: p. 448). It requires at least two disks. It offers the best performance but low fault tolerance.

RAID-1

This type is also known as mirror disc creation and formed of at least two drives that duplicate the storage of data. No banding. Reading performance is enhanced either because readable discs simultaneously. The write performance is the same as when stored in a single disc. It provides the best performance and the best fault tolerance in a multi-user system.

RAID-5

RAID 5 stores parity information but not redundant data (but parity information can be used to reconstruct data) (Patterson: pp 604-605). The RAID-5 requires at least three and usually five disks in the set. It's better for multiuser systems where performance is not critical or performing few write operations.

RAID 1+0

It is a combination of levels 1 and 0. The disks are grouped together in clusters according to RAID 1. These clusters are then grouped according to level 0. This technology requires a minimum of 4 drives. It should not be confused with RAID 01 (RAID clusters are gathered by 0 RAID 1), which is less interesting in terms of reliability because it cannot handle multiple failures. (Gupta: p. 79).

This level allows to combine dependability and performance. Reliability is particularly interesting because it requires that all the elements of a cluster are defective RAID 1 for touch operation. Data reconstruction is efficient because it involves hard one cluster. RAID 10 can cope with multiple hard disk failure (Patterson: pp 604-605).

RAID 0 +1

As shown ...