The disks used for data storage are mechanical parts which are susceptible to failures. When any disk fails, all data stored in the disk is lost. This becomes a serious issue when critical data is stored in the disk. RAID stands for redundant array of independent disks. Redundancy is the key behind RAID where data is stored in multiple disks to prevent data loss. RAID is now used to increase the speed & performance of data storage & retrieval method. It is clear that retrieving data sequentially from a single disk is slower than retrieving the same data from parallel disks. To control the operation, host adapter in RAID provides centralized control.
Multiple RAID levels named as RAID0 to RAID5 define the way the storage & retrieval of data is done. Each of these levels concentrates on minimizing data loss & increasing the performance. However, either of these goals dominates RAID levels because both the qualities are mutually exclusive.
In RAID, multiple disks are used for storing data. When a write operation is initiated, the host adapter accesses multiple connected drives to write the data. The data to be written is split into blocks. Odd blocks are written in to disk & even blocks to the other if there is disks. When a read operation is initiated, the host adapter fetches data from multiple disks & stores in the buffer. The RAID levels are used to increase reliability & performance.
RAID 0 strips the data & writes data fragments to multiple disks on the same sector. Even if of the disks fails, the stored data will be lost. RAID 0 is used where speed is the major concern. RAID 1 makes use of mirroring to minimize data loss. The same data is written on multiple disks to be positive that the data is still obtainable even if of the disks fails. The probability of failure of all the disks simultaneously is less & hence, you can expect the data to be present on a regular basis. You ought to buy more number of disks in the event you need to increase the storage capacity with RAID1.
 |
| The RAID 0 and 1 |
RAID 2 makes use of error correction code to minimize data loss. Apart from storage disks, ECC disks are used in this RAID level. The ECC disks are dedicated to store error correction codes. The data to be stored is split across multiple disks & the correction codes are directed to the ECC disk. RAID 3 makes use of byte level stripping with parity in lieu of error correction codes. RAID 4 is similar to RAID 3 but block level stripping of data is done here. The correction code is based on parity.
 |
| RAID 2 |
With levels 2, 3 & 4, failure of a single ECC disk ends in loss of whole data. This issue is eliminated in RAID 5. RAID 5 also makes use of block parity but the parity data are also distributed across multiple disks. The parity can be used for data recovery meaning that the data can be recovered even when a single disk fails.
 |
| RAID 3 |
 |
| RAID 4 |
 |
| RAID 5 |
RAID 6 is fundamentally an extension of RAID level which allows for additional fault tolerance by using a second independent distributed parity process (two-dimensional parity)
Data is striped on a block level across a set of drives, like in RAID, and a second set of parity is calculated & written across all the drives; RAID 6 provides for an high data fault tolerance & can maintain multiple simultaneous drive failures.
Data is striped on a block level across a set of drives, like in RAID, and a second set of parity is calculated & written across all the drives; RAID 6 provides for an high data fault tolerance & can maintain multiple simultaneous drive failures.
0 komentar:
Posting Komentar