RAID Storage
Redundant arrays of independent disks (RAID) are used to overcome
two bottlenecks associated with disk storage:
By using additional redundant disks we can increase both.
The following are some typically identified RAID levels:
- 0: Nonredundant
- Data is distributed across different disks to increase performance
- 1: Mirrored
- A second disk set keeps a copy of the data (50% overhead)
- 0+1 (or 10): Striping and mirroring
- Data is distributed across a second disk set to increase performance
- 2: Error correcting codes
- Additional check disks (4:3, 10:4, 25:5) are used to provide redundancy
with a smaller overhead (e.g. 57%, 71%, 83%)
- 3: Bit-interleaved parity
- A single additional check disk is used to recover the data by distributing bits across all disks
- 4: block-interleaved parity
- The check disk contains the parity on a block level: higher read throughput
- 5: block-interleaved distributed parity
- Block parity is distributed across all disks: higher read and write throughput
- 6: higher redundancy
- Like 5 with an additional check disk guarding against a second failure