RAID
• Redundant Array of Independent Disks
• Redundant Array of Inexpensive Disks
• 7 levels numbered 0 through 6 – does not imply a hierarchical relationship
• Set of physical disk drives viewed by the OS as a single logical drive
• Data are distributed across the physical drives of an array
• Redundant disk capacity used to store parity information – guarantees data
recoverable
RAID Level 0 - Striping
• Data striped across all disks
• Uses Round Robin striping
o Multiple data requests found on different disks
o Multiple disks can be seeked in parallel
o A set of data is likely to be striped across multiple disks
RAID Level 1 - Mirroring
• Mirrored disks
• Data is striped across disks
• 2 copies of each stripe exist on 2 separate disks
• Read request handled by either disk – minimum seek plus rotation latency
• Write requests go to both disks
• Recovery from failure is simple – swap out bad disk and re-mirror
• Expensive – requires twice the disk space
RAID Level 2 – Parallel access
• Disks are synchronized – heads in same position on every disk
• Stripes are very small – single byte or word
• Error-correcting code calculated across corresponding bits on each disk
• Multiple parity disks store the code in corresponding positions – Hamming code
• Too much redundancy – expensive – overkill so not implemented
RAID Level 3 – Parallel access
• Similar to RAID 2
• Only a single redundant disk – no matter what the array size is
• Simple parity bit is computed for the set of individual bits in the same position
• Data on a failed disk reconstructed from surviving data and parity info
• Very high transfer rates – small stripes yield parallel transfer from all disks
RAID Level 4 – Independent access
• Each disk operates independently
• High I/O request rates – separate I/O request satisfied in parallel
• Large stripes are used
• Bit-by-bit parity strip is calculated across corresponding strips on each disk
• Parity strips stored in the corresponding strip on the parity disk
• Every write must involve a parity disk – potential I/O bottleneck
RAID Level 5 – Independent access
• Similar to RAID 4
• Distributes the parity strips across all disks
• Uses Round Robin allocation scheme
• Avoids RAID 4 I/O bottleneck
• Commonly used in network servers
RAID Level 6 – Independent access
• Two parity calculations
• Stored in separate blocks on different disks
• Requires N+2 disks – where N is number of disk required for user data
• Extremely high data availability
o Three disk must fail before loss of data
o Substantial write penalty – affects two parity blocks
posted by; Hazyanty Shahira Abd. Halim
0 comments:
Post a Comment