7.10.4 - 7.10.5 RAID through Culstering

Redundant Array of Independent Disks (RAID)

Data storage technology that combines an array of physical disk drives into one logical unit for the purposes of data redundancy, performance improvement, or both.

CH: Trade-off mechanism between availability, performance, and cost.

RAID 0

AKA Striping. Requires 2 or more drives. Does NOT provide redundancy of data. Purpose is solely performance. Data is split and written across drives simultaneously.

CH: High performance, zero fault tolerance.

Striping

RAID storage method where data sent to the RAID controller is split and fragments are written across multiple drives to improve performance.

CH: Performance gain comes at the cost of resilience.

RAID 1

AKA Mirroring. Requires 2 or more drives. Provides redundancy of data by writing full copies to each drive.

CH: Simple redundancy, fast recovery.

RAID 10

Combination of RAID 0 and RAID 1. Requires 4 or more drives. Provides redundancy and performance but uses many disks. Can lose up to 1 drive.

CH: Performance plus fault tolerance, high cost.

RAID 5

Requires 3 or more drives. Data is striped for performance and parity is distributed to allow recovery from a single disk failure.

CH: Balance of cost, performance, and availability.

RAID 6

Similar to RAID 5 but stores additional parity allowing two disk failures without data loss. 4 drives minimum.

CH: Higher fault tolerance for large arrays.

Parity Bits

Provide fault tolerance by XORing data across drives so lost data can be recomputed if a drive fails.

CH: Mathematical recovery mechanism.

Clustering

A group of multiple systems working together to support a workload.

CH: Improves availability and scalability.

Redundancy

The presence of additional systems or components kept in standby to take over if a primary system fails, ensuring continued availability.

CH: Eliminates single points of failure.