Skip to main content
Definition Website Computer and Technology

Redundant Array of Independent Disks

Subject: Computer and Technology noun /rɪˈdʌndənt əˈreɪ əv ˌɪndɪˈpɛndənt dɪsks/ 1 definition

  1. Definition (noun)

    Entry #219 · Updated

    A Redundant Array of Independent Disks (RAID) is a data storage virtualization technology that combines multiple physical drives into one or more logical units to provide data redundancy, higher I/O performance, or both. Depending on the selected RAID level, a hardware controller or software layer distributes data across the member drives through striping, mirroring, or parity calculations while presenting the array to the host operating system as a unified volume. While redundant configurations keep data accessible during individual drive failures, RAID only protects against physical disk faults and does not replace off-system backups against accidental deletion, data corruption, or catastrophic hardware loss.

    Common Types

    • RAID 0 (Striping): Splits data blocks evenly across two or more drives to maximize read and write throughput, utilizing full combined capacity with zero fault tolerance if any drive fails.

    • RAID 1 (Mirroring): Writes an exact duplicate of all data across two or more drives, allowing immediate survival of drive failures and faster parallel reads while limiting usable capacity to the size of a single member drive.

    • RAID 5 and RAID 6 (Distributed Parity): Stripes data along with rotating parity blocks across at least three (RAID 5) or four (RAID 6) drives, tolerating one or two simultaneous drive failures respectively with higher storage efficiency than mirroring.

    • RAID 10 (Nested Mirroring and Striping): Stripes data across mirrored drive pairs using at least four drives, combining high random I/O performance with the ability to survive multiple drive failures provided no single mirror pair loses all its drives.

    Example

    A database administrator configures a production PostgreSQL server with eight 4 TB NVMe drives in a RAID 10 array, yielding 16 TB of usable high-IOPS storage. When one drive fails during peak traffic, the array continues serving queries from its surviving mirror while automatically rebuilding onto a hot spare with zero downtime.