RAID (Redundant Array of Independent Disks)

Definition

RAID is a data storage virtualization technology that combines multiple physical disk drives into a single logical unit for performance, redundancy, or both. The term was originally “Redundant Array of Inexpensive Disks,” later changed to “Independent.”

RAID levels define how data is distributed across disks, with different trade-offs between performance, capacity, and fault tolerance.

Common RAID Levels

Level Name Min Disks Redundancy Performance Use Case
0 Striping 2+ None Read/Write fastest Performance, no fault tolerance
1 Mirroring 2 1 disk Read fast, Write same Small volumes, boot drives
5 Striping + Parity 3+ 1 disk Read fast, Write slower General-purpose storage
6 Striping + Dual Parity 4+ 2 disks Read fast, Write slower Critical data, 2-disk failure tolerance
10 Mirror + Stripe 4+ Half disks Read/Write fast High-performance, redundant storage
50 Stripe + Stripe 6+ 1 per stripe High performance Large-scale storage arrays

Key Concepts

  • Hot spare: Unused disk that automatically rebuilds if another fails
  • Rebuild: Reconstructing data on replacement disk after failure
  • ZFS/Btrfs: Software RAID alternatives with additional features (snapshots, compression)
  • Hardware RAID: Dedicated RAID controller card
  • Software RAID: OS-level implementation (Linux mdadm, ZFS)
  • Nas — different from RAID; RAID protects against disk failure, not data loss

References