💾

RAID Storage Overview

Jun 6, 2025

Overview

This lecture covers RAID (Redundant Array of Independent Disks), explaining different RAID levels, their redundancy, and performance features for data storage reliability.

Importance of Storage Redundancy

  • Hard drives and SSDs can fail, risking data loss.
  • Combining drives can create redundancy to protect data availability.
  • RAID is not a backup solution; a separate backup process is still needed.

RAID Concepts and Levels

  • RAID stands for Redundant Array of Independent (or Inexpensive) Disks.
  • RAID levels differ in redundancy and performance characteristics.
  • Common RAID types include RAID 0, RAID 1, RAID 5, RAID 6, and RAID 10.

RAID 0 (Striping)

  • Requires at least two drives; splits data across all drives.
  • Improves speed by writing data to multiple drives at once.
  • Provides no redundancy; losing one drive causes total data loss.

RAID 1 (Mirroring)

  • Uses at least two drives; duplicates data on each drive (mirroring).
  • Doubles storage need but protects data if one drive fails.
  • Keeps data available during a drive failure until the mirror is restored.

RAID 5 (Striping with Parity)

  • Needs at least three drives; stripes data and stores parity for redundancy.
  • Parity is distributed across drives, not duplicated data.
  • Allows one drive to fail without data loss, but recovery uses more CPU.

RAID 6 (Striping with Double Parity)

  • Similar to RAID 5 but adds an extra disk for a second parity block.
  • Can withstand failure of up to two drives.
  • No extra storage capacity from added drive; used only for additional parity.

RAID 10 (Nested RAID: RAID 1+0)

  • Combines striping (RAID 0) and mirroring (RAID 1).
  • Requires at least four drives; stripes are mirrored for high performance and redundancy.
  • Can lose one drive in each mirrored pair without losing data.

Key Terms & Definitions

  • Redundancy — extra components ensuring system operation if one part fails.
  • RAID (Redundant Array of Independent Disks) — method of combining multiple drives for data redundancy/performance.
  • Striping — splitting data into blocks and spreading them across multiple drives.
  • Mirroring — duplicating data exactly on two or more drives.
  • Parity — calculated data used to rebuild lost information after a drive failure.
  • Nested RAID — combining different RAID levels, e.g., RAID 10 (striped mirrors).

Action Items / Next Steps

  • Remember: implement a separate backup routine, as RAID alone does not prevent data loss from all failure types.

Full transcript