CSE Roadmap

Specialist tier · about 2 hours · 4 parts

Part 2 of 4 · 1 min read

RAID, in outline

Combining multiple disks for either speed or protection against failure.

Think of a warehouse splitting its inventory across several smaller storage rooms. This lets several workers restock different rooms at once. It can also keep a second full copy in case one room floods.

In plain words: RAID combines multiple physical disks into one logical storage unit. Different RAID levels trade off speed, usable capacity, and protection against a disk failing. RAID 0 stripes data across disks for speed, with no redundancy at all. RAID 1 mirrors data across disks for protection, at the cost of half the usable capacity. RAID 5 stripes data with added parity information, tolerating one disk's failure without losing data.

A worked example. A database server storing critical order data might use RAID 1 or RAID 5, accepting less raw capacity or added parity overhead, specifically to survive one disk failing without losing any data.

The key lines. No SQL; RAID configuration is a storage-infrastructure decision, made outside the database itself.

RAID, striping, mirroring, parity

RAID combines multiple disks into one logical unit. Striping spreads data across disks for speed. Mirroring duplicates data across disks for protection. Parity stores extra, calculated information letting one disk's data be reconstructed if it fails.

Example: RAID 0 uses only striping; RAID 1 uses only mirroring; RAID 5 uses striping plus parity, balancing speed, capacity and some protection.