Hong Life CalculatorsUnderstanding RAID capacity — concept and principle
RAID (Redundant Array of Independent Disks) combines several physical drives into one logical volume. Depending on the level, it spreads data across disks for speed, stores extra parity or mirror copies for protection, or both. The protection is paid for in space: some of the raw capacity is used to store redundancy instead of your files. This calculator shows how much space you actually get and how many disks can fail before data is lost.
The principle: striping, parity and mirroring
Data is cut into blocks and written across the disks in rows called stripes. In RAID 5 each stripe has one parity block, and its position rotates from disk to disk so no single drive becomes a bottleneck. Parity is the XOR of the data blocks in that stripe: if any one disk disappears, XOR-ing the survivors gives back exactly the missing bits. RAID 6 stores two independent parity blocks per stripe (P and Q), so any two disks may fail. RAID 1 and RAID 10 skip parity and simply keep identical copies.
The formulas
| Level | Usable capacity | Disk failures tolerated | Min. disks |
|---|---|---|---|
| RAID 0 | n × size | 0 | 2 |
| RAID 1 | 1 × size | n − 1 | 2 |
| RAID 5 | (n − 1) × size | 1 | 3 |
| RAID 6 | (n − 2) × size | 2 | 4 |
| RAID 10 | ⌊n ÷ 2⌋ × size | 1 per mirror pair | 4 |
Efficiency is usable ÷ raw. It grows with more disks for RAID 5 and 6 — 3 disks in RAID 5 give 67%, 10 disks give 90% — while RAID 1 and RAID 10 stay fixed.
Worked example, step by step
- Raw capacity: 6 disks × 8 TB = 48 TB
- RAID 5 reserves one disk’s worth for parity: (6 − 1) × 8 = 40 TB usable
- Efficiency: 40 ÷ 48 = 83.3%
- In binary units: 40 × 10¹² ÷ 2⁴⁰ = 36.38 TiB — what Windows or Linux will report
- Fault tolerance: any 1 disk may fail; a second failure before the rebuild finishes loses the array
Common mistakes
- Forgetting the TB → TiB difference and thinking 3.6 TB “went missing”.
- Choosing RAID 5 for many large disks — the long rebuild window is the risky part.
- Counting a hot spare as usable space.
- Using an odd number of disks for RAID 10 — the last disk is left unused.
Frequently asked questions
Is RAID a backup?
No. RAID protects against a disk dying, not against deletion, ransomware, fire, theft or a controller bug — all of which hit every disk at once. Keep a separate copy (the 3-2-1 rule: 3 copies, 2 media, 1 off-site).
Can I mix disks of different sizes?
Classic RAID treats every disk as the size of the smallest one, so mixing 8 TB and 12 TB wastes 4 TB per larger disk. Some NAS systems (e.g. hybrid RAID modes) work around this; this calculator assumes equal disks.
RAID 5 or RAID 6 for large disks?
With disks of roughly 8 TB and up, a rebuild can take a day or more, and a second failure or unreadable sector during that time loses a RAID 5 array. RAID 6 (or RAID 10) is the common recommendation for big drives and arrays of 6+ disks.
Why is RAID 10 capacity always half?
Every block is written to two disks (a mirror pair), and the pairs are striped together. You get twice the write safety and fast rebuilds, but only 50% of raw space.
How many disks does each level need?
RAID 0: 2+, RAID 1: 2+, RAID 5: 3+, RAID 6: 4+, RAID 10: 4+ in an even number.
Sources: Patterson, Gibson & Katz, “A Case for Redundant Arrays of Inexpensive Disks (RAID)”, ACM SIGMOD 1988 · SNIA Common RAID Disk Data Format (DDF) Specification · IEC 80000-13 (binary prefixes).