Storage

RAID Calculator

See how much disk space remains for data after mirroring or parity. Enter an equal-capacity disk count, size, and RAID level to calculate raw and usable storage, redundancy overhead, and disk-failure tolerance. Results update instantly in your browser. This estimate excludes formatting, system reservations, and spare drives, and it does not replace a backup plan.

Your RAID capacity

Raw storage capacity: 16 TB16,000 GB
Usable storage capacity: 12 TB12,000 GB
Storage used by redundancy/parity: 4 TB4,000 GB
Storage efficiency: 75%
Disk failure tolerance: 1 disk.
Minimum disks for RAID 5: 3

RAID is not a backup. These theoretical capacities exclude formatting, snapshots, system reservations, and hot spares. Disk-failure tolerance assumes an otherwise healthy array and readable surviving data.

How RAID capacity and redundancy work

RAID combines physical disks into an array. Striping distributes data across disks; mirroring stores another copy; parity stores information used to reconstruct data after a supported disk failure. The chosen layout changes both usable space and how failures affect the array. This calculator models conventional RAID 0, 1, 5, 6, and two-way-mirror RAID 10, not vendor-specific hybrid storage layouts.

Let N be the disk count and C the capacity of one disk. Raw capacity = N × C. Usable capacity is N × C for RAID 0, C for classic two-disk RAID 1, (N − 1) × C for RAID 5, (N − 2) × C for RAID 6, and (N ÷ 2) × C for RAID 10. Redundancy overhead = raw − usable. Efficiency = usable ÷ raw × 100. GB and TB use decimal units: 1 TB = 1,000 GB. They are not GiB or TiB.

Raw capacity is the sum of the participating disks before redundancy. Usable capacity here is the theoretical space after RAID overhead, not the exact size of a formatted volume. Filesystem metadata, snapshots, controller reservations, and operating-system partitions can reduce what you can store. Do not count a dedicated hot spare as an active disk in these formulas.

RAID 0 versus RAID 1: striping provides all modeled capacity but tolerates no disk failures; a failed member can make the array's data inaccessible. A classic two-disk mirror uses half of the raw space and can tolerate one disk failure. This version rejects RAID 1 arrays with more than two disks instead of assuming a multi-way mirror or a controller-specific arrangement.

RAID 5 versus RAID 6: RAID 5 reserves one disk-equivalent for distributed parity and tolerates one failed disk. RAID 6 reserves two disk-equivalents and tolerates two failed disks. Parity is spread across the array rather than necessarily placed on dedicated parity drives. Minimums in this calculator are three disks for RAID 5 and four for RAID 6; your controller may impose additional limits.

RAID 10 stripes across mirrored pairs and uses half the raw capacity. It tolerates any single disk failure. It can survive additional failures only while every pair retains a working disk: up to N ÷ 2 failures if each is in a different pair. Losing both members of one pair can lose the array even if all other disks work. Mirrored rebuilds copy from a surviving partner, and RAID 10 can suit write-intensive workloads, but performance depends on the hardware and workload.

Redundancy is not backup. RAID can preserve availability during some disk failures, but it does not provide an independent recovery copy against accidental deletion, ransomware, corruption, theft, or loss of the whole system. Keep separate backups, retain recovery history suited to your needs, and test restoration. A redundant live copy is not a substitute for a recoverable historical copy.

A degraded array needs attention. Rebuilding reads surviving disks and writes a replacement while the system may still be serving requests. Duration depends on drive size and speed, workload, controller behavior, and recovery settings; this calculator does not predict it. Another failure or an unreadable sector can complicate recovery. Replace failed members according to the device instructions and verify backups before maintenance.

Version 1 assumes all disks have equal capacity. Conventional equal-member RAID layouts generally limit each member's contribution to the smallest disk, leaving extra capacity on larger disks unused in that array. For example, treating mixed 4 TB and 8 TB members as 4 TB each can estimate a conventional layout, but it is not a model of all hybrid NAS schemes. Matched capacities simplify planning; check the vendor's compatibility and expansion rules as well.

When planning a recording system, first estimate the required recording space using the related CCTV Storage Calculator, then compare that target with RAID usable capacity. RAID redundancy overhead and the recording tool's planning allowance serve different purposes. Leave room for the filesystem and operational free space rather than filling the theoretical volume to its last byte.

Examples

2 × 4 TB in RAID 1

Input: Two equal 4 TB disks, classic RAID 1 mirror.
Result: 8 TB raw; 4 TB usable; 4 TB used by redundancy; 50% efficiency. Tolerates one failed disk. Exactly two disks are supported for RAID 1 in this calculator.

4 × 4 TB in RAID 5

Input: Four equal 4 TB disks, RAID 5.
Result: 16 TB raw; 12 TB usable; 4 TB parity overhead; 75% efficiency. Tolerates one failed disk.

4 × 4 TB in RAID 6

Input: Four equal 4 TB disks, RAID 6.
Result: 16 TB raw; 8 TB usable; 8 TB parity overhead; 50% efficiency. Tolerates two failed disks.

4 × 4 TB in RAID 10

Input: Four equal 4 TB disks, arranged as two mirrored pairs.
Result: 16 TB raw; 8 TB usable; 8 TB mirroring overhead; 50% efficiency. Any one disk may fail. Two failures are survivable only if they affect different mirrored pairs.

8 × 8 TB in RAID 6

Input: Eight equal 8 TB disks, RAID 6.
Result: 64 TB raw; 48 TB usable; 16 TB parity overhead; 75% efficiency. Tolerates two failed disks. This does not include filesystem reservations or a hot spare.

Frequently asked questions

What RAID level gives the most usable storage?

Among the levels modeled here, RAID 0 uses all raw capacity before formatting. It provides no disk-failure redundancy. Choose based on recovery and availability requirements as well as usable space; capacity alone is not a protection strategy.

Is RAID 5 safe?

RAID 5 tolerates one disk failure in the modeled array, but that is not a promise of data safety. During a degraded period, a further failure can lose the array, and read errors can disrupt recovery. Evaluate rebuild exposure, device guidance, and backups rather than relying on the RAID name alone.

What is the difference between RAID 5 and RAID 6?

RAID 5 uses one disk-equivalent of distributed parity and tolerates one failed disk. RAID 6 uses two disk-equivalents and tolerates two failures, so it has less usable space for the same disks. Both still require separate backups.

How much storage do I lose with RAID?

RAID 0 has no redundancy overhead. Two-disk RAID 1 and paired RAID 10 reserve half the raw space for mirrors. RAID 5 reserves one disk's capacity and RAID 6 reserves two. This is capacity devoted to redundancy, not defective or missing storage; formatting overhead is separate.

Does RAID replace backup?

No. Deletion or corruption can affect data throughout the live array, and a whole-system failure can affect every member. Backups should provide separate, recoverable copies or versions appropriate to your recovery requirements.

Can I use disks of different sizes?

Some systems permit it, but conventional RAID commonly uses only the smallest member capacity from each disk. Extra space may be unavailable. This calculator assumes equal sizes; hybrid storage systems can behave differently, so use their vendor-specific capacity guidance.

How many disks do I need for RAID 10?

This calculator requires at least four disks and an even total, grouped into two-disk mirrors. Any single failure is tolerated. Multiple failures are tolerated only if no mirrored pair loses both disks; N ÷ 2 is a conditional maximum, not a guarantee for arbitrary failures.

Which RAID level is best for NAS storage?

There is no universal choice. Compare the NAS-supported layouts, number of bays, capacity target, workload, rebuild exposure, and tolerance for downtime. A two-disk mirror, dual-parity array, or paired mirror layout addresses different tradeoffs. Confirm the exact device's rules and maintain backups whichever layout you choose.

Last reviewed: September 22, 2026