Halkali merkez neighborhood fatih st ozgur apt no 46 , Kucukcekmece , Istanbul , 34303 , TR
Enter disk count, size, RAID level and hot spares to compare raw capacity, usable capacity, redundancy overhead, efficiency and fault tolerance. Mixed-size arrays use a conservative smallest-disk method for cross-controller planning.
RAID is not backup. RAID improves availability during disk failure; separate backups are still required for deletion, ransomware, controller failure and disaster scenarios.
| Level | Min. disks | Usable formula | Fault tolerance | Capacity efficiency | Typical use |
|---|---|---|---|---|---|
| RAID 0 | 2 | N × S | 0 | 100% | Scratch space, temporary high throughput |
| RAID 1 | 2 | 1 × S | N−1 | 1/N | OS, simple mirror, high availability |
| RAID 5 | 3 | (N−1) × S | 1 | (N−1)/N | Read-heavy balanced storage |
| RAID 6 | 4 | (N−2) × S | 2 | (N−2)/N | Larger arrays, stronger parity protection |
| RAID 10 | 4, even | (N/2) × S | Pair-dependent | 50% | Database, VM and write-heavy workloads |
With N active disks and smallest active disk size S, approximate usable capacity is (N−1) × S. One disk equivalent is reserved for parity.
Approximate usable capacity is (N−2) × S. Two disk equivalents are reserved for dual parity.
In a typical mirror-pair + stripe layout, capacity is N/2 × S. Losing both disks in the same mirror pair can fail the array.
Standard RAID 5 uses single parity and tolerates one member disk failure. A second failure during rebuild can cause data loss.
RAID 6 stores two independent parity blocks, reserving the equivalent of two disks while tolerating two failures.
RAID 10 stripes mirrored pairs. Typical layouts require at least four disks and an even member count.
No. A hot spare normally holds no active array data and waits to replace a failed member.
Most classic RAID layouts are constrained by the smallest active disk. Some modern systems differ; this calculator uses the conservative common-denominator method.
If the controller or software reads from mirror members in parallel, read throughput can improve. Writes remain closer to single-disk behavior because every copy must be written.
For NVMe, SSD or HDD storage, capacity must be planned together with IOPS, latency, controller cache, rebuild time and backup targets.
Turkey (Turkish)
Germany (German)
Worldwide (English)
