Usable RAID capacity is the raw total minus whatever the level spends on redundancy. The raw total is simply the number of disks times the size of each disk. From there, each RAID level keeps a different share: RAID 0 keeps everything, RAID 5 gives up one disk to parity, RAID 6 gives up two, RAID 1 and RAID 10 mirror and so halve the usable space. The RAID calculator applies the right formula for any level and also shows efficiency and fault tolerance.
Here is how each common level works, and why the usable number lands where it does.
Start from the raw total
Every calculation begins with the raw capacity, which assumes all your disks are the same size:
raw total = number of disks × size of each diskSix 2 TB disks give 12 TB raw. That figure is the ceiling. No RAID level gives you more than the raw total, and every level except RAID 0 gives you less, because some of the space goes to keeping a copy or a parity calculation.
Capacity by level
RAID 0 stripes data across the disks with no redundancy, so all of the raw total is usable. Four 1 TB disks give 4 TB usable. The catch is that losing any single disk loses the whole array.
RAID 1 mirrors the same data onto every disk, so you get one disk worth of space no matter how many disks you add. Two 2 TB disks give 2 TB usable. Extra disks add resilience, not capacity.
RAID 5 uses single parity, which costs one disk worth of space spread across the set. Usable space is (disks − 1) × size. Four 4 TB disks give (4 − 1) × 4 = 12 TB. It survives one disk failure.
RAID 6 uses double parity, costing two disks worth. Usable is (disks − 2) × size. Six 2 TB disks give (6 − 2) × 2 = 8 TB. It survives two failures, which matters during the long rebuild after one disk dies.
RAID 10 stripes across mirrored pairs, so half the disks hold copies. Usable is half the raw total. Four 1 TB disks give 2 TB usable. It is fast and tolerant but expensive on space.
RAID 50 and 60 stripe several RAID 5 or RAID 6 groups together. Each group loses one disk (50) or two disks (60) to parity. Eight 1 TB disks split into two groups of four under RAID 60 give two groups of (4 − 2) × 1 = 4 TB usable.
Efficiency and the cost of protection
Storage efficiency is the usable share of the raw total, written as a percentage:
| Level | Disks | Usable | Efficiency |
|---|---|---|---|
| RAID 0 | 4 × 1 TB | 4 TB | 100% |
| RAID 5 | 4 × 1 TB | 3 TB | 75% |
| RAID 6 | 6 × 1 TB | 4 TB | 67% |
| RAID 10 | 4 × 1 TB | 2 TB | 50% |
Efficiency rises with more disks in parity levels, because the fixed one- or two-disk overhead is shared more widely. A RAID 5 set of ten disks is 90% efficient; a set of three is 67%. That is the trade behind choosing a wider array.
Fault tolerance is the other half
Capacity alone does not tell you whether the array is safe. Fault tolerance is how many disks can fail before data is lost. RAID 0 tolerates none. RAID 5 tolerates one, RAID 6 two. RAID 10 tolerates at least one and often more, as long as both disks in the same mirror do not fail together. RAID 50 tolerates one failure per group, RAID 60 two per group.
The reason RAID 6 exists is the rebuild window. When a disk in a RAID 5 array fails, the array is exposed until the replacement finishes rebuilding, which can take many hours on large drives. A second failure in that window loses everything. RAID 6 keeps you protected through it.
Getting the numbers in one step
The formulas are worth knowing, because they show why a level costs what it costs. Day to day, enter your level, disk count and disk size into the RAID calculator and it returns usable capacity, efficiency and fault tolerance together, so you can compare layouts before you buy drives. And whatever you pick, keep a separate backup: RAID guards against a dead disk, not against the file you wish you had not deleted.