About the RAID Calculator
RAID (Redundant Array of Independent Disks) combines multiple physical disks into one logical volume to improve performance, add fault tolerance, or both. Different RAID levels trade raw disk capacity for protection against drive failure in different ways. This calculator applies the standard capacity formula for RAID 0, 1, 5, 6, and 10 so you can see exactly how many usable terabytes an array delivers and how many disks can fail before data is lost.
The formulas by RAID level
Let N be the number of disks and C be the capacity of each disk. Raw capacity is always N × C. Usable capacity and fault tolerance depend on the level:
- RAID 0 (striping): usable = N × C. No redundancy — data is split evenly across every disk with no duplication, so losing any one disk loses the entire array. Minimum 2 disks.
- RAID 1 (mirroring): usable = C. Every disk holds an identical copy, so usable space equals a single disk's capacity no matter how many disks you add. Tolerates all but one disk failing. Minimum 2 disks.
- RAID 5 (single parity): usable = (N − 1) × C. One disk's worth of space is reserved for distributed parity data, which can reconstruct any single failed disk. Minimum 3 disks.
- RAID 6 (dual parity): usable = (N − 2) × C. Two disks' worth of space go to parity, allowing the array to survive any two simultaneous disk failures. Minimum 4 disks.
- RAID 10 (mirrored stripes, a.k.a. RAID 1+0): usable = (N ÷ 2) × C. Disks are mirrored in pairs, then the pairs are striped together. This gives half the raw capacity and tolerates one failure per pair. Requires an even number of disks, minimum 4.
Storage efficiency
Storage efficiency is simply usable capacity divided by raw capacity, expressed as a percentage. RAID 0 is 100% efficient but offers zero protection. RAID 5 efficiency rises as you add disks (a 3-disk array is 67% efficient; a 10-disk array is 90% efficient) because the one-disk parity overhead is spread across more disks. RAID 6 follows the same pattern with two disks of overhead. RAID 1 and RAID 10 are fixed at 50% efficiency because exactly half the raw capacity is a mirrored copy.
Choosing a RAID level
RAID 0 suits scratch space or temporary data where speed matters more than safety. RAID 1 suits small, critical volumes like boot drives. RAID 5 is a common general-purpose choice balancing capacity and single-drive protection, though rebuild times on very large disks raise the risk of a second failure during recovery. RAID 6 addresses that risk for large arrays by tolerating two failures. RAID 10 offers the fastest rebuilds and strong performance, at the cost of losing half the raw capacity to mirroring.