How to use the RAID calculator
- Pick HDD or SSD, then click a capacity to drop that drive into a bay: hard drives from 1 TB to 24 TB, SSDs from 0.48 TB to 15.36 TB, up to 24 drives. Click a filled bay to pull the drive back out, or use Reset to start over.
- Mix capacities if you need to. Every drive counts at the size of the smallest one, the calculator shows the difference as Unused, and it flags a build that combines hard drives and SSDs, which RAID controllers keep in separate arrays.
- Choose a RAID level. If the build has fewer drives than that level’s minimum, the calculator tells you how many more to add.
- Read the bar: Available is the capacity you can store data on before formatting, Protection is the capacity redundancy uses, and Unused is capacity stranded by the smallest-drive rule.
- Step three lists up to four chassis with at least as many drive bays as your build, smallest first. It matches on bay count only, so check that your drives (2.5-inch or 3.5-inch) suit the chassis.
How usable RAID capacity is calculated
For every level except JBOD, usable capacity is the capacity of the smallest drive multiplied by the number of drives that hold data rather than redundancy. RAID 0 keeps them all, RAID 5 gives one drive’s worth to parity, RAID 6 two, RAID 10 half the drives, and RAID 50 and RAID 60 one or two drives per group. These are the formulas in the HPE Smart Array SR Gen10 User Guide and the Broadcom 12Gb/s MegaRAID Tri-Mode Software User Guide, applied per group for RAID 50 and RAID 60, and the ones this calculator uses.
Scroll the table sideways to see every column.
| Level | Usable capacity | Used for protection | Usable share of raw | 8 × 4 TB drives | Minimum drives |
|---|---|---|---|---|---|
| RAID 0 | n × smallest drive | None | 100% | 32 TB | 1 |
| RAID 1 | the smaller of the two drives | One drive (the mirror copy) | 50% | 4 TB from 2 drives | 2 |
| RAID 5 | (n − 1) × smallest drive | One drive’s worth (parity) | 75% of 4 drives, 87.5% of 8, 95.8% of 24 | 28 TB | 3 |
| RAID 6 | (n − 2) × smallest drive | Two drives’ worth (parity) | 50% of 4 drives, 75% of 8, 91.7% of 24 | 24 TB | 4 |
| RAID 10 | (n ÷ 2) × smallest drive | Half the drives (mirror copies) | 50% at any drive count | 16 TB | 4 |
| RAID 50 | (n − 2) × smallest drive, two groups | One drive’s worth per group | 66.7% of 6 drives, 75% of 8, 91.7% of 24 | 24 TB (2 groups of 4) | 6 |
| RAID 60 | (n − 4) × smallest drive, two groups | Two drives’ worth per group | 50% of 8 drives, 66.7% of 12, 83.3% of 24 | 16 TB (2 groups of 4) | 8 |
| JBOD | sum of every drive | None | 100% | 32 TB | 1 |
n = the number of drives; smallest = the capacity of the smallest drive. Shares and examples assume drives of equal size. RAID 50 and RAID 60 use the calculator’s two-group layout; a controller needs the drive count to divide evenly into groups, and more groups store less (HPE). Minimums from the Dell PERC 11 User’s Guide; shares and examples derived from the formulas.
Usable capacity by RAID level
Each level below gives the usable-capacity formula, a worked example with eight 4 TB drives and the protection the level provides. n is the number of drives in the array and smallest is the capacity of the smallest drive in it. To choose between levels, read our RAID 5 vs RAID 6 vs RAID 10 guide, which covers drive minimums, rebuild risk and write performance.
RAID 0 — Striping, no redundancy
Usable capacity is every drive at the size of the smallest: n × the smallest drive, so 8 × 4 TB drives give 32 TB. Nothing is set aside for redundancy, and per SNIA, RAID 0 loses data when any one drive in the set fails.
RAID 1 — Two-drive mirror
Usable capacity is one drive, the smaller of the two, because the second drive holds a full copy: 2 × 4 TB drives give 4 TB. Dell’s PERC guides build RAID 1 from exactly two drives; with more than two drives mirrored in pairs, HPE calls the layout RAID 10. A two-drive mirror survives one drive failure.
RAID 5 — Single distributed parity
Usable capacity is (n − 1) × the smallest drive: one drive’s worth of capacity holds parity, spread across all of the drives, whatever the drive count. 8 × 4 TB drives give 28 TB. RAID 5 survives one drive failure.
RAID 6 — Double distributed parity
Usable capacity is (n − 2) × the smallest drive: RAID 6 keeps two independent parity sets, and HPE notes that each uses the capacity of one drive. 8 × 4 TB drives give 24 TB. RAID 6 survives any two drive failures.
RAID 10 — Striped mirrors
Usable capacity is (n ÷ 2) × the smallest drive, because every drive is mirrored to a partner: 8 × 4 TB drives give 16 TB, and the share stays at 50% however many drives you add. Drives work in pairs, so the calculator shows an odd extra drive as Unused. RAID 10 survives one failure in each mirrored pair.
RAID 50 — Striped RAID 5 groups
Each RAID 5 group gives one drive’s worth of capacity to its own parity, so usable capacity is groups × (drives per group − 1) × the smallest drive. The calculator models two groups, (n − 2) × the smallest drive: 8 × 4 TB drives give 24 TB. Controllers build groups of equal size, and more groups store less. RAID 50 survives one failure in each group.
RAID 60 — Striped RAID 6 groups
Each RAID 6 group gives two drives’ worth of capacity to its parity, so usable capacity is groups × (drives per group − 2) × the smallest drive. In the calculator’s two-group model that is (n − 4) × the smallest drive: 8 × 4 TB drives give 16 TB. RAID 60 survives two failures in each group.
JBOD — Just a bunch of drives
Every drive counts at its full size, so mixed sizes leave nothing unused: 8 × 4 TB drives give 32 TB. HPE’s MR controller guide describes a JBOD drive as one exposed to the host operating system as a physical drive, outside any RAID configuration. ZFS and Ceph expect drives presented this way; the Proxmox VE Administration Guide says ZFS needs to communicate directly with the disks.
Why usable capacity is lower than the raw total
Four things reduce the capacity printed on the drive labels, and they stack: the smallest drive, redundancy, decimal versus binary units, and formatting.
The smallest drive sets the size of every member. Dell’s enterprise RAID FAQ says RAID controllers “allocate space on hard drives into arrays by the smallest drive in the array.” Put one 4 TB drive alongside seven 12 TB drives and all eight count as 4 TB drives; the other 56 TB is stranded, which the calculator shows as Unused. Use matched drives, which HPE’s server guides say give “the greatest storage space efficiency”, or put each size in an array of its own. JBOD is the one option here that counts every drive in full.
Redundancy takes its share. A mirror keeps a second copy, single parity uses one drive’s worth of capacity and double parity two. That is the Protection share of the bar: not wasted, but not available for your data.
Decimal terabytes are not binary terabytes. Drive makers count 1 TB as 1,000,000,000,000 bytes. Toshiba’s MG11 product manual explains that a computer operating system “reports storage capacity using powers of 2”, where 1 TB is 240 = 1,099,511,627,776 bytes, “and therefore shows less storage capacity.” Each decimal terabyte shows as about 0.91 TiB, roughly 9% less. This calculator works in decimal TB, so treat its result as the ceiling.
Formatting and the controller take some too. Seagate’s Exos 7E10 datasheet notes that “some of the listed capacity is used for formatting and other functions, and thus will not be available for data storage.” Each RAID member also stores the controller’s configuration: HPE’s MR controller guide says a drive in a logical drive “contains valid Disk Data Format (DDF) metadata”, the SNIA standard for RAID configuration data, and Dell’s PERC 12 guide offers drive coercion to give drives “that are slightly different in sizes” a consistent size.
Mixing drive sizes and drive types
You can mix drive sizes in one array, but not drive types. Every drive counts at the size of the smallest member, and RAID controllers keep hard drives and SSDs in separate arrays.
Sizes and speeds. Dell’s enterprise RAID FAQ calls it “perfectly valid” to use drives of different makes, models, sizes and rotational speeds, with two limits: capacity above the smallest drive “is not usable within that array”, and slower drives “may adversely affect the performance of the array.” HPE’s Smart Array guide requires any drive added to an array to have “a capacity no less than the capacity of the smallest drive in the array.”
Types. Dell’s PERC 10 guide states that mixing SAS and SATA drives, or hard drives and SSDs, within a virtual disk is not supported, and the PERC 11 User’s Guide repeats the hard drive and SSD rule. HPE requires an added drive to be “of the same type as existing drives in the array (for example, SAS HDD, SAS SSD, SATA HDD, or SATA SSD).” The calculator lets you add both types so you can count bays, then flags the mix: size each drive type as an array of its own. For SAS and SATA drives in more detail, see our SAS vs SATA guide.
From your numbers to a configured server
Step three, Systems that fit this build, lists the Dell PowerEdge servers, HPE ProLiant servers and Dell PowerVault arrays we sell with at least as many drive bays as your build. Open one to configure it.
Many of our Dell PowerEdge configurators (Dell PERC controllers) and our quote-based HPE ProLiant Gen10 Plus and Gen11 configurators (HPE Smart Array, SR and MR controllers) offer an optional RAID Configuration build service. In the option’s own words: “Optional build service. We configure the selected RAID level before the server ships.” Nothing is pre-selected, and the levels offered depend on the controller you choose.
The RAID level and the drives are separate choices in the configurator, so check the drive count and usable capacity here first. For a build the calculator does not model, such as more than 24 drives or more than one array, request a quote and describe the layout.
Frequently asked questions
- How do you calculate usable RAID capacity?
- Multiply the smallest drive's capacity by the number of drives that hold data. With n drives, that is n for RAID 0, n − 1 for RAID 5, n − 2 for RAID 6 and n ÷ 2 for RAID 10; a two-drive RAID 1 keeps one. RAID 50 and RAID 60 give up one or two drives per group. HPE's and Broadcom's controller guides publish these formulas.
- How do you calculate RAID 5 capacity, and how much storage do you lose?
- RAID 5 usable capacity is (n − 1) × the smallest drive, where n is the drive count, per HPE's and Broadcom's controller guides: one drive's worth of capacity holds parity. You lose 25% of the raw capacity with four drives, 12.5% with eight and about 4% with 24. Drives larger than the smallest also lose their extra capacity, which the calculator shows as Unused.
- How do you calculate RAID 6 capacity?
- RAID 6 usable capacity is (n − 2) × the smallest drive, where n is the drive count. HPE's Smart Array guide explains that each of RAID 6's two parity sets uses the capacity of one drive. Four drives keep 50% of the raw capacity, eight keep 75%, 12 keep 83.3% and 24 keep 91.7%. Eight 8 TB drives give 48 TB.
- How much usable space do 4 × 8 TB drives give in RAID 5?
- 24 TB. RAID 5 keeps the capacity of every drive but one, so four 8 TB drives keep three drives' worth and one drive's worth holds parity. An operating system that counts in binary units shows 24 TB as about 21.8 TiB, before formatting. Four 4 TB drives give 12 TB the same way. In RAID 6 or RAID 10, four 8 TB drives give 16 TB.
- Why does my operating system show less space than the calculator?
- Units and formatting. The calculator uses decimal terabytes, as drive makers do: Toshiba defines 1 TB as 1,000,000,000,000 bytes and notes that operating systems count in powers of 2, where 1 TB is 1,099,511,627,776 bytes, so each terabyte shows as about 0.91 TiB. Seagate adds that some of the listed capacity is used for formatting and other functions.
- Can you mix different drive sizes in one RAID array?
- Yes, but every drive counts as the smallest one. Dell's enterprise RAID FAQ says controllers allocate space by the smallest drive, so a RAID 5 of 500 GB, 200 GB and 100 GB drives uses 100 GB of each; the calculator shows the excess as Unused, and JBOD counts each drive in full. Dell PERC controllers keep hard drives and SSDs in separate virtual disks.
- How do RAID 50 and RAID 60 groups change usable capacity?
- Each group pays for its own parity: one drive's worth per RAID 5 group in RAID 50 and two per RAID 6 group in RAID 60, so more groups store less, as HPE's Smart Array guide notes. Twenty-four drives keep 22 drives' capacity as RAID 50 in two groups, 20 as RAID 60 in two groups of 12 and 18 in three groups of 8.
For drive minimums, how many failures each level survives, rebuild risk and which level to choose, see our RAID 5 vs RAID 6 vs RAID 10 guide.
