Last updated: 2026-06-29

IOPS to Throughput

Storage performance is reported in two units that are easy to confuse: IOPS, which counts operations, and throughput, which measures the volume of data moved per second. This calculator converts between them using the I/O block size — the amount of data carried by a single operation. Enter IOPS to get throughput, or a target throughput to find the required IOPS.

Two views of the same device

A storage device does work one operation at a time. IOPS (input/output operations per second) counts how many of those operations it completes each second — the right metric for workloads made of many small, independent accesses, such as a database serving random lookups. Throughput measures how many bytes per second flow through the device — the right metric for streaming large files. Neither number alone tells the whole story; the block size connects them.

The conversion

Each operation transfers one block of data, so throughput is simply the operation rate multiplied by the bytes per operation:

$\text{Throughput} = \text{IOPS} \times \text{block size}$

Rearranged, the operations needed to sustain a given data rate are:

$\text{IOPS} = \dfrac{\text{Throughput}}{\text{block size}}$

Because the two are linked through the block size, an IOPS figure quoted without a block size cannot be turned into a data rate, and a throughput figure cannot be turned into IOPS.

Worked example

A solid-state drive is rated at 10,000 IOPS for 4 KiB operations. Its throughput is:

The same drive, asked to move 1 MiB blocks instead, would need only about 40 operations per second to reach the same 40 MB/s — a reminder that IOPS and throughput trade off directly against block size.

Why block size dominates

Small blocks push IOPS up and throughput down, because each operation carries little data; large blocks do the reverse. This is why benchmarks always state the block size, and why a database tuned for 4 KiB random reads and a backup job streaming 1 MiB sequential writes stress completely different limits of the same hardware. When sizing storage, convert using the block size that matches your real workload, and bear in mind that a device's rated IOPS usually assumes a particular access pattern. For moving a known quantity of data, the Data Transfer Time Calculator estimates how long a transfer takes at a given rate.