Practical guide
IOPS vs throughput: formula and worked examples
Convert IOPS and block size to MiB/s, understand the arithmetic ceiling, and avoid confusing throughput with storage latency.
Updated
IOPS counts completed input/output operations per second. Throughput measures data transferred per second. Block size connects them: throughput in MiB/s = IOPS × average KiB per operation ÷ 1,024.
Worked examples
At 3,000 IOPS and 16 KiB per operation, theoretical throughput is 3,000 × 16 ÷ 1,024 = 46.875 MiB/s. At the same IOPS with 256 KiB operations, the arithmetic result is 750 MiB/s—but a volume, instance, or network throughput cap may prevent that result.
Input the observed average block size in the IOPS and throughput calculator. The output is a unit conversion, not a performance guarantee.
Common mistakes
- Treating provisioned IOPS as observed application IOPS.
- Ignoring queue depth and latency.
- Mixing decimal MB with binary MiB.
- Assuming random and sequential access behave identically.
- Calculating a value above a documented volume or instance cap and treating it as attainable.
Use the EBS cost calculator or Persistent Disk cost calculator after determining the required performance tier.