Write Amplification Calculator

Work out write amplification instantly with clear inputs, formula shown and shareable results.

About Write Amplification Calculator

Write amplification is physical bytes written divided by logical bytes. A B-tree writes the log plus the page, plus each index. A levelled LSM rewrites data once per compaction level, so amplification grows roughly with level count — often ten times or more. This matters for SSD endurance and for provisioned IOPS, both of which are billed on physical writes.

Formula

Write amplification

B-tree ~ 2 + 0.5 x indexes; levelled LSM ~ 1 + 1.1 x levels + 0.4 x indexes; size-tiered ~ 1 + 0.5 x levels + 0.4 x indexes; physical = logical x factor

Worked example

With the default values below, Write Amplification Calculator returns Write amplification factor of 8.10. Change any input above and every figure here is recalculated in your browser.

Inputs used

Logical write rateminimum 0.001 MB/s: 20 MB/s
Secondary indexesminimum 0: 4
LSM compaction levelsaccepts 0–12: 5
Storage engineone of B-tree with write-ahead log, LSM tree with levelled compaction, LSM tree with size-tiered compaction: LSM tree with levelled compaction

Results

Write amplification factor: 8.10 x
Physical write rate to storage: 162.00 MB/s
Physical bytes written per day: 12.73 TiB
Extra bytes written per logical byte: 7.10

Step by step

1Values used
2Write amplification
3Write amplification factor= 8.10 x
4Physical write rate to storage= 162.00 MB/s
5Physical bytes written per day= 13,996,800,000,000
6Extra bytes written per logical byte= 7.10

Assumptions and limits

The result is derived only from the 4 inputs listed above. Anything not among them is not part of the calculation.
In particular, this calculator does not take inflation, tax, fees or charges or compounding frequency into account.
Figures are rounded for display to 2 decimal places, so totals built from several rounded parts can differ slightly from the unrounded arithmetic.

How these calculators are built and tested is described on the Methodology page. If a figure looks wrong, please report it.

Frequently Asked Questions

Why does levelled compaction amplify writes so much?

Each level is roughly ten times the previous one, and merging a level into the next rewrites the overlapping data. The trade is much better read amplification and space efficiency.

Which engine should I pick for write-heavy workloads?

Size-tiered LSM has the lowest write amplification but the worst space and read amplification. Levelled is the usual default; tiered suits append-heavy time-series data.