TOON is worth evaluating when JSON payload size matters and the data shape is structured enough to benefit from a more compact representation.
Current benchmarks (.NET 10.0.6, April 2026, 1,000 items, 15 warm iterations):
Speed:
That is the core value proposition: less data to move, less data to store, and better efficiency on real structured business payloads.
Benchmarked on .NET 10.0.6, 1,000 items per dataset.
| Dataset | JSON bytes | TOON bytes | Raw saved | Gzip saved | Brotli saved | JSON enc ms | TOON enc ms | JSON dec ms | TOON dec ms |
|---|---|---|---|---|---|---|---|---|---|
| Flat Orders | 151,066 | 55,142 | 63.5% | 15.7% | 7.7% | 0.794 | 2.026 | — | — |
| Catalog | 283,047 | 257,070 | 9.2% | 6.9% | 11.7% | 1.957 | 5.202 | — | — |
| Products | 2,878,419 | 1,253,844 | 56.4% | 14.6% | 9.0% | 36.550 | 18.050 | 35.626 | 35.049 |
| Warehouses | 849,281 | 200,114 | 76.4% | 20.1% | 20.0% | 2.051 | 2.350 | 3.630 | 2.997 |
| Change Fields | 300,822 | 176,965 | 41.2% | 7.4% | 4.0% | 0.718 | 0.838 | 1.676 | 0.854 |
| Configuration | 15,366 | 9,299 | 39.5% | 4.4% | 5.2% | 0.059 | 0.068 | — | — |
prods.json vs prods.toon2,051 actual products encoded to disk, decoded from disk, then re-encoded.
| JSON | TOON | Saved / Speedup | |
|---|---|---|---|
| Raw bytes | 11,520,246 | 2,568,353 | 77.7% |
| Gzip bytes | 421,934 | 354,875 | 15.9% |
| Brotli bytes | 274,797 | 246,025 | 10.5% |
| Avg decode (ms) | 26.1 | 18.6 | 1.40× |
| Avg encode (ms) | 16.2 | 18.0 | 0.90× |
In-process ASP.NET Core measurement, 12 warm iterations.
| Dataset | JSON (ms) | TOON (ms) | Delta | Raw saved | Gzip saved |
|---|---|---|---|---|---|
| Catalog | 12.369 | 11.628 | −0.741 ms | 56.3% | 14.5% |
| Products | 11.396 | 10.827 | −0.569 ms | 56.3% | 14.5% |
TOON is especially compelling when payloads include:
These are common enterprise shapes, not edge cases.
TOON should be positioned as a strong fit for the right workloads, not as a universal replacement story.
Strongest claim: TOON materially reduces payload size on repeated-row and nested collection data, and this package gives .NET teams a production-friendly way to adopt it.
Important caveats:
catalog data (9.2% raw saving) shows that not every nested shape is a strong fit.DevOp.Toon is the package that turns TOON from an idea into an application integration:
ToonNode or strongly typed models