Performance · methodology before claims
No figures on this page until they are measured.
Headline benchmarks in this market are typically measured at large packet sizes with subscriber features disabled, and with limited methodology disclosure. Netavo publishes its test protocol first; measured results follow, reproducible on your own hardware with the evaluation build.
MEASURED ONLY. Every figure carries the exact CPU, NIC, memory, traffic profile, session count, and loss criterion. No extrapolation, no "up to".
IMIX BESIDE EVERY LARGE-PACKET FIGURE. A 1500-byte result in isolation overstates real-world capacity. Both figures, always, bidirectional.
FEATURES ON. Headline numbers run with QoS enabled at full session scale — the configuration you'd actually operate. Bare forwarding appears only as a secondary line.
REPRODUCIBLE. Each row links its runbook and raw results. Re-run it on our eval build and hold us to it.
For calibration: the most prominent per-server throughput claim in this market — one terabit — was produced at 1,024-byte packets in a co-branded laboratory test. Results of that kind are standard industry practice, and they are why operators discount vendor benchmarks.
We would rather publish a smaller figure you can reproduce than a larger one you cannot. Every cell in the table below will meet rules R1–R4.
| Reference hardware | Sessions | IMIX bidir | 1500B bidir | QoS | Runbook |
|---|---|---|---|---|---|
| entry · 8-core / 2×10G | 5,000 | measuring… | measuring… | on | pending |
| mid · 16-core / 2×25G | 25,000 | measuring… | measuring… | on | pending |
| high · 32-core / 2×100G | 100,000 | measuring… | measuring… | on | pending |
Reference configs will match what real deployments run — the telemetry hardware-inventory program keeps this list honest. Want your hardware class represented? Tell us what you run.
The same counters the benchmark protocol records — visible live on any deployment, including the evaluation build.
Benchmark it on your own hardware.
The evaluation build is the production build. Deploy it on your own x86 hardware, drive it with your traffic generator, and compare your results with ours.