Skip to content

Known limits

netverdict should be clear about its limits. A simulator is only useful if you know where the model ends.

netverdict does not boot real vendor operating systems. It models network behaviour directly. That makes labs fast, shareable, and repeatable, but it also means vendor-specific edge cases may not exist.

  • NAT: static NAT and UDP PAT work; TCP / ICMP PAT is not complete yet.
  • RIP: some config surface exists, but there is no full RIP routing process.
  • Some control-plane details are modelled for behaviour rather than exact packet-by-packet fidelity.
  • Vendor profiles vary in interactive CLI depth. Config import is sometimes deeper than live terminal coverage.

The most useful report includes:

  1. The lab or topology you used.
  2. The command or config line that surprised you.
  3. What you expected.
  4. What netverdict did instead.

Repeatability is the point: if we can replay the same state, we can fix the model.