Unit and integration proof
Tests show a parser, detector, isolation boundary, or workflow behaves as specified under controlled inputs.
Evidence methodology
Every claim needs a provenance label.
A test result, a virtual-device result, and a customer field result answer different questions. AssureMesh treats them as different evidence tiers.
Evidence tiers
Recorded fixture proof
Sanitized telemetry or device responses exercise a real parsing and detection path repeatably.
Licensed or simulated network
A compiled product runs against virtual devices or protocol simulators. Useful, but not physical-hardware certification.
Named physical platform
The path is tested against a physical device with model and firmware recorded.
Authorized customer environment
The system observes a real operating environment, with publication limited to evidence the customer permits.
Independent review
A qualified third party validates a control or outcome. AssureMesh does not use this label without the assessment.
How benchmark numbers are read
When AssureMesh publishes a benchmark number, it is labeled “test-measured” if it comes from repeatable test scenarios, not from a statistically representative customer fleet. The homepage currently shows no AssureMesh benchmark numbers; the figures on it are third-party industry statistics, not AssureMesh results.
Noise-reduction scenarios
A scenario defines a known set of related and unrelated observations. The expected result preserves independent faults and groups only the observations that belong together. Any reduction we report describes that scenario, not a guaranteed customer percentage.
RCA scenarios
A scenario supplies a known fault and evidence sequence. The expected result checks whether the deterministic diagnosis identifies the intended cause and whether unrelated evidence is incorrectly linked.
What is not claimed
- No universal reduction percentage across every network.
- No guaranteed mean-time-to-resolution improvement.
- No claim that every device alarm has a producer.
- No claim that virtual-device success equals physical-hardware certification.
- No claim that AI is required for deterministic RCA.
How field evidence will be published
Field results should include the environment, sample size, observation period, starting condition, measurement definition, important exclusions, and customer permission. A testimonial without enough context will not be presented as a benchmark.
Last reviewed: October 6, 2026.
Questions or challenges to a public product claim can be sent to support@assuremesh.com.
Questions or challenges to a public product claim can be sent to support@assuremesh.com.