Physical units under test – ECUs, microcontrollers, sensors, actuators. Firmware revision, hardware variant, serial number.
Handoff: Device identity must reach every upstream layer. If it doesn't, results can't be traced to a specific unit.
The Stack
Six layers. One orchestration flow. Your existing tools, still in place.
From physical devices (L0) to test evidence (L5) – every handoff visible, every run reproducible. Not a tool replacement. A diagnostic and orchestration layer around what you already own.
Architecture
The L0-L5 model describes what every test lab already has – whether named or not.
CAN transceivers, RS-485 gateways, JTAG probes – e.g. dSPACE HIL, Vector interfaces, signal conditioners. The physical bridge between DUT and data.
Handoff: Calibration date, instrument ID, and configuration should ideally travel with each measurement – most standard exports omit this context.
Network switches, VLANs, DHCP, mDNS – how signals travel from instruments to the orchestration layer.
Handoff: Stand topology maps from L2 must be discoverable by L3. In many labs this configuration is maintained outside the toolchain.
Discovery, session management, config deduction, dependency resolution, container launching, health monitoring, capability enforcement. The orchestrator.
Handoff: L3 resolves which driver and frontend to load per session. If L3 has no structured info from L2, the orchestration cannot resolve the correct driver – the session fails before it starts.
ROS2 topics, DDS transport, HTTP control plane. Typed pub/sub messaging between components.
Handoff: Commands and telemetry flow over typed topics with QoS. If the lab has no structured messaging layer, every handoff is manual and unrecoverable.
Test scripts (pytest), CLI tools, evidence capture, CI/CD integration. The interface between engineers and the test infrastructure.
Handoff: Evidence capture links every result to the session context – firmware version, DUT serial, instrumentation state, operator identity. This is what makes results defensible in a technical review.
Not a product pitch
The model is a way of seeing, not a platform migration.
Finnegans Test is being built. The L0-L5 model is a Finnegans framework concept – not an industry standard – but it maps to real lab structures and is already useful today as a diagnostic lens – map one workflow against the six layers, find where handoffs break, decide what to fix. That exercise costs nothing and requires no change to your existing tools.
- We do not replace LabVIEW, TestStand, Vector, or dSPACE.
- We do not certify compliance by ourselves.
- We do not claim production-ready remote hardware testing.
- We do not claim zero integration effort – the configuration process takes time.
- We commit to start with one workflow, one diagnostic – not a platform migration.
Start here
Map one workflow against the six layers.
Use the diagnostic kit – free, 30-45 minutes, no data shared. If the gaps are structural, the next step is a paid diagnostic workshop (EUR 2k-5k, 1-2 weeks).