Hardware is the easy part to demonstrate. The platform is where an installation either becomes a system the plant keeps using or turns into a dashboard nobody opens after month three. Everything here is built around one idea: a single tag model that a 1998 controller and a 2024 breaker can both live inside.
We run it. You log in.
Multi-tenant, US-hosted, per-site subscription. Fastest to start and the cheapest at one site, and the option most plants choose for energy and lighting where there is no sensitive video involved.
Your hardware, your network.
A rack server in the plant or the customer's own VM. Required where video must not leave the building, or where the corporate IT policy simply forbids an outbound connection from the OT network.
Edge for video, cloud for the rest.
Video analytics run at the edge and only event metadata is published; energy and alarm data go to the hosted platform. This is the common answer once a security review has been through the proposal.
What a security questionnaire asks.
Role-based access, per-site tenancy, MQTT over TLS 1.2, signed firmware, full audit log of every control action, and a documented data-retention period per customer.
Modbus RTU and TCP as master, BACnet/IP, DL/T 645 and ONVIF ingest. The gateway polls; nothing new needs to be installed on a controller that has been running since 1998.
OPC UA server, MQTT topics per tag, and a REST API with the same authentication as the UI. Existing SCADA keeps its own screens and simply reads one more server.
Tag definitions export to CSV and JSON, historical data exports in bulk, and every device still speaks its native protocol without our gateway in the path. That is stated in the contract, not just here.