The software platform behind connected smart lighting

A single greenhouse can hold five thousand LED grow lights, and a smart building thousands more. Our client, a European wireless lighting-controls company, makes the hardware that connects them. Toobler built the cloud platform that turns all of them into one system a grower or facility manager can see, control and automate from anywhere.

Customer and context

The client is a wireless-controls and IoT company for smart lighting. Its hardware turns individual LED fixtures, in horticultural greenhouses and commercial buildings, into connected, controllable endpoints on a wireless mesh. The client builds that hardware. Toobler built the software: the enterprise, multi-tenant, multi-site cloud platform that sits above the network and makes the whole estate manageable. Toobler has been the client’s software engineering partner since 2022, and the platform is still being developed today.

The challenge

The hardware could make thousands of fixtures wirelessly addressable. The software had to make them operable, at scale, across sites, for non-technical operators:

• Thousands of endpoints, one system. A single greenhouse can hold thousands of fixtures and sensors across zones, and a customer might run several sites in different countries. The platform had to model and manage all of it centrally.
• Real-time control and monitoring. Operators needed to control lighting (zones, dimming, schedules, multi-channel spectrum) and see live state, faults and energy as they happen.
• Multi-tenant, multi-site. Many customers on one platform, each with its own sites, buildings, zones, users and permissions.
• Two very different settings. The same system had to serve greenhouse growers and commercial-building managers, with different structures and different needs.

The solution

Toobler designed and built the platform as the cloud and intelligence layer above the client’s wireless network.

A model of the physical estate. The platform represents the real world the devices live in:

customer → site → building or greenhouse → floor or section → zone → fixtures and sensors

An operator can navigate from a whole portfolio down to a single lighting zone, with floor-plan visualization of where everything is.

Central control and real-time monitoring. From one place, operators control lighting (on/off, dimming, zoning, scheduling and multi-channel spectrum), commission devices, and monitor live state, faults and alerts across the entire deployment.

A real-time data pipeline. Devices publish over MQTT (Mosquitto); Telegraf collects the stream into InfluxDB for time-series storage, which feeds the platform’s dashboards; and AWS WebSocket carries live, two-way updates. Lambda functions aggregate the time-series data and archive it to S3.

Dashboards and analytics beyond lighting. Energy dashboards, occupancy and environmental information, reports and metrics, and alerts and fault monitoring turn a lighting network into operational data for the building or greenhouse.

Integrations. REST APIs and third-party, BMS and climate-computer integrations let the platform fit into the systems a building or greenhouse already runs on.

A cloud-native architecture built to scale. The platform runs on Amazon EKS (Kubernetes) with an Ambassador API gateway routing HTTP, WebSocket and MQTT traffic, Node.js services and MongoDB, containerized and orchestrated to grow with the number of devices, sites and tenants.

Built continuously. Toobler has developed the platform from 2022 to today, from the first sprints through successive releases and a DevOps pipeline (EKS setup, InfluxDB and MQTT on Kubernetes, CI/CD).

A digital twin of the lighting estate

The platform keeps a live digital model of the physical estate. Every fixture, sensor and zone has a digital identity; its real-time state (on or off, dimming level, energy, occupancy, environmental readings) is represented in the platform; and commands flow back the other way, from the model to the physical device. An operator can move from a customer’s whole portfolio down to a single zone on a floor plan, see what is happening there now, and change it.

That loop between the physical estate and its digital representation (sense, understand, act) is the connected-asset foundation a digital twin is built on. The platform doesn’t simulate or predict the physics of a space, so it isn’t a simulation-based twin. It is the live, controllable digital layer a twin needs beneath it, and it follows the same direction as Toobler’s digital-twin work.

Engineering worth naming

• The software layer of IoT. The client makes the wireless nodes; Toobler built the platform that makes thousands of them usable.
• A live digital model of the physical estate. Customer, site, building, zone and fixture, with floor-plan visualization, live state and control.
• Real-time, device to dashboard. MQTT → Telegraf → InfluxDB, plus WebSocket, carries control and telemetry live, end to end.
• One platform, two industries. The same multi-tenant system serves greenhouse horticulture and commercial smart buildings.

Results

• Thousands of fixtures and sensors, managed as one system — controlled, monitored and automated from a single cloud platform.
• Real-time visibility and control across zones, sites and tenants, with energy, occupancy and environmental data.
• A platform that spans greenhouses and smart buildings — multi-tenant, multi-site, and integrated with the systems around it.
• A multi-year software partnership — Toobler has built and evolved the platform as the client’s software team since 2022.

What this means for IoT software

In a connected-device product, much of the value sits in the software that turns thousands of devices into something a person can operate: modeling the estate, controlling it in real time, making sense of its data, and integrating it with everything around it. This platform is that layer for connected lighting, engineered on a cloud-native, real-time foundation. If you make connected hardware and need the platform that turns it into a product, talk to Toobler’s engineering team.