Services
Industrial IoT & automation
Production data is plentiful. What is usually missing is not the measurement - it is the path from measurement to decision.
When we are the right fit - and when we are not
Where we carry weight - and where we do not.
A fit when …
- Heterogeneous machines and controllers need to come under a shared data base.
- Data has to be pre-processed at the edge before it flows anywhere.
- There are data sovereignty requirements and the platform must be able to run on premises.
- Existing ERP or MES systems should be connected, not replaced.
Not a fit when …
- You are looking for an off-the-shelf IIoT product - we build custom.
- The controllers are not reachable and are not meant to be.
- This is solely about hardware selection and wiring.
- Below 10 person-days our ramp-up effort does not pay off for you.
What we actually do
- Real-time data acquisition - sensor and machine data via OPC UA, MQTT and Modbus, including customer-specific transport protocols for PLCs.
- Edge computing - pre-processing, plausibility checking and filtering directly at the machine, before data flows to a cloud or on-premises platform.
- Event streaming - Kafka and Redpanda as the backbone for high data volumes, with clean decoupling of acquisition and analysis.
- Predictive maintenance and anomaly detection - analysing machine data for early failure detection, with detection and alerting logic on the event stream.
- Factory digitalisation - integration with ERP and MES systems, real-time dashboards for operations.
- Deployment models - on premises, cloud or hybrid, depending on requirements for data sovereignty, latency and availability.
Why decoupling is the decisive design move
The most common mistake in IIoT platforms is coupling acquisition and analysis directly. If the analysis fails, measurements are lost - irrecoverably. We therefore insist on an event stream between the two: acquisition writes, analysis reads at its own pace. New analyses can be retrofitted onto historical data later without touching acquisition.
Approach & collaboration models
Usually a scoped proof of concept on a real machine first, then expansion into a platform. Alternatively team reinforcement in an ongoing project, or an architecture review of the existing connectivity.
Technologies
Project reports from this field
EcoPulse
An event-driven microservice platform that consolidates inverters, storage, wallboxes and meters into a shared time series base and derives load decisions from it. Energy · IIoT · e-mobilityIndustrial data platform
Scalable platform for industrial data exchange with end-to-end secured communication. Industrial automation · metalsTypical project size
What a project costs with us.
Day rate €480-960 depending on the type of engagement
Below 10 person-days our ramp-up effort does not pay off for you. A scoped proof of concept on a real machine is the usual entry point.
Frequently asked questions
Which protocols do you connect?
OPC UA, MQTT (EMQX, Mosquitto), Kafka or Redpanda, Modbus and customer-specific PLC transport protocols. For proprietary controllers we have built our own transport implementations several times.
Cloud or on premises?
Both, and hybrid. Because acquisition, processing and analysis are decoupled via MQTT and Kafka, the boundary can be drawn freely - acquisition stays on site, analysis may run externally or not at all.
What happens at the edge?
Pre-processing and filtering directly at the machine: plausibility checks, aggregation, outlier handling. Only reliable values move on. That reduces bandwidth and storage and keeps measurement errors out of analyses.
Contact
Your contact
Software architecture & technical project management