The problem
Factories run blind until something breaks
Industrial plants across Africa and emerging markets operate without real-time visibility into equipment state. Sensors exist but are not connected. Data is collected manually — if at all. When a motor overheats or a bearing fails, the first alert is the sound of metal grinding.
The major vendors — AVEVA, Siemens, Rockwell — price their platforms for Fortune 500 budgets. A basic AVEVA PI deployment costs €500,000+. The result: factories that need this infrastructure the most cannot access it.
Unplanned downtime costs manufacturing firms between 5% and 20% of annual revenue. The fix is not more hardware. It is a data pipeline that connects existing sensors to dashboards operators can act on.
Our approach
Sensor-to-dashboard pipelines built for production
We deploy MQTT-based telemetry collection through EMQX, a production-grade broker that handles the throughput industrial environments demand. Sensor data streams through Kafka into TDengine — a time-series database purpose-built for industrial volumes.
Every deployment follows ISA-95 architecture. We do not invent new patterns. We implement the industry standard — Level 0 through Level 4 — and integrate with existing plant systems via OPC-UA connectors.
Security is zero-trust from day one. SPIFFE/SPIRE for workload identity, mTLS between every service, secrets managed by Vault. We do not bolt security on after deployment. It is part of the architecture.
Visualization runs through Grafana with custom dashboards per production line. Alerts fire before operators need to intervene. The system self-heals — services restart, pipelines resume, data is never lost.
Deliverables
- Complete sensor-to-dashboard pipeline — MQTT collection, stream processing, time-series storage, visualization
- ISA-95 compliant architecture documentation
- Zero-trust security layer with SPIFFE/SPIRE workload identity
- Grafana dashboards with per-line alerting and anomaly detection
- Infrastructure as code — Terraform and Ansible for repeatable deployments
- Operational runbooks and incident response procedures
Proof of work
Forgelink — zero-trust IIoT platform
Production-grade IIoT platform simulating a complete steel plant process — electric arc furnace, continuous casting, rolling mill, finishing. 68+ sensors publishing via MQTT to EMQX. Streaming through Kafka to TDengine historian. Visualized in Grafana and Flutter mobile HMI. Zero-trust security via SPIFFE/SPIRE throughout. Infrastructure as code via Terraform and Ansible.