Skip to content
All work

Industrial IoT platform

Smart Cloud

A telemetry platform for real-time machine-health monitoring: data from the shop floor travels through gateways, MQTT and queue-based processing into dashboards and alerts.

Role
Engineering at DCC Group — platform and backend, dashboards, gateway integration.
Stack
MQTT · RabbitMQ · PostgreSQL · Node.js · Next.js · Raspberry Pi · Modbus
Link
No public URL
VIBRATIONTEMPERATURECURRENT DRAWALERTMACHINE HEALTH · ILLUSTRATIVE

Original illustrative artwork, not a screenshot of a real system.

The problem

Machines on a factory floor produce a continuous stream of readings in industrial protocols that web software cannot consume directly. Operators need those readings turned into machine-health views, alerts and asset tracking, reliably and in near real time.

My role & responsibilities

  • Designing the ingestion path from field devices to the cloud.
  • Building backend services and APIs that process and store telemetry.
  • Building the dashboards that surface machine health, alerts and asset status.
  • Working with gateway devices (Raspberry Pi) and Modbus integrations.

Architecture

  1. 01

    Field layer

    Machines and PLCs expose registers over Modbus. A gateway polls them on a schedule.

  2. 02

    Transport

    The gateway publishes readings to an MQTT broker, a lightweight protocol suited to constrained devices and unreliable links.

  3. 03

    Processing

    Messages are handed to RabbitMQ and consumed by workers that validate, normalise and enrich them.

  4. 04

    Storage & API

    Normalised telemetry lands in PostgreSQL and is exposed through REST APIs.

  5. 05

    Dashboards

    A web dashboard reads the API to show machine health, alerts and asset tracking.

Explore the interactive pipeline

Engineering decisions

  • Decouple ingestion from processing with a queue, so a burst of messages or a slow consumer does not lose data.
  • Keep gateways thin: poll, timestamp, publish. Interpretation happens centrally where it is easier to change and observe.
  • Treat every device payload as untrusted input and validate it before it reaches the database.

Challenges

  • Bridging industrial protocols and web infrastructure without losing data on flaky site networks.
  • Keeping dashboards responsive while telemetry volumes grow.
  • Making device faults distinguishable from application faults when something looks wrong.

What shipped

  • Real-time machine-health monitoring, asset tracking, telemetry and alerting platform (in development at DCC Group).

Architecture is conceptual. Components listed vary by deployment, and no customer names or metrics are published.

Next case study

Service CRM & Field Service Management