IoT Protocol

Industrial MQTT: The Messaging Protocol for the IoT

MQTT for Industrial IoT: How It Works, QoS Levels, MQTT Brokers, TLS Security, Sparkplug B, and a Comparison with HTTP/AMQP.

Quick answer
MQTT (Message Queuing Telemetry Transport) is the standard messaging protocol for industrial IoT. Designed for low-bandwidth, high-latency connections, it is based on a model publish/subscribe : The gateways publish data to topics, and the cloud platform subscribes to these topics to receive the data in real time. Ultra-lightweight (2-byte header), it supports three levels of quality of service (QoS 0, 1, 2) and TLS encryption.

Why MQTT Has Become the Standard in Industrial IoT

The Industrial IoT generates millions of data points every day: sensors that report every minute, PLCs that send status updates every second, and meters that transmit hourly readings. These data streams must be transmitted reliably, efficiently, and securely.

MQTT meets these needs exactly:

  • Ultra-lightweight — 2- to 5-byte headers. HTTP/REST generates hundreds of bytes of headers for a payload of just a few bytes.
  • Persistent connection — the gateway connects to the broker once and maintains the connection. No TCP handshake is required for each message.
  • Publish/Subscribe — multiple subscribers (cloud monitoring, ERP, BI) can receive the same data without the gateway being aware of it.
  • Configurable QoS — depending on the criticality of the data.
  • Resilience — messages are stored if the client is temporarily disconnected (retain, persistent session).

Industrial MQTT Architecture

MQTT Topics: Structure and Best Practices

An MQTT topic is a hierarchical chain separated by /:

eziwan/{site_id}/{equipment}/{variable}

Exemples :
eziwan/pumping-station-01/control-panel/discharge-pressure
eziwan/site-pompage-01/compteur/energie_kwh
eziwan/batiment-paris/cvc/temperature_zone_a

Best Practices for Naming:

  • Always include the site ID to filter by site.
  • Separate data types (measurement, status, alarm, command) into distinct topics.
  • Use human-readable names, not obscure codes.
  • Avoid overly broad topics (eziwan/#) for production subscriptions.

QoS Levels: What’s the Best Choice for Industry?

QoSGuaranteeOverheadRecommended Use
0At most once (best-effort)MinimalFrequent measurements (1/s), loss OK
1At least once (with acknowledgment)LowStandard monitoring, alerts
2Exactly once (4 exchanges)MediumCommands, critical data

For typical industrial IoT monitoring (measurements every 5 to 60 seconds), QoS 1 offers the right balance: sufficient reliability and acceptable overhead.

MQTT vs. HTTP for the IoT: A Comparison

CriterionMQTTHTTP/REST
ModelPublish/SubscribeRequest/Response
Overhead size2–5 bytes200–400 bytes
ConnectionPersistentPer request (or keep-alive)
Data pushNative (subscribe)Polling or WebSocket
Latency< 50 ms100–500 ms
Multiple subscribersNativeWebhooks/external queues
Suitable for low-bandwidth IoT★★★★★★★

The Eziwan Approach

The Eziwan Gateway implements an MQTT client that publishes the collected Modbus/OPC UA data to the Eziwan Cloud MQTT broker. Topics are auto-generated based on each device’s configuration. Transmission is secured via TLS with client certificate authentication—there are never any unsecured MQTT connections.

Learn more: Modbus Cloud, Industrial IoT Gateway, and Remote Site Monitoring.

Frequently Asked Questions

You might also like