Building an Industrial IoT Pipeline with ESP32, MQTT & AWS

In industrial environments, reliability is non-negotiable. This guide walks through building a production-grade data pipeline that collects sensor data at the edge and delivers it to AWS for storage, alerting, and visualization.

Architecture Overview

[Sensors] → [ESP32 Gateway] → [MQTT/TLS] → [AWS IoT Core] → [DynamoDB] → [Grafana]

Each component is chosen for resilience and scalability.

Edge: ESP32 as Gateway

The ESP32 reads data from multiple sensors over I2C and SPI, then publishes JSON payloads to an MQTT broker.

// Sensor reading structure
typedef struct {
    float temperature;
    float humidity;
    float vibration;
    uint32_t timestamp;
} SensorPayload;

void publishData(SensorPayload *data) {
    char json[256];
    snprintf(json, sizeof(json),
        "{\"temp\":%.2f,\"hum\":%.2f,\"vib\":%.2f,\"ts\":%lu}",
        data->temperature, data->humidity,
        data->vibration, data->timestamp);

    mqttClient.publish("factory/line-1/sensors", json);
}

Communication: MQTT with TLS

For industrial deployments, always use TLS mutual authentication:

• Device certificates provisioned via AWS IoT
• Root CA certificate embedded in firmware
• Automatic reconnection with exponential backoff

Cloud: AWS IoT Core

AWS IoT Core handles:

1. Device Authentication — X.509 certificates
2. Message Routing — IoT Rules Engine
3. Data Storage — DynamoDB or TimeStream
4. Alerting — SNS notifications on threshold breaches

IoT Rule Example

SELECT temperature, humidity, timestamp()
FROM 'factory/+/sensors'
WHERE temperature > 80

This rule triggers an SNS alert whenever temperature exceeds 80°C.

Visualization: Grafana Dashboards

Connect Grafana to your TimeStream database for real-time dashboards:

• Line charts for temperature trends
• Gauge widgets for current values
• Alert annotations for threshold events
• 30-day data retention policies

Lessons Learned

1. Always buffer locally — If the network drops, store readings in SPIFFS and replay later
2. Use QoS 1 minimum — At-least-once delivery prevents data loss
3. Monitor device health — Publish heartbeat messages every 60 seconds
4. Version your firmware — OTA updates are essential for production deployments

Want to build this system hands-on? Join our Industrial IoT Architecture track.