Skip to main content

IoT Edge Aggregation & Telemetry

In industrial environments, smart homes, and edge computing scenarios, hardware resources are heavily constrained. Running a full Java Virtual Machine for Kafka or the Erlang runtime for RabbitMQ on a Raspberry Pi or an industrial IPC (Industrial PC) is often overkill and wastes valuable memory.

This is where TinyMQ's 13MB static binary shines.

The Challenge​

You have dozens of local sensors (temperature, humidity, motion, RFID) publishing telemetry data via MQTT. You need a local broker to:

  1. Aggregate the local traffic.
  2. Provide a buffer if the upstream internet connection drops.
  3. Securely forward the data to a centralized cloud application via HTTP REST.

The TinyMQ Solution​

Deploy TinyMQ on the edge device using the Docker scratch image.

  1. Ingestion: TinyMQ's built-in MQTT v3.1.1 Gateway listens on port 1883 and accepts connections from all local sensors.
  2. Buffering: The Write-Ahead Log (WAL) persists messages to the SD card. If the edge device loses its 4G/Wi-Fi connection, TinyMQ continues to ingest messages up to the TINYMQ_MAX_MESSAGES limit.
  3. Forwarding: Using TinyMQ's Webhook Subscriptions, the broker is configured to automatically POST messages to your cloud ingestion API.

Edge Configuration Example​

Configure TinyMQ with aggressive memory limits to ensure the edge device never runs out of RAM, while relying on the disk for durability:

services:
  tinymq-edge:
    image: ghcr.io/x-name15/tinymq:latest
    restart: always
    ports:
      - "1883:1883"
    environment:
      # Enable strict disk sync for durability on power loss
      - TINYMQ_FSYNC=true
      # Protect RAM: Max 10,000 messages in memory per topic
      - TINYMQ_MAX_MESSAGES=10000
      # Drop oldest messages if the buffer fills up during a long network outage
      - TINYMQ_DEFAULT_POLICY=drop-oldest
    volumes:
      - ./tinymq-data:/home/tinymq/data

Cloud Forwarding (Webhook)​

Use the TinyMQ CLI to register the cloud webhook. The webhook features built-in exponential backoff, meaning it will gracefully handle upstream cloud downtime.

tmq webhook add \
  --topic "sensors.*" \
  --url "https://api.mycloud.com/ingest" \
  --secret "cloud_hmac_secret"

Why it Shines Here​

  • Zero OS Dependencies: Can run directly on a minimal Alpine Linux or Yocto build.
  • Resource Limits: Hard limits on memory usage prevent the broker from causing Out-Of-Memory (OOM) kernel panics on the host.
  • Resiliency: Dropping the oldest messages ensures the system recovers gracefully from prolonged outages without operator intervention.