$30 Pocket Router as an AWS IoT Edge Device

How to flash a ~$30 GL.iNet MT300N-V2 pocket router with a custom OpenWrt image that includes the AWS IoT Device SDK and an aws-iot-pubsub-demo application — turning it into an AWS IoT Edge device that can feed local sensor data to the cloud.

GL.iNet MT300N-V2 pocket router — ready to become an AWS IoT Edge device.

Quick overview

Three steps to turn your pocket router into an AWS IoT Edge device:

1. Flash the OEM firmware with the custom gl-mt300nv2-awsiot-demo.bin
2. Upload your AWS-generated device certificate and private key via the web UI
3. Reboot and watch Hello World messages published to your AWS IoT Core

Flashing the firmware

Step 1 — Connect your PC to the pocket router via Ethernet and power it on. Wait for the LED to stop blinking.

Setup: PC connected to the pocket router via Ethernet cable.

Step 2 — Open http://192.168.8.1/cgi-bin/luci/admin/system/flashops in your browser (or follow GL.iNet's onboarding process).

Step 3 — Find the firmware upgrade menu and flash with gl-mt300nv2-awsiot-demo.bin.

Important: Disable "Keep settings" — you want to start with default settings.

Step 4 — Wait about 2 minutes until the LEDs stop blinking.

Step 5 — Disconnect and reconnect the Ethernet cable on your PC so it gets a new IP in the 192.168.20.x range.

Step 6 — Navigate to http://192.168.20.1 — you should see the new web UI:

Custom firmware web UI after successful flash.

AWS IoT configuration

Step 7 — Upload your device certificate and private key files through the web UI:

Upload your AWS-generated device certificate and private key.

Step 8 — Go to AWS-IoT → Service Settings, enter your endpoint, and click Save & Apply.

Before clicking Save & Apply, ensure your security policies on console.aws.amazon.com are set up correctly (see Step 10 below).

Service Settings — enter your AWS IoT endpoint here.

Step 9 — Check the Service Log. If everything is configured correctly, you should see a "connection success" message:

Service Log — "connection success" confirms the router is connected to AWS IoT Core.

Step 10 — On console.aws.amazon.com, ensure your security policies are set correctly:

AWS IoT security policies — required permissions for the device.

Testing publish and subscribe

Step 11 — On console.aws.amazon.com, subscribe to topic test/topic to see Hello World messages published from your pocket router. By default, the demo publishes 10 messages at 5-second intervals. To publish continuously, increase the Publish Count in the Service Settings page.

AWS IoT Core — Hello World messages arriving from the pocket router.

Step 12 — To test the subscribe action, publish a JSON message to topic test/topic_led from the AWS console to control the router's LED:

{"powerstate" : "on"}
{"powerstate" : "off"}

Publishing LED control commands from the AWS IoT console.

Other use cases

The pocket router can act as a gateway between AWS IoT Core and local Wi-Fi or USB-connected devices:

Possible use cases — the pocket router bridges local devices to AWS IoT Core.

SOURCE CODE

github.com/hackboxguy/openwrt-wrapper — build instructions and sources

gl-mt300nv2-awsiot-demo.bin — pre-built firmware image

Related posts

• Getting Started with AWS IoT on OpenWrt Routers
• Remote-Kit: Pocket Router for Remote Hardware Bring-up
• AWS IoT on Raspberry Pi with 25 MB Buildroot Linux

Measuring PREEMPT_RT Latency with an Oscilloscope

A ready-to-use bootable SD card image for BeagleBone Black/Green with a fully preemptible Linux kernel (PREEMPT_RT) — includes a gpio-test program that mirrors an input signal to a GPIO output pin for measuring scheduler latency with an oscilloscope.

Oscilloscope view — yellow is the input signal, blue is the GPIO output. The phase difference indicates scheduler latency.

Thanks to Maxime Chevallier of Bootlin for providing the gpio-test demo utility during PREEMPT_RT training.

Setup diagram

Setup diagram — signal generator feeds the BeagleBone, oscilloscope measures input vs output.

Items needed

• BeagleBone Black or Green
• +5 V power adapter
• Micro SD card (1 GB or larger)
• Programmable signal generator board
• 2-channel oscilloscope
• Jumper wires

Instructions

1. Prepare the setup as shown in the diagram (set signal generator to 100 Hz, 50% duty cycle)
2. Download sdcard-beaglebone-preemptrt-demo.img.xz (~25 MB)
3. Write the image to a micro SD card using Balena Etcher
4. Insert the SD card into the BeagleBone and power on
5. After boot, the oscilloscope shows two square waves with a phase difference — yellow (input) and blue (output from BeagleBone)
6. The phase difference indicates the latency introduced by the scheduler while running the preemptrt-gpiotest program

On startup, /etc/init.d/S99PreemptrtGpioTester invokes:

chrt -r 99 /usr/sbin/preemptrt-gpiotest gpiochip0 28 gpiochip0 17 h p

This starts the gpio-test utility at real-time priority 99.

Measuring latency under load

The gpio-test program mirrors a GPIO input to an output — the phase difference between the two signals shows how fast the scheduler delivers context to your application. Jitter on the output signal (measured in oscilloscope persist mode) indicates the best and worst case scheduler latency under different load conditions (network traffic, interrupts, etc.).

Persist mode — showing best and worst case scheduler latency under load.

For further details on PREEMPT_RT, see Bootlin's PREEMPT_RT training material.

Building from source

Detailed build instructions are available in the git repository. Cross-compile instructions for modifying and running a customized gpio-test utility are also available.

SOURCE CODE

github.com/hackboxguy/preemptrt-gpiotest — GPIO latency test program

github.com/hackboxguy/br-wrapper — Buildroot config and build instructions

sdcard-beaglebone-preemptrt-demo.img.xz — pre-built SD card image (~25 MB)

Related posts

• CAN Bus Multi-Sensor Node with ESP32 and KiCad PCB
• AWS IoT on Raspberry Pi with 25 MB Buildroot Linux
• DIY Passive PoE for Raspberry Pi Under $2