In this subchapter we will move the processing from our sensor and actuator nodes to the Raspberry Pi, acting as our fog node.

To do so we will use custom functions, written in javascipt, in Node-RED.

The Video below gives you an overview of how a function node is added to Node-RED.

This is the code snipped pasted into the function block 45 seconds into the video:

var led_status= context.get('led_status');

if (led_status === undefined) {
    led_status = 'off';
}

if (msg.payload === 'toggle') {
    if (led_status === 'on') {
        led_status = 'off';
    }
    else {
        led_status = 'on';
    }
}

context.set('led_status', led_status);

var out = {'payload': led_status};

return out;

Task: Recreate the flow shown in the video and, by reading the sourcecode and analyzing the behaviour, find out what the function node does.

Task: Change the function block so that, in addition to the toggle functionality, injecting on and off messages sets the LED state to on and off, regardless of previous state.

Now we want to integrate the function block into our hardware setup.

Task: Edit the program on your publishing microcontroller to send pressed to the /button topic whenever the button input goes from unpressed to pressed.

Task: Edit the function block in Node-RED to toggle the LED state when the string pressed is received instead of toggle.

Task: Add MQTT subscription and publishing nodes to the flow so that pressing the button on the publishing microcontroller toggles the LED on the subscribing microcontroller.

Task: Add a delay node to delay the reaction between pressing the button an changing the LED by two seconds.

TODO: Do more complicated stuff with, for example, delays or triggers.

TODO: Change publisher to read another sensor, like ultrasonic distance or temperature and publish measurements in a generic format like cm or °C.

TODO: Change subscriber to control another actuator than the LED and take commands in a generic format.