In this lab, we learn how to use a publish-subscribe protocol to send data between devices. In particular, we will be using the Paho-mqtt library for Python and MQTT.fx to connect the two Raspberry Pis to each other, and to a laptop.
Before we start, just as a repetition, since it’s easy to lose the overview when so many different parts are involved:
- MQTT is the publish-subscribe protocol that we are using today.
- MQTT Broker is the software to run your own MQTT broker.
- Paho-mqtt is a library for Pyhton to run an MQTT client, which can publish messages and subscribe to topics.
- MQTT.fx is a helping program to manually publish messages and subscribe to messages.
For the Report: Last week we used HTTP. Describe in a few words the difference of MQTT and HTTP. Do you see situations where each of them has their strengths? When would you use which protocols?
Starting Your Own MQTT Broker
Goal: Install your own broker on a Raspberry Pi, and use MQTT.fx to verify that all is setup correct. Send some messages back and forth.
Tasks:
- Install the MQTT broker called Mosquitto on the Raspberry Pi.
- Find again which IP address your Pi has, you need this when connecting to your broker.
- Use MQTT.fx (on your laptop) and connect to the Raspberry Pi. Use the IP address of the Pi, and usually Port 1883.
- Use MQTT.fx on another laptop, and also connect to the broker.
- Send messages from laptop to laptop via the broker. (Select any topic you want.)
For the Report:
- Show a sequence diagram of a message exchange between laptop 1, broker, laptop 2 when you send a few messages back and forth.
- Include a screenshot of MQTT.fx.
MQTT in Python
Goal: You can publish a message to an MQTT broker from Python, and you can subscribe to an MQTT topic from Python.
Tips:
- Use the MQTT broker that you just installed on the Raspberry Pi.
- You can work with Python on the same Pi that runs the broker, or use the other one. (We will later connect both Pis, so it does not matter on which of them the broker runs.)
Tasks:
- Install Paho MQTT on the Pi. This is a Python library for an MQTT client that can connect to the broker.
- https://pypi.python.org/pypi/paho-mqtt/1.1
- Write a simple Python program that subscribes to a topic, and prints out whatever you sent as payload in your message.
- Send something to the chosen topic from your laptop using MQTT.fx. Can you see the output in your Python program?
- Write another simple Python program that publishes a message with some payload to a topic.
- Subscribe with MQTT.fx on your laptop to that topic and check if you can receive the message sent via Python.
For the Report:
- Include the (few) lines to publish a message and to subscribe to a topic, and document your work, also using some screenshots of MQTT.fx.
Back to the Sense HAT
Goal: Create a program on the Sense HAT that uses the joystick and the LED. We will use this program later for some remote communication.
Tips:
- The example Python programs for the Sense HAT from week one also contained a Snake-like example. Have a look at it, if it helps. Your solution can be much simpler: Just move a simple pixel around on the LED matrix.
Remote Control of the Sense HAT
Goal: You can remotely control the moving Pixel of the Pi of another team with the joystick on your Pi, and the other team can control the pixel on your screen with their joystick.
Tips:
- Use different topics for up, down, left, right. (In this way, you can ignore any payload of the message.)
- Use the joystick to send these commands as MQTT messages to different topics.
- Let the pixel be controlled by subscribing to the corresponding topics.
- Make first a version that runs with your MQTT broker, and which control your pixel with your joystick, simply by publishing and subscribing to the corresponding topics.
- Use MQTT.fx for testing, either by sending messages that should move the pixel, or by checking that the joystick triggers the correct messages.
- Agree with another team which topics to use. You can use a single MQTT broker, of either team