Skip to content

Latest commit

 

History

History
90 lines (56 loc) · 3.12 KB

MQTT.md

File metadata and controls

90 lines (56 loc) · 3.12 KB
Raw

squeeze-alexa with MQTT

MQTT is a lightweight pub/sub messaging protocol with similarities to AMQ, used a lot in IoT situations.

Amazon AWS now has good (and mostly free) support for this via AWS IoT.

With MQTT transport, we'll be using this to act as the interface between your Alexa skill and your LMS server.

squeeze-alexa needs a tunnel (conceptually similar to stunnel or HAProxy for SSL) to relay traffic. For this you can use mqtt-squeeze (BETA).

Set up mqtt-squeeze

  • This needs to run on a server (typically the same as your LMS one).
  • Needs Python 3.6 (or 3.5 maybe), just like the skill (in fact it shares some code)
  • Admin access to your server (e.g. via SSH).

Copy the files

  • You'll need the
    • the script, mqtt_squeeze.py
    • the etc/ directory with your certificates (see below)
    • the squeezealexa directory

TODO: script this too.

Create a new directory somewhere on your server and copy these there. For Synology, I've chosen /volume1/mqtt-squeeze Make sure mqtt_squeeze.py is executable (chmod +x mqtt_squeeze.py).

Create a service

  • To start and stop this, it's best to use your OS's service manager. On Linux, this might be SysV (traditional), Upstart, or systemd (most modern Linux).

Using Upstart on Synology

For convenience find an Upstart script suitable for Synology, which you can copy to /etc/init/mqtt-squeeze.conf.

You can then reload the daemon: sudo initctl reload-configuration.

You can then start it with: sudo start mqtt-squeeze

And the status with: sudo initctl status mqtt-squeeze

...or manually

You can just do it oldschool and run nohup mqtt_squeeze.py &, but you'll have to do this every time your server starts, and it doesn't take care of connections dying like Upstart etc.

Set up MQTT with Amazon IOT

Create a new certificate

This convenient AWS CLI command will create the certs in the right place (assuming you're in the project root, e.g. ~/workspace/squeeze-alexa/). You'll need to be logged in first, as with all the other aws commands. Use --profile if you've got lots of accounts.

aws iot create-keys-and-certificate --set-as-active --certificate-pem-outfile etc/certs/iot-certificate.pem.crt --private-key-outfile etc/certs/iot-private.pem.key

Set up permissions for MQTT

Go to the AWS IoT section (make sure to select the right region), and you start the setup.

You'll need an IAM policy to grant MQTT access to the squeeze-alexa Lambda.

Use the helpful included IAM policy to permission topics - remember to make sure these match your MQTT settings.

Test

You can use local_test.py to test once your settings are configured. You can also debug using AWS IoT test console.

Troubleshooting

TODO: Add troubleshooting