Illuminator Cluster Set Up

The Illuminator cluster setup is an optional execution of Illuminator scenarios in different Raspberry Pis. This offers a more hands-on approach to running Illuminator cases with the option to have one or multiple models of a scenario running on different Raspberry Pi devices. The Illuminator Cluster is also used for the Illuminator Demonstator Kit, a physical setup in which flows and interactions between components of a specific scenario are visualised for educational purposes.

Master-Worker Raspberry Pis

For the cluster setup the Raspberry Pis need to be connected to the same network. The master Raspberry pi holds the yaml file of the scenario to be run. It connects to the worker Raspberry Pis through ssh and “tasks” them with running specific models of the scenario.

alt text Example setup of a Raspberry Pi cluster running the Illuminator

All Raspberry Pis can run multiple models, specifically they can run one model per available port.

Raspberry Pi Setup

The following 2 steps need to be done to all Raspberry Pis, in the future the Install Illuminator step will also be done through the master to make deployment smoother.

Flash SD card

Install a Raspberry Pi imager and flash the SD card with Raspberry Pi OS. In the config file of the imager add the following configurations:

Parameter

value

Username

your_username

Password

your_password

ssh

Enabled

Your defined username and password will be needed to connect to the worker Raspberry Pis through your master. Note that the username field is case sensitive.

Install Illuminator

Clone the github repository by running git clone https://github.com/Illuminator-team/Illuminator.git. Enter the Illuminator directory and switch to the cluster branch by running git checkout cluster.

Note that the Illuminator directory needs to have a capital I.

Master Pi Setup

The following steps are all conducted through the Master Raspberry Pi. First we need to ensure our connection to the worker Raspberry Pis for remote execution of our scripts.

IP Address Setup

The IPs of all Raspberry Pis need to be static for us to know where to connect. Navigate to the Illuminator/cluster_setup directory and run static_ip.py. The script will automatically make the device’s IP static.

To make the worker Raspberry Pis’IPs static navigate to Illuminator dir and edit and run setup.py. This script generates keys for ssh connections without needing a password. Make sure your username, password and IPs reflect those of the worker Pis you wish to set up.

Simulation Preparation

Normally the Illuminator scenarios are defined as yaml files. Similarly, those very same files are used in the cluster setup with some additional steps which allow us to run specific models in the worker Raspberry Pis.

Within your yaml file define the additional fields

  connect:
    ip: 145.94.213.75 # IP of the worker Pi we want to run this model
    port: 5123 # port of the worker Pi that will run this model

To prepare the scenario you wish to run in the cluster run Illuminator cluster your_scenario.yaml.

Simulation Preparation

Execute run.sh and run Illuminator scenario run your_scenario.yaml.

Warnings

  • Slaves might need to be updated after adding a model.

  • Slaves run .sh files to actually start the host (available in Illuminator/configuration/runshfile), if changed/added they need to be made executable through chmod +x <file.sh>.

This section of the documentation is a work in progress, please consult the related guide of the Illuminator Demonstrator Kit repository found here.

Note

The ideas above can be used by contributors to participate in the development of the Illuminator. Feedback and solutions are always welcome. Please contact the Illuminator Development Team if you would like to contribute.