Skip to content

A LoRaWAN Security Evaluation Framework

License

Notifications You must be signed in to change notification settings

leinher/chirpotle

 
 

Repository files navigation

ChirpOTLE

ChirpOTLE is a practical LoRaWAN security evaluation framework that provides tools for the deployment and management of a LoRa testbed based on COTS hardware. It allows managing LoRa field nodes from a central controller and to orchestrate experiments and tests using a Python 3 interface.

By collocating the nodes with a LoRaWAN network, the built-in functions for receiving, transmitting, jamming, and sniffing can be used to study their effects on the network under test. With its dissector and predefined building blocks like wormholes, the framework allows for quick vulnerability assessment in LoRaWAN networks as well as for the evaluation of attempts for their mitigation.

Basic Setup

The framework is managed through the chirpotle.sh shell script. It creates and manages a virtual environment, node configurations and it takes care of creating the RPC stubs for communication with the nodes.

To get started with the controller, just run the install task and you are ready to go:

./chirpotle.sh install

If you want to be able to use the framework independent of the working directory, you can add it to your .bashrc:

echo "alias chirpotle=\"$(pwd)/chirpotle.sh\" \"\$@\"" >> ~/.bashrc

You will then be able to call it from any directory like:

chirpotle interactive

By default, the virtual environment is created in the env folder in the repository and the configurations are stored in conf. If you want a clean install, you can delete the env folder without losing your configuration.

Configure Your Setup

As ChirpOTLE is meant to be deployed in the field, it uses a star topology with the controller in the center and nodes in the field. Controller and field nodes are connected through SSH for deployment and RPC calls for control during experiments.

Note: The tool assumes a secure network connection between controller and nodes, the RPC traffic is not secured and SSH host keys are trusted by default.

For this description, we assume you have the following setup:

  • The controller is running on your desktop machine, and you've installed ChirpOTLE as mentioned above
  • You have two Raspberry Pis running a fresh installation of Raspbian, and they are available as loranode1.example and loranode2.example
  • To each of the Pis, you've got a LoPy 4 connected via a USB-to-Serial bridge, so it's available as /dev/ttyUSB0

Note: Other LoRa hardware can be used with the framework, see Makefile.preconf of the companion application for more options.

Prepare the Nodes

First, you need to make the nodes available for root access via SSH. Copy your SSH key to /root/.ssh/authorized_keys on each of the Raspberry Pis.

Note: Your SSH public key is usually located in ~/.ssh/id_rsa.pub. If that file does not exist, run ssh-keygen -t rsa -b 4096 on your desktop machine.

After having deployed the keys, you can start installing Python 3 and pip on the Pis. All other software installation will be managed by the framework.

ssh [email protected] apt-get update
ssh [email protected] apt-get update
ssh [email protected] apt-get install python3 python3-pip
ssh [email protected] apt-get install python3 python3-pip

Create a Configuration

Now it is time to setup the configuration, so that your controller knows which nodes are avaialble.

All configurations are stored in the conf folder of the repository after running chirpotle.sh install, but the easiest way for most cases is to use the interactive editor.

Run ./chirpotle.sh confeditor, and you'll be greeted with the main menu:

What do you want to do?
        
   📝  List/edit controller configurations 
   📝  List/edit node profiles
   Save changes and quit 
  • Select List/edit controller configuration.
  • Select Create new configuration
  • Enter the name testconf
  • Select Add Node
    • Name: alice (This is how you will address the node in scripts)
    • Host: loranode1.example
    • Node profile: uart-lopy4
  • Select Add Node
    • Name: bob (This is how you will address the node in scripts)
    • Host: loranode2.example
    • Node profile: uart-lopy4

Your configuration should now look like this:

   🖥️  Node: alice (loranode1.example, uart-lopy4) 
   🖥️  Node: bob (loranode2.example, uart-lopy4) 
   ➕ Add node                                          
   🏷️  Rename this confiugration                       
   Delete this configuration                           
   Go back 

If everything seems alright, select go back in all menus, and your configuration will be saved as testconf.

Deploy ChirpOTLE to the Nodes

Now, you can test if everything is set up correctly and if the nodes have all required software installed:

./chirpotle.sh deploycheck --conf conftest

Note: Most of the CLI commands support the --conf option to select the configuration that you want to use. If you omit this option, the CLI will try to use a configuration with the name default.

You should now see green checkmarks for everything that is working, a warning sign for optional software that is not required in every case and a red x for unfulfilled requirements. If you see errors, please re-check the instructions above.

If all requirements are fulfilled, you can start deploying ChirpOTLE to the nodes:

./chirpotle.sh deploy --conf conftest

This command will do the following:

  • Build and bundle the TPy node in submodules/tpy/node
  • Bundle the add additional ChirpOTLE modules in node/remote-modules
  • Build and bundle the companion application for the remote MCUs from node/companion-app
  • For each node:
    • Copy everything to the nodes
    • Install additional software (globally, using pip)
    • If MCUs are connected to the nodes (the LoPy 4 boards in this case): Flash the MCUs with the companion application

Start the Nodes

As a last step before starting with experiments, you need to launch the node deamon on each node:

./chirpotle.sh restartnodes --conf conftest

Now you should be ready to go!

Getting Started

To familiarize with the framework, the best way is to start an interactive session. After you have set everything up, you can run the interactive task to start such a session:

./chirpotle.sh interactive --conf testconf

If you used the setup mentioned above, you can now try to communicate between both LoRa boards:

# Assure both use the same channel setup
bob_lora.set_lora_channel(**alice_lora.get_lora_channel())

# Configure IQ inversion to default values
alice_lora.set_lora_channel(invertiqtx=True)
bob_lora.set_lora_channel(invertiqrx=False)

# Set Bob in receive mode
bob_lora.receive()

# Transmit a frame
alice_lora.transmit_frame([int(b) for b in b'Hello, World'])

# Check if Bob received it
bobframe = bob_lora.fetch_frame()
if bobframe is not None:
    print("Bob received: {payload_str} (RSSI={rssi} dB, SNR={snr} dB)".format(**{
      **bobframe,
      "payload_str": "".join([chr(b) for b in bobframe['payload']]),
    }))
else:
    print("No frame received")

To run one of the example scripts, you can use the run task with a python script as parameter:

./chirpotle.sh run --conf testconf example.py

Getting Started with Jupyter Notebooks

The framework also comes with an integration for Jupyter notebooks. After creating you configuration like mentioned above, you can just run:

./chirpotle.sh deploy --conf testconf
./chirpotle.sh restartnodes --conf testconf
./chirpotle.sh notebook --conf testconf

On the first run, the notebook action will install Jupyter notebook in the virtual environment. The default notebook folder is called notebook and created in the repository root. It also contains an examples folder with notebooks that show you how to setup your experiments and how to integrate the framework with data visualization tools like matplotlib to create a seamless workflow.

System Requirements

Most of the software that is required to run the framework is managed by the framework in the virtual environment. However, some preparations have to be made to bootstrap the management.

Controller

For the basic installation of the controller, Python >=3.9 is required, pip and the venv module must be present. Earlier versions might still work but are no longer supported. Everything else will be fetched by the installer and be placed in the virtual environment.

If you do not want to use your system's default python (the installer will check first for python3, then for python on your path), you can specify the PYTHON environment variable during installation to point to a specific executable:

PYTHON=/opt/my-python/bin/python ./chirpotle.sh install

The framework was tested on Debian Bullseye and we check basic functionality in GitHub Actions for Ubuntu 22.04 and 20.04, but it should work on most other Linux distributions as well.

Nodes

Calling ./chirpotle.sh deploy will install the framework globally on the node using an SSH connection as user root. Therefore, the public SSH key of the user running the ChirpOTLE controller must be added to root's authorized_keys file on the node. Furthermore, you need to install Python3 with pip, git, make and gcc on each node. For Debian-based systems, you can run:

apt install python3 python3-pip git build-essential

You can check if your nodes meet the requirements by calling ./chirpotle.sh deploycheck. The output will also suggest quick fixes in case some of the requirements are not met.

Currently supported hardware:

Development

If you want to modify the framework, you need to install it in development mode for changes to be immediately available. Therefore, the install task supports a --dev flag:

./chirpotle.sh install --dev

If you've already installed the framework in the default virtual environment (env in the repository root), you can just delete that folder and install again.

Read Our Paper

The ChirpOTLE framework has been published at ACM WiSec '20 with our paper:

Frank Hessel, Lars Almon, and Flor Álvarez. 2020. ChirpOTLE: A Framework for Practical LoRaWAN Security Evaluation. In 13th ACM Conference on Security and Privacy in Wireless and Mobile Networks (WiSec '20), July 8–10, 2020, Linz (Virtual Event), Austria. ACM, New York, NY, USA, 11 pages. https://doi.org/10.1145/3395351.3399423

A preprint is available. Scripts and data for all experiments from the paper can be found in experiments/wisec2020. If you use our work for your research, please cite the paper:

@inproceedings{chirpotle2020,
  title     = {ChirpOTLE: A Framework for Practical LoRaWAN Security Evaluation},
  author    = {Hessel, Frank and Almon, Lars and Álvarez, Flor},
  booktitle = {Proceedings of the 13th Conference on Security and Privacy in Wireless and Mobile Networks},
  date      = {2020},
  month     = jul,
  address   = {Linz (Virtual Event), Austria},
  doi       = {10.1145/3395351.3399423},
  publisher = {ACM},
  series    = {WiSec '20},
  url       = {https://doi.org/10.1145/3395351.3399423},
}

License

We provide the ChirpOTLE framework under the GNU General Public License, Version 3. However, the repository contains (modified) third-party code and tools which has been published using a different licenses:

Component License Directory/Files
esp-idf Apache License, Version 2¹ submodules/esp-idf
RIOT GNU Lesser Public General License, Version 2.1 submodules/RIOT
TPy n/a submodules/tpy
xtensa-esp32-elf for RIOT n/a submodules/xtensa-esp32-elf
ubjson (deprecated RIOT module) GNU Lesser Public General License, Version 2.1² `node/companion-app/riot-modules/{incude/ubjson.h
ChirpStack Docker config MIT License² experiments/wisec2020/infrastructure/network/chirpstack
LoRaMAC node Revised BSD license¹² experiments/wisec2020/infrastructure/node/LoRaMAC-node/

¹ The submodule may contain submodules on its own, which again is published under different licenses, so please also check the description of the submodule.

² We modified this component and publish the changes under the same license.

Powered by

Logo: TU Darmstadt Logo: Secure Mobile Networking Lab Logo: CYSEC Logo: ATHENE Logo: emergenCITY

About

A LoRaWAN Security Evaluation Framework

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 33.9%
  • Python 28.3%
  • C 22.6%
  • TeX 10.2%
  • Shell 4.1%
  • Makefile 0.9%