Use Case 4 - CyberSecurity

UseCase 4 of the project focuses on Cybersecurity and highlights the utility of the COGNIT framework through the implementation of an anomaly detection scenario within a fleet of rovers (vehicles). The current architecture of the scenario includes the following components: Rover Data Collection: Rovers collect system logs data. Data Transfer: The collected data is transmitted to the anomaly detection, which is deployed at the cluster level. This step aims to ensure fast (low latency) and secure transmission of data to the detection system. Anomaly Detection: Anomaly detection is performed using a Serverless Runtime. This component analyses incoming data to identify any significant deviations from normal behaviour patterns. Migration Management: A crucial aspect of UseCase 4 is to demonstrate the COGNIT framework's ability to manage the migration of serverless runtimes based on the itinerary and movement of the rovers. This functionality ensures service continuity and operational efficiency even in dynamic and constantly evolving environments.

Here is a high level diagram of the architecture :

We implemented an architecture represented in the diagram, which illustrates the interaction of our use case with the framework.

In the examples/ folder one can find the example of the anomaly detection functionality. Refer to examples README.md file for further information.

Technical overview

This repository contains the Python implementation of the Device Runtime, an SDK designed to enable devices to communicate seamlessly with the COGNIT platform for task offloading. The Device Runtime provides a highly abstracted layer of communication between the COGNIT platform and end devices, simplifying the interaction process for users.

At the core of this abstraction is a state machine, which autonomously manages all stages of communication without requiring direct user intervention. This allows for efficient and automated handling of the different states involved in the task offloading process.

The communication with the Device Runtime involves two key components: the Cognit Frontend Client and the Edge Cluster Frontend. The Cognit Frontend Client allows the Device Runtime to offload functions that the device may wish to execute in the future, along with a comprehensive set of requirements, policies, and dependencies. Meanwhile, the Edge Cluster Frontend is responsible for executing those previously uploaded functions.

Developer Setup

This repository has been built using Python v3.10.6, therefore it is highly recommended to maintain this Python version for development purposes.

For setting it up it is recommended installing the module virtualenv or, in order to keep the dependencies isolated from the system.

pip install virtualenv

After that, one needs create a virtual environment and activate it:

python -m venv runtime-env
source runtime-env/bin/activate

The following installs the needed dependencies from the requirements.txt file:

pip install -r requirements.txt

Setting up COGNIT module

To set up the COGNIT module the following needs to be executed:

python setup.py sdist

In such a way that for installing it in an empty environment, one should:

pip install dist/cognit-0.0.0.tar.gz

Once done that, COGNIT module's installation will be fully completed. Now is possible to instantiate device runtimes in the same way as follows:

from cognit import device_runtime

my_device_runtime = device_runtime.DeviceRuntime("./examples/cognit-template.yml")

User's manual

There are several folders that might be interesting for a user that is getting acquainted with COGNIT:

Examples

The anomaly detection function can be found in the examples/ folder, see README.md.

Configuration

The configuration for your COGNIT Device Runtime can be found in cognit/test/config/cognit.yml, with an example for running the tests.

Tests

The cognit/test/ folder holds the tests for the COGNIT module. More info about how to run them in the README.md file.