Seguridad para redes IoT usando Machine Learning : implementación de un sistema de detección de amenazas utilizando redes definidas por software.

Abstract

This study presents the process of the design of a threat detection system for simulated SDN network in Cisco Packet Tracer, starting with a topology that facilitates the classification of frames which are sent and monitored by an SDN controller. Python programming language is used for communication with the network controller’s API and datasets processing that will be used for designing and implementing prediction models based on Machine Learning. In the first chapter of this project, all theorical foundations about IoT networks, cybersecurity and Machine Learning oriented towards prediction models are studied and explored to start the second chapter that encompasses the project methodology, defining tools, resources, and Python scripts to capture traffic data from the simulator, creating a dataset and evaluating it with different prediction models. Finally, in the last chapter, a comparison is made between all proposed algorithms and models according with the theorical stages, drawing conclusions and providing recommendations based on the gained knowledge during the development of this project.