Desarrollo de un sistema de detección de ataques de denegación de servicio (ddos) en redes iot, con especial enfoque en el ataque udp flood, utilizando técnicas de machine learning y el conjunto de datos ton iot como base

Abstract

Since the creation of IoT networks, the number of devices connected to them has grown exponentially, and likewise, the exposure to vulnerabilities for devices connected to an IoT network has increased. According to DataReportal, in 2024, Colombia has 39.51 million internet users, with a penetration rate of 75.7%. Additionally, 77.02 million active mobile connections were registered, exceeding the country’s total population by 147.5%, reflecting the high level of connectivity and the rapid adoption of IoT devices across various areas. One of the main security concerns for IoT networks is distributed denial-of-service (DDoS) attacks, among which the UDP Flood attack stands out. This type of attack seeks to overload a server’s resources through multiple requests, severely affecting service availability. Furthermore, in relation to 5G infrastructure, the Ministry of ICT (MinTIC) has authorized several companies to conduct pilot tests in cities such as Bogotá, Medellín, and Cali to evaluate private networks and IoT applications. These trials include temperature sensors, thermal cameras, and remote health monitoring, demonstrating how IoT is increasingly integrated across diverse sectors, thereby expanding cybersecurity requirements.