A methodical examination of single and multi-attacker flood attacks using RPL-based approaches

Abstract

The Internet of Things (IoT) constitutes an intricate network of physical entities, ranging from vehicles to home appliances, each ingrained with electronics, software, sensors, and internet connectivity to facilitate data exchange and collection. This novel realm of interconnectivity, while yielding considerable advantages, also invites concerns over cybersecurity, as the vast amount of sensitive data gathered by IoT devices necessitates safeguarding against potential cyber breaches. In this context, the focus of cybersecurity in IoT involves the deployment of diverse technologies, standards, and optimum practices including, but not limited to, encryption, firewalls, and multi-factor authentication. Hence, while IoT contributes significantly to societal advancement, addressing the associated security concerns remains an imperative task. This study therefore delves into an examination of flood attacks, a prevalent form of cyber assault aimed at IoT devices. The study explored the ramifications of such an attack on an IoT system by analyzing network traffic in scenarios of singular and multiple attackers. A benchmark model devoid of an attack was employed for comparative purposes. To circumvent additional stress on the operational system, network packets were mirrored through cloud infrastructure and subsequently relayed to artificial intelligence (AI) and forensic analysis tools for real-time examination. To assure the integral cybersecurity component of continuity within IoT systems, the attacking entities were identified through AI, and forensic tools were employed to conduct real-time data analysis, thereby enabling continuous network monitoring. This study introduces an innovative approach to detecting flood attacks on IoT systems by leveraging a novel AI-based technique that integrates the '6LoWPAN.Pattern' feature, previously unexplored in this context. Extensive simulations were conducted to analyze the impact of flood attacks using both single and multiple-attacker scenarios. Our method demonstrated a 99.9% success rate in attacker identification, distinguishing it from existing techniques. This research contributes to the cybersecurity field by enhancing the detection capabilities within IoT networks and offering a robust model against a prevalent form of cyber threats.