Abstract
The Internet of things (IoT) is an all-encompassing web of devices whose evolution is reaching unimaginable levels of convenience and data-driven insights. IoT networks are becoming more transformative, and this is happening amidst concerns on their security, especially when routing attacks are involved and pose threats to data integrity as well as the entire system security. To overcome these concerns, we propose a comprehensive and multi-layered security model for IoT routing that is designed specifically for this task. The main elements of our proposed solution comprise a solid data validation, encryption, and intrusion detection system – which are the cornerstone of IoT networks’ reliability and security. However, IoT systems are resource-limited in essence, which means that the approaches different from the traditional ones should be sought. RSA encryption with SHA3-512 integrity verification algorithm for the purposes of secure data transmission is the method we apply. A Thermal Exchange Optimization (TEO) algorithm is used to deal with the generation of optimal cryptographic keys, hence outperforming the traditional key generation method. Intrusion detection, as a fundamental element of IoT security, is tackled through deployment of a Convolutional Neural Network (CNN) solution. An intrusion detection system that is additionally enhanced using a Q-learning-based Whale Optimization Algorithm is being used to make the system more precise and efficient in adapting to the dynamic IoT environment. A complete set of simulations was performed with Python to ensure the efficiency of our proposed solution. This evaluation report shows that our approach is more resistant to the routing attacks, and, moreover, it surpasses the existing model in both security and functionality. This study is not just focused on the issue of routing security in IoT but also presents a highly reliable and innovative solution, thereby, contributing to the ongoing discussions on the security of the network infrastructure of IoT