A Secure AI Framework for Intelligent Traffic Prediction and Routing in SDN-Based Consumer Internet of Things

With the rapid increase of Consumer Internet of Things (CIoT) and advancements in communication technologies, both are generating a huge amount of imbalance data. Traditional network architectures struggle to handle the complex and heterogeneous nature of CIoT devices, as well as the imbalance and unpredictability of traffic flows. Software Defined Networking (SDN) is a novel networking paradigm. By decoupling the data plane from the control plane, it efficiently manages the complexity and heterogeneity of CIoT devices. However, challenges such as imbalance data security, future traffic load prediction, and optimized routing still persist in CIoT environment. Advancements in Deep Learning (DL) algorithms, along with their extensive application in CIoT domain, have enabled resolving SDN-CIoT security and performance issues. To address the above mentioned challenges, in this article, we propose an AI-based framework which comprises two modules: 1) DL-based security and traffic load prediction module and 2) DRL-based routing optimization module. In addition, the proposed framework employs the CNN based intelligent load balancing strategy among SDN controllers to reduce the computational burden on the main controller. The performance of the proposed model is evaluated through simulation and results demonstrate that the proposed framework achieved excellent performance compared to the state-of-the art methods.