SDN-Assisted Spatial Encoded Sequence Enabled BLSTM-Based Zero-Trust Anomaly Detection Model for Consumer Electronics of Smart Cities

The proliferation of Consumer Electronics (CE) within smart cities has introduced both unprecedented convenience and significant cybersecurity vulnerabilities. The growing complexity, heterogeneity, and increasing number of CE devices have led to a significant increase in data flow and security issues. Furthermore, traditional static network infrastructure methods need customized CE device administration and human configuration. This research presents a unique Deep Learning (DL) system that is coordinated by Software-Defined Networking (SDN) to address these issues. The proposed novel model, leveraging SDN and DL, efficiently manages the heterogeneity of CE networks and detects attacks with high accuracy. First, by separating the control and data planes, the SDN architecture is used as a versatile solution that allows reconfiguration over static network infrastructures and manages the distributed nature of smart city CE networks. Second, integrating spatial-temporal context encoding with a Bidirectional Long Short-Term Memory (BLSTM) network captures both spatial and temporal dependencies. The BLSTM bidirectional processing improves anomaly detection by analyzing patterns from past and future states. The proposed method is validated by simulation results using the CICDDoS-2019 dataset, showing that it outperforms current state-of-the-art security methods and is a viable choice for next-generation smart city CE networks.