Task Offloading in SDN-Fog Architecture Utilizing Hybrid Deep Learning Models for Internet of Medical Things

Abstract

The critical, complex, and dynamic nature of Internet of Medical Things (IoMT) devices poses significant challenges to traditional network architectures, particularly in terms of security, latency, throughput, and energy. Identifying efficient fog nodes and minimizing unnecessary data transmissions further complicates these issues. Software-defined networking (SDN), with its decoupling of control and data planes, offers a robust solution to address the heterogeneous and dynamic nature of IoMT systems. This study integrates deep learning into an SDN controller to improve fog-node selection and task scheduling in consumer IoMT networks. Using patient medical histories, the DL model improves data transfer efficiency to the fog layer and strengthens general security measures. The proposed scheme aligns with emerging trends in healthcare, incorporating advanced technologies such as DL to improve patient care, optimize resource management, and protect sensitive data. The simulation results indicate that the proposed method outperforms previous techniques, achieving a low latency of 10 ms, an efficiency energy of 60 W, and an optimal Fog Node (FN) selection rate of 90%. These results highlight the significant potential of the Internet of Medical Things (IoMT) in facilitating smart, energy-efficient, and low-latency delivery of services.