Modeling a Fog Computing Network Architecture for Secure IoT Data Processing

Abstract

This article presents a fog computing network architecture designed for secure processing of Internet of Things data. The proposed architecture consists of four layers: the cloud layer – a central server for analytics and long-term data storage; the proxy layer – responsible for caching, routing, and load balancing; four fog nodes that provide real-time data processing; and the device layer – representing Internet of Things endpoints. This structure enables flexible load distribution and enhances the system’s resilience to component failures. To verify the effectiveness of the developed architecture, simulations were performed in the iFogSim environment. Scenarios were created with different numbers of smart cameras (from 16 to 48). The modelling results showed that, with 16 cameras, the data processing latency of the proposed architecture was 286 ms, while in the traditional cloud-based architecture was 811 ms. These results demonstrate an overall 64.7% reduction in data processing latency in the developed architecture. The fog computing network architecture also achieved a 17-fold reduction in network resource utilization under minimum load (16 cameras) and a 4.3-fold reduction under maximum load (48 cameras). This translates to up to 90% savings in bandwidth and a significant decrease in the risk of network congestion. The proposed architecture ensures a high level of protection of users’ personal data through the local processing of video streams. Sensitive information is processed on fog nodes without being transmitted to external networks, which minimizes the risk of personal data leakage. The modeled fog computing network architecture provides a solid foundation for further development of fog computing technologies in the Internet of Things domain.