Publication
Analyzing and Predicting the Power Consumption of a Publish/Subscribe IoT-Broker
Franc Pouhela; Maryam Arabshahi; Hans Dieter Schotten
In: 2024 IEEE Cybermatics Congress. IEEE International Conference on Internet of Things (iThings-2024), Everything Interconnected via Cyberspace, August 19-22, Kopenhagen, Denmark, IEEE Xplore, 8/2024.
Abstract
The rapid expansion of the Internet of Things (IoT) has underscored the need for energy-efficient communication protocols, with the Publish/Subscribe model standing out for its efficiency and scalability. This paper presents a comprehensive analysis of the power consumption patterns of a Publish/Sub-scribe IoT broker, identifying key factors influencing energy consumption. By employing advanced predictive modeling techniques, we developed accurate power consumption models. Our findings highlight the significant impact of message throughput, thread count, and network conditions on power consumption. By comparing a variety of predictive models using different algorithms, we can identify the most suitable model for the task. These predictive models can assist in designing energy-efficient IoT systems, thereby contributing to the development of sustainable IoT infrastructures.