Communication in IoT Environments Decentralization and Resilience via Brokers

Abstract

The rapid expansion of the Internet of Things (IoT) has led to an increasing demand for communication protocols that are efficient, scalable, and reliable—especially in environments with limited resources. Among these, MQTT (Message Queuing Telemetry Transport) has emerged as a lightweight protocol particularly well suited for constrained IoT systems. However, as devices become more distributed geographically, new challenges arise in ensuring communication that is both decentralized and resilient. This thesis addresses these challenges by exploring two real-world scenarios where MQTT is used to ensure system robustness and flexibility. The first involves a critical healthcare environment, where we implemented a fault-tolerant architecture using a Master/Slave model over Apache ActiveMQ with shared persistent storage. This setup ensures automatic broker failover with no interruption in data transmission or alert delivery, which is vital in medical emergencies. The second scenario focuses on decentralization through broker bridging, using Mosquitto to interconnect remote sites—a retirement home and a hospital. In this case, critical alerts from the retirement home are selectively forwarded to the hospital’s dashboard, reducing traffic while maintaining responsiveness. Our contribution lies in the concrete design and implementation of these two architectures: a highly available broker system for resilience, and a decentralized communication model for distributed monitoring. We also explore the integration of WebSockets to deliver real-time alerts to web-based dashboards, further enhancing usability. Through this work, we demonstrate how MQTT can support both robust failover mechanisms and scalable, autonomous communication across heterogeneous IoT environments