Titolo della tesi: Edge2LoRa: A New Paradigm for Enabling Cloud Edge Computing Continuum over LoRaWAN
Low-Power Wide-Area Network (LPWAN) offer a low-cost solution for deploying large-scale Internet of Things (IoT) infrastructures by adhering to a classic producer/consumer model with minimal requirements. However, as these deployments scale, the need for low-latency, distributed, and collaborative data aggregation models becomes inevitable. The Cloud Edge Computing Continuum (CECC) has been proposed as an evolution of traditional centralized, ultra-high-end processing clouds into a continuum of collaborative processing elements, distributed from the cloud to the network edge. Integrating existing centralized and monolithic LPWAN architectures into the CECC framework, however, poses several security challenges. To address these challenges, I propose Edge2LoRa (E2L), a comprehensive and secure solution that integrates LPWAN architectures into the CECC. This integration facilitates faster data processing while minimizing the transmission of sensitive data. E2L enhances network performance by enabling data pre-processing, optimizing traffic flow, and providing real-time local analysis. By progressively transforming existing LPWAN deployments into agile and versatile infrastructures, E2L ensures seamless and efficient data processing across the CECC while maintaining service continuity and full backward compatibility. The E2L implementation is available as open-source and is compatible with hardware that adheres to the Things Stack and Long-Range Wide-Area Network (LoRaWAN) v1.0.4 and v1.1 specifications. The system’s performance was evaluated in terms of network and computing resource utilization, Quality of Service (QoS), and security. Results demonstrate significant improvements in both public and private LoRaWAN infrastructures, without any disruption or degradation of existing legacy services. In particular, for public LoRaWAN deployments supporting large-scale IoT data streams and big data analytics, a reduction in core network bandwidth usage by up to 90% is observed, along with data processing latency improvements of up to 10×. Additionally, as a further contribution, Edge4LoRa (E4L) incorporates a distinct computing module capable of processing data streams at the network edge. This module utilizes a Map/Reduce engine based on Apache Spark, enabling the execution of various processing applications, including anomaly detection and data reduction techniques. Furthermore, E4L enables efficient traffic routing across LoRaWAN gateways, addressing the dynamic nature of IoT data traffic and the mobility of source devices. The proposed architecture ensures modularity, reliability, scalability, and robustness. The evaluation demonstrates its effectiveness under different testbed configurations. Performance assessments were conducted using a hardware setup in a laboratory, and we evaluated the architecture across three scenarios: data reduction, scaling activation of edge gateways, and mobility-aware scenarios.