Design, Implementation, and Assessment of a Testbed for Scalable IoT Architectures for Smart Renewable Energies Systems

Abstract

This paper presents a comparative performance evaluation of three containerized Internet of Things (IoT) middleware platforms, Dojot, FIWARE, and Node-RED, deployed to support a photovoltaic Digital Twin within the EMOB-AMAZON research project. Moving beyond descriptive resource monitoring, the study combines descriptive metrics with inferential statistical analysis based on pooled experimental observations. Pearson’s correlation coefficient was used to investigate the relationships between infrastructure resource utilization and application-level performance under nominal and stress workloads. The results indicate that the evaluated platforms exhibit distinct resource utilization and scalability profiles. FIWARE maintained the highest realized throughput together with stable sub-millisecond latency under the evaluated workloads, while requiring higher CPU utilization than the other platforms. Node-RED consistently exhibited the lowest infrastructure resource consumption, making it well suited for resource-constrained Edge deployments, although its scalability decreased under the highest evaluated workload. In contrast, the evaluated Dojot deployment showed higher idle resource consumption together with reduced throughput, increased latency, and elevated request timeout rates during workload execution. Overall, the combined descriptive and inferential analyses provide complementary evidence for understanding the relationship between middleware architecture, resource utilization, and application-level performance, supporting middleware selection according to the computational requirements of photovoltaic Digital Twin applications.