Testbed de arquiteturas IOT escaláveis para aplicações de sistemas inteligentes

Abstract

This dissertation presents the design, implementation, and evaluation of Internet of Things (IoT) architectures for applications in intelligent systems, using a testbed developed in the context of the Research and Development (R&D) Project Technological Development and Innovative Extension of Advanced Materials in Energy and Mobility applied to the Amazonian context (EMOB-AMAZON). The proposal seeks to offer a modular and scalable approach, based on an architecture composed of four layers: physical (hardware), middleware, processing, and software. To this end, virtualization (Proxmox VE) and containerization (Docker) technologies were employed in the implementation of four virtual machines. Three of them were used to configure development environments based on the Dojot, FIWARE, and Node-RED IoT platforms, while the fourth was dedicated to the implementation of an observability and monitoring system, structured from the Prometheus-Grafana stack, which includes the Prometheus, Grafana, InfluxDB, cAdvisor, and Grafana k6 tools. Scalability was assessed through continuous monitoring of technical system metrics (CPU, RAM, disk, and network), while the robustness of the architectures was verified with load tests that simulated typical and extreme usage scenarios. The results revealed relevant operational and performance aspects, demonstrating that the platforms maintained acceptable consumption levels, according to the hardware resources allocated to each virtual machine, in addition to presenting stability and latency compatible with the simulated scenarios and their respective established thresholds, highlighting patterns, advantages, and limitations of each development environment. Finally, the analyses provided important technical insights for understanding the scalability and robustness of the deployed IoT architectures, supporting both the sizing and targeting of each IoT solution within the scope of the project and the expansion of the observability and monitoring system.