Avaliação de aglomerado de computadores de placa única que utilizam Apache Flink

Abstract

The advancement of the Internet of Things has driven the demand for computational solutions capable of processing large volumes of data with low latency. In this context, Fog Computing emerges as an alternative to Cloud Computing by bringing processing resources closer to end devices; however, its implementation often entails high costs. This research aims to verify the feasibility of using Single-Board Computer (SBC) clusters as a low-cost, energy-efficient alternative for distributed applications in Fog Computing. In this work, the performance of clusters composed of Raspberry Pi 4 nodes was evaluated using the Apache Flink framework and Docker Swarm orchestration. To achieve this, an SBC cluster and a traditional server cluster were implemented to perform a continuous data stream processing test, allowing for a comparison of energy consumption during processing. The results showed that the cluster composed of Raspberry Pi devices achieved superior energy efficiency compared to the traditional server.