Análise do consumo de recursos na execução dos algoritmos Dijkstra e A* em nós simulados de névoa e nuvem para o gerenciamento de tráfego com semáforos inteligentes

Abstract

This study aims to analyze the performance and resource consumption of simulated fog and cloud computing nodes responsible for managing vehicle traffic in urban environments with smart traffic lights, through the calculation of shortest paths. The methodology involved the use of Docker containers to represent the fog environment, with limited hardware resources, and the cloud environment, with expanded computational capacity. The Dijkstra and A* algorithms were used as workloads in these scenarios to determine the shortest path between predefined points, as well as to evaluate their performance and computational resource usage during execution. Data were collected regarding the algorithms’ execution time, CPU and RAM usage, and the inbound and outbound traffic on the containers’ network interfaces, in order to identify which algorithm is more suitable for each context based on defined criteria. The results indicate that the A* algorithm is more suitable for the fog computing context due to its consistent performance, and also for the cloud context, for offering greater resource efficiency.