Um estudo da arquitetura Edge-Fog-Cloud com foco na otimização distribuída do processamento e na redução da latência

Abstract

This study investigates the effectiveness of Edge-Fog-Cloud architectures in distributed environments, focusing on optimizing data processing and reducing latency in critical applications such as smart cities, factories, and hospitals. The main objective is to assess how different configurations of these architectures influence the system’s responsiveness in situations that require real-time data processing and agile decision-making. This work explores the dynamics of large-scale data processing and communication by using a methodology that involves simulating environments with a variation of 1 to 25 Edge devices and maintaining a constant number of three Fogs and a single Cloud. The employed method includes the analysis of different databases, each representing a specific type of workload in terms of data volume and complexity, from simple health parameters to complex industrial monitoring data. By quantifying response time and the system’s overall performance in various configurations, it was possible to identify critical saturation points and efficiency in the Fog and Cloud layers. The results show that while Fog Computing offers significant improvements in latency and local data processing, its capacity is limited by the amount of available computational resources, leading to the need for scalability and proper planning of Fog. The study’s conclusions highlight the importance of a well-dimensioned and adapted Edge-Fog-Cloud infrastructure to the specific needs of each application. Furthermore, it is emphasized that the potential independence from the internet connection offered by Fog Computing can be decisive in environments with limited or unstable connectivity and requiring critical decision-making responses.