A novel cloud and fog-based architecture to support spatial analytics in smart cities

• Autor(a) principal: João Paulo Clarindo dos Santos
• Data de Publicação: 2021
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Biblioteca Digital de Teses e Dissertações da USP
• Texto Completo: https://doi.org/10.11606/D.55.2021.tde-21012022-170246

Providing an infrastructure to accommodate a large number of people in cities is a major challenge for public authorities and private companies. Thereby, the concept of smart cities emerged, which use technologies like sensors and Internet of Things (IoT) devices to aid in urban growth. These devices generate spatial data that can be used for spatial analytics by smart city managers to improve the populations quality of life. However, these IoT devices quickly generate a large volume of spatial data, causing big data problems. A smart city manager can benefit from using concepts such as fog computing, spatial data warehouses, data lakes, and parallel and distributed storage and processing environments to handle this massive amount of data. Based on a systematic review, there are no studies in the literature that consider all of these concepts in the context of smart cities. Therefore, we propose a novel architecture that aims smart city managers in spatial analytics. This architecture is composed of four layers: (i) terminal, which consists of a network of IoT devices; (ii) fog computing, which contains data lakes for real-time data processing; (iii) cloud computing, in which spatial data warehouses are used to support SOLAP (Spatial Online Analytical Processing) queries carried out in batch; and (iv) analytical tools, which incorporate data visualisation and analysis tools. Furthermore, we introduce a set of guidelines to aid smart cities managers to implement the proposed architecture, by describing and discussing important issues and examples of tools and technologies. The proposed architecture and guidelines were validated through two case studies that use real data generated by IoT devices disposed in smart cities. We investigated the execution of three categories of spatial queries, as well as the execution of queries in the fog, in the cloud, and in both environments. These case studies demonstrated the architectures efficiency and effectiveness to support spatial analytics in the context of smart cities.