Decentralized, distributed and fault-tolerant context recognition architectures for smart cities: A systematic mapping
Leonardo Vianna do Nascimento, José Palazzo Moreira de Oliveira · SBC Reviews on Computer Science · 2021
An essential aspect of smart city applications is how to obtain context data about their users and understand them. Several works in the last decades have presented context recognition software architectures that can be applied in smart cities. This study aims to overview the state of the art in distributed context recognition software architectures suitable for smart cities and identify research opportunities. The researchers conducted a systematic mapping that provides an overview of such architectures applied in the smart cities domain. This review focuses on the following aspects of these approaches: decentralization, plugin support, resilience, data fusion, dynamic composition, privacy, and security. The researchers searched for relevant papers using four search engines (ACM Digital Library, IEEE Xplore, Scopus, and Springer Link). The papers returned by this search passed by a two-step filtering process: the analysis of title, abstract, and keywords; and an analysis of sections that describe the proposed solution. The selected papers include 87 works of the total 1977 papers returned in the search. The analysis of these papers has shown that only a few works explored resilience, data fusion, security, and privacy, and fewer integrated these aspects in decentralized architecture with plugin support and dynamic composition. The development of architectures that integrate these aspects is a research opportunity, especially those that present resilience and security-related mechanisms.