Uma arquitetura flexível de segurança para compartilhamento de contexto em internet das coisas

• Ano de defesa: 2022
• Autor(a) principal: Martins Júnior, Francisco Luciano Castro
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Ceará
• Link de acesso: http://www.repositorio.ufc.br/handle/riufc/70001

Despite the evolution of the Internet of Things (IoT) paradigm in recent decades, concerns about security and privacy are still a necessity and a challenge, due to the heterogeneity, large amount of data and devices and their restrictions, among other aspects. Additionally, some applications have specific demands, such as context-aware applications, which can use sensitive information from different entities to provide context for various applications. Thus, it is necessary that new security features be developed. Technologies such as virtualized network functions, cloud, fog and edge computing have been explored in order to create security mechanisms. In another aspect, attribute-based access control (ABAC) and the use of context information in security decisions have shown to be promising for IoT. This work presents a flexible implementation architecture called Flexibility for Context-Aware Applications Security in IoT (FCAAS-IoT) that aims to provide confidentiality and privacy for context sharing in IoT environments. The security functions defined for the modules allow controlling access to information and encrypting it before sending it, based on the ABAC model and using the request context to choose algorithms and cryptographic keys, by defining and analyzing policies, conferring a potential for adaptability to various context request scenarios. The results obtained with the proof of concept demonstrated the correct functioning of the functions designed for different variations of contexts and attributes, and that the architecture performance increases the response time in providing information. Although this difference grows for larger contexts, the addition is not so high and can be reduced with adjustments in policies and in the implementation of modules according to the operating scenario, thanks to the flexibility of the architecture. Extensions of the research will be carried out to verify the functioning of FCAAS-IoT in different implementations and improve the use of request context, through the consolidation of a policy model and insertion of machine learning in the process.