Sistema de apoio a decisão para implantar uma aplicação baseada em microsserviços em um ambiente multi-cloud

• Ano de defesa: 2019
• Autor(a) principal: Carvalho, Juliana Oliveira de
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: Seleção de múltiplos provedores; Múltiplos requisitos; Microsserviços
• Link de acesso: http://www.repositorio.ufc.br/handle/riufc/46144

Cloud computing has become a trendy model of service delivery, bringing various benefits. However, to apply the cloud model in specific scenarios, some challenges must be overcome. One of these problems is to deploy and run applications in various providers, and each one comprises several services with similar functionalities and different capabilities. Thus, dealing with issues of application distributions in multiple providers is a complex task for a software architect, since the components of an application have different characteristics. Solutions have been proposed to face this problem, but most of them focus on service providers. Therefore, we propose a cost-effective decision-making system to deploy a distributed application across multiple cloud providers. We consider in this work applications based on microservices, for offering greater flexibility. Thus, the proposed solution select providers that best meet the microservices and software architect requisites, in a manner that the microservices can be deployed in many providers. We propose three selection models for the system to serve a variety of scenarios. To reach the objectives, we also offer a definition, a classification and taxonomies for the management of resources in multiple providers from the perspective of a software architect, and a definition of microservices in the context of multi-cloud. Further, we propose PacificClouds, an architecture for managing the deployment and execution of applications based on microservices distributed multi-cloud. In the end, we accomplished a comparative analysis of the three proposed models, which one shows the feasibility concerning the performance of the solutions applied in each of the models.