Um framework para suportar de forma semiautomática a atividade de desenvolvimento de software para mapreduce utilizando MDE

Detalhes bibliográficos
Ano de defesa: 2017
Autor(a) principal: SOUSA JUNIOR, Osvaldo Silva de lattes
Orientador(a): LOPES, Denivaldo Cícero Pavão lattes
Banca de defesa: LOPES, Denivaldo Cícero Pavão lattes, SILVA, Aristófanes Corrêa lattes, DEL FABRO, Marcos Didonet lattes, SANTOS NETO, Pedro de Alcântara dos lattes, GUTIÉRREZ, María del Rosario Girardi lattes
Tipo de documento: Tese
Tipo de acesso: Acesso aberto
Idioma: por
Instituição de defesa: Universidade Federal do Maranhão
Programa de Pós-Graduação: PROGRAMA DE PÓS-GRADUAÇÃO EM ENGENHARIA DE ELETRICIDADE/CCET
Departamento: DEPARTAMENTO DE ENGENHARIA DA ELETRICIDADE/CCET
País: Brasil
Palavras-chave em Português:
Palavras-chave em Inglês:
Área do conhecimento CNPq:
Link de acesso: https://tedebc.ufma.br/jspui/handle/tede/2048
Resumo: The need to analyze a large volume and variety of data to extract information has been increasing investments in Big Data. One example would be investments targeted at software engineering for Big Data platforms. These investments are recent and emerging, so several challenges and opportunities are found in the literature, but few approaches have been proposed to support them. In this work, a framework based on Model-Driven Engineering (MDE) and Weaving is proposed to support the software development activity in a semiautomatic way, using the MapReduce model of the Big Data platform. This framework was called F2BD and uses MDE to assist in controlling the complexity of software development through models; and uses Weaving to unify the view between different models. An activity process is proposed to guide the use of F2BD. In addition, a metamodel based on Action Language for Foundational UML (Alf) and a graphical notation called VisualAlf are proposed to complement UML, aiming to support the description of the actions modeled in the bodies (i.e. body field) of methods of diagram class UML. Metamodels for Platform-Description Model (PDM) based on MapReduce and metamodels for abstract Platform-Specific Model (PSM) based on Spark are provided. Transformation definitions of models written in Atlas Transformation Language (ATL) are proposed. The applicability of F2BD was demonstrated through the construction of a tool (TF2BD) and the feasibility of TF2BD was demonstrated through the construction of two illustrative examples and an experimental evaluation. TF2BD supports the tasks involved in software development activity, providing editors for manual manipulation of models and transformation definitions for automatic generation of PSM as well as full source code. This is possible because TF2BD was built based on the F2BD architecture. Thus, it is concluded that F2BD is feasible and can be used for the construction of other tools.