Processo de ETC orientado a serviços para um ambiente de gestão de PDS baseado em métricas

• Ano de defesa: 2007
• Autor(a) principal: Silveira, Patrícia Souza
• Orientador(a): Ruiz, Duncan Dubugras Alcoba
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Pontifícia Universidade Católica do Rio Grande do Sul; Porto Alegre
• Palavras-chave em Português: INFORMÁTICA; ENGENHARIA DE SOFTWARE; SOFTWARE - TÉCNICAS DE AVALIAÇÃO; QUALIDADE EM INFORMÁTICA; DATA WAREHOUSE
• Link de acesso: http://hdl.handle.net/10923/1616

The search for quality is a constant value in corporate environments. With this aim, software development organizations utilize metrics to measure quality of their products, processes and services. These metrics should be collected, consolidated and stored in a single central repository typically implemented as a Data Warehouse (DW). The definition of extraction, transformation and loading (ETL) of metrics that will be stored in DW, considering the software development environment (heterogeneity of sources, of process models, of project classes and of level of isolation) is no trivial task. This paper presents a data warehousing environment called SPDW+ as a solution to the automation of the ETL metrics process. This solution introduces a comprehensive and streamlined analytical model for the analysis and monitoring of metrics, and is built on a service-oriented approach that utilizes the Web Services technology (WS). Moreover, SPDW+ addresses the low-intrusion incremental load and the high frequency and low latency present in metrics collection. The main components of SPDW+ are specified, implemented and tested. The advantages of SPDW+ are: (i) flexibility and adaptation to meet the requirements of the constant changes in business environments; (ii) support to monitoring, which allows the run of frequent and incremental loads; (iii) the capacity to make less burdensome the complex, time-consuming task of capturing metrics; (iv) freedom of choice regarding management models and the support tools used in projects; and (v) cohesion and consistency of the information contained in the metrics repository needed to compare the data of different projects.