Visualização de dados de monitoramento de requisitos de qualidade contextualizados coma arquitetura de software

Abstract

The problem of the large volume of data is present in the software industry, especially in the software monitoring log data, which is generated in real time, as the software is used. Some factors make it difficult to manually analyze log data. One is the large volume of data generated; another factor that makes it difficult to manually analyze the logs and the fact that they are usually scattered in one or several log files, presenting itself in a disorderly way. For this reason, a manual log analysis has been shown as an error-prone activity. Therefore, it is necessary to use some computational tool to support log analysis. In addition, in order to have a good interpretation of the analyzed data, it is necessary to contextualize them with abstract software models, such as software architecture. This work aims to develop a visualization approach to facilitate the data analysis process of monitoring the requirements of quality, efficiency and reliability, providing a greater contextualization with abstractions of the software architecture and a tool, called VisRQ, supporting this approach. This tool was evaluated by a semi-structured experiment, in which the users analyzed the monitoring logs and answered questionnaires in order to verify the fulfillment of the desired requirements. We also evaluated the tool learning time and whether it makes the log analysis process more accurate and faster. The experiments participants were students of the Informatics technical course of the Federal Institute of Alagoas - Campus Arapiraca and of the course of Computer Science of the Federal University of Alagoas - Campus Arapiraca, besides some professionals that work with Information Technology. The tool acceptance was positive, according to the responses of the evaluative experiment, 100% of the participants stated that they prefer to use the VisRQ tool instead of the previous process that is less automated and 97.6% considered that it facilitates the analysis process. The experiment shows that the log analysis process using VisRQ tool is more efficient, besides allowing the visualization of information about the nonfunctional requirements in a contextualized way with diagrams of the software architecture.