Proposal of a real time microservice architecture applied to business intelligence for the financial sector

Abstract

This work proposes to i) generate a microservice, which allows the capture of massive transaction information from a relational database, ii) develop the extract, transform and load process (ETL), and iii) store proper and valuable information for decision-making in a non-relational database. The entire process also uses business intelligence mechanisms to convert data into information. Three phases were used for the implementation of this process: the first one, a mechanism within the relational database for change data capture (CDC)- Then, the second one, the design and implementation of a consumer in charge of listening to the messages, generate the transformation process and load it to the non-relational data-base, ensuring fault tolerance, reliability and scalability. Finally, through defined indicators, a control panel was designed, which offers operators and managers the necessary guidelines to generate decisions and measure the achievement of objectives and goals. The entire process, implemented through a visual and intuitive tool, allows converting the data into knowledge, which generates a competitive advantage in companies’ management and operational processes. The proposed integration, results in an architecture design that efficiently manages thousands of records in real-time, offering greater speed, performance, and optimization in searches of large volumes of data for effective decision-making. © 2022, Springer Nature Switzerland AG.