MiMoQ: A System for Experimentation of Microservice-Based Applications

Microservices have emerged as an architectural style, where software architectures are broken up into loosely coupled components or services that can be developed, tested, deployed, scaled, operated, and upgraded independently. Thus many companies are re-architecting monolithic systems with microservices-based architecture to improve quality attributes (QA) such as scalability and performance. Once the refactoring is concluded, designers have to evaluate whether the expectations raised regarding quality attributes are met. This evaluation must be done in controlled environments and involves knowing the ecosystem used for deployment, as well as the available tools for monitoring and workload generation. In this paper, we present the MiMoQ tool for experimentation of microservice-based applications, which are deployed in Kubernetes. MiMoQ (Microservices Metrics of Quality) aims to facilitate the development of the experimental process, ranging from the generation of workloads to the collection and analysis of metrics, thus freeing experimenters from tasks such as the selection, installation, configuration, and integration of different types of tools for monitoring and workload generation. In this way, performance analysts can focus on the design of experiments, analysis of results, and continuous improvement of applications.