Browsing by Author "Seifermann, Valentin"

Now showing 1 - 2 of 2

Open Access
Application performance monitoring in microservice-based systems
(2017) Seifermann, Valentin
Nowadays, cloud computing including its functionality and service delivery model, turned into a common paradigm to provide on-demand IT resources and scalability. This also leads to a change of architecture from monolithic systems to a microservice-based architecture with high scalability and elasticity as well as independent developments and deployments. Due to this change, IT environments are getting more complex and highly distributed. Thus, the requirements of Application Performance Management (APM) for these types of systems are also changing. While monitoring monolithic applications and infrastructures focuses on a few single components with their callstacks and the surveillance of the general state of health, the monitoring of microservice-based systems requires other approaches. This work will elaborate the actual state-of-the-art of APM in microservice-based systems on the basis of an industrial case study. Furthermore, the challenges of monitoring microservice-based systems will be elaborated. As part of the industrial case study, an existing microservice monitoring tool will be evaluated on different environments and integrated into the state-of-the-art business-oriented monitoring strategy "System Management: Inform-Locate-Escalate" (SMILE), which consists of various monitoring and service management tools. The evaluation will be done by conducting use cases, defined in the context of the thesis to meet the challenges. In addition, the work proposes an experimental concept for APM in microservice-based systems. This concept consists of a selected monitoring stack with different open-source tools and an existing microservice monitoring solution. Furthermore, it contains different dashboards with decisive metrics and other monitoring-specific data.
Open Access
How to strangle systematically : an approach and case study for the continuous evolution of monoliths to microservices
(2021) Seifermann, Valentin
In today’s enterprise IT, a growing number of applications are being developed based on microservice architectures to meet enterprise requirements for highly scaled, highly available and resilient systems running in cloud environments. However, monolithic architectures have been the traditional way of developing applications for many years, which results in IT environments still consisting of applications based on these types of architectures. To overcome the limitations of these monolithic architectures, microservice migrations are performed with the goal of gradually decomposing the monolith into independent services at granular levels. In order to achieve this process, different decomposition approaches are presented in various publications, often focusing only on the technical aspects of a microservice migration. However, during this long-term process, various factors occur on different dimensions that are not only technical in nature. These non-functional factors are often not considered in existing publications. This leaves software engineers in need of a systematic migration guide on how to proceed with a continuous migration that also takes into account the decision-making processes based on the factors that influence the migration process. To address this lack on methodological guidance, this work provides a systematic approach for a continuous and gradual evolution of monolith to microservices based on the strangler pattern. The approach is based on interviews with software engineers and project managers who have participated in microservice migrations. Based on these experiences, the influencing factors that occur during the various steps of a migration are identified. Moreover, the decisions made based on these factors are considered in order to provide a decision-making guidance.