Master Thesis MSTR-2021-114

BibliographyStieß, Sarah Sophie: Tracing the impact of SLO violations on business processes across a microservice architecture with the saga pattern.
University of Stuttgart, Faculty of Computer Science, Electrical Engineering, and Information Technology, Master Thesis No. 114 (2021).
64 pages, english.
Abstract

Context. Applications in the microservice architecture style consist of many individual services. SLOs describe the quality at which they provide a functionality, as an example their responsetime or availability. Patterns exist to handle recurring problems better. Among these is the saga pattern [Ric18], which deals with transactions distributed across multiple services. Problem. SLO violations may propagate across the architecture and cause unintended behaviour in the business process. Patterns may hide the cause of a business process’ behaviour. As for the saga pattern, an SLO violation may trigger the rollback of a transaction. The process owner notices the rollback. But a rollback is an acceptable behaviour for a transaction, such that they either do not not question the rollback’s origin at all or they do question it but cannot find any fault in it. Thus, the connection to the SLO violation remains unidentified. Objective. This thesis’ objective is to expose the impact of SLO violations on a business process in the presence of the saga pattern. Method. The means to achieve this objective is a notification that informs a user about the SLO violation and its possible impacts on the business process and a modelling language for models that capture all knowledge required to create such notifications. The user received the notification as an issue. An expert survey evaluates the language and the concept and a experiment assures that the modelling language fulfils its purpose. The experiment employs the T2-Project as a reference architecture. Result. This thesis’ results include the aforementioned modelling language, the concept for the notification, and a prototype to calculate impacts. The language connects models of architectures with those of business processes while reusing existing models. The prototype uses models according to the designed language and calculate the impact of SLO violations upon receiving notice about a violation. It also creates issues for these impacts. There is an expert survey and an experiment that attempt to validate this thesis content. Conclusion. The designed modelling language is of use when connecting existing models of architectures and business processes while also representing an instance of the saga pattern. Models according to the language are useable to calculate the impact of SLO violations on the business process. Notifications about impacts that are not apparent in the business process are helpful.

Full text and
other links
Volltext
Department(s)University of Stuttgart, Institute of Software Technology, Software Quality and Architecture
Superviser(s)Becker, Prof. Steffen; Breitenbücher, Dr. Uwe; Speth, Sandro
Entry dateJuly 2, 2024
   Publ. Computer Science