Explaining the performance characteristics of complex software systems is not trivial, even more so when exposed to multiple sources of uncertainty, e.g., heterogeneous workloads leading to fluctuations in resource demands. The goal of this talk is to present some methodologies on how to explain the performance analysis results while quantifying the impact of suitable architectural alternatives. One viable solution is to make use of software (anti)patterns since they are known to affect the system performance and bring quantitative information. Recent results on real-world java applications and microservices will be presented, along with challenges and open research directions.
Join at: imt.lu/seminar