The adoption of architectural patterns has recently been investigated in connection to their impact on some quality attributes, such as the system performance. Since the benefit of choosing an architectural pattern is not guaranteed, it becomes of key relevance to collect data about the parameters that contribute to the effective use of patterns.

This talk aims to present a data-driven approach to assess the quantitative impact of design decisions for a given pattern, thus understanding the relationships between design choices and performance requirements. The approach has been tested on a dataset including parameters related to three microservices patterns and their performance characteristics. We do apply machine learning techniques, i.e., PRIM and CART, to infer constraints on the parameter values that help to understand and reduce the performance sensitivity of pattern configurations. This is joint work with J. Andres Diaz-Pace and David Garlan.

Catia Trubiani is Associate Professor in Computer Science at the Gran Sasso Science Institute (GSSI), Italy. Before joining the GSSI, she has been with various international research institutes like the Imperial College of London in UK, and the Karlsruhe Institute of Technology in Germany. Among various national and international projects, she is currently scientific coordinator for the MUR-PRIN project DREAM (modular software Design to Reduce uncertainty in Ethics-based cyber-physicAl systeMs), under the Young Line action, and the MUR-PRO3 project on Software Quality. Her main research interests include the quantitative modelling and analysis of interacting heterogeneous distributed systems.