Distributed quality-attribute optimization of software architectures

ALEJANDRO RAGO; SANTIAGO A. VIDAL; JORGE ANDRES DIAZ PACE; SEBASTIAN FRANK; ANDRE VAN HOORN

Fortaleza

Simposio; 11th Brazilian Symposium on Software Components, Architectures, and Reuse (SBCARS 2017); 2017

CBSoft

A key challenge of software architecture design is how to satisfyquality-attribute requirements, which often con ict with each other.This is usually a complex task, because there are several candidatesfor architectural solutions meeting the same requirements, andquality-attribute tradeo s of those solutions need to be consideredby the architects. In this context, we present the SQuAT frameworkto assist architects in the exploration of design solutions and theirtradeo s. This framework provides a modular approach for inte-grating quality-attribute analyzers and solvers, and also features adistributed search-based optimization. In this paper, we report on anexperience using SQuAT with Palladio architectural models, whichintegrates third-party tools for performance and modi ability, andshows the tradeo s among candidate solutions to the architect. Fur-thermore, we enhance the standard search schema of SQuAT witha distributed negotiation technique based on monotonic conces-sion, in order to provide better tradeo s for the architect?s decisionmaking.