Configuring the Variability of Business Process Models Using Non-Functional Requirements Emanuel Santos 12 , Jo˜ao Pimentel 1 , Jaelson Castro 1 , Juan S´ anchez 2 , and Oscar Pastor 2 1 Universidade Federal de Pernambuco, Centro de Informatica, Cidade Universitaria. S/N, 50741-000 Recife, Brazil {ebs,jhcp,jbc}@cin.ufpe.br 2 Universidad Politecnica de Valencia, Camino de Vera, S/N, 46022. Valencia, Spain {jsanchez,opastor}@dsic.upv.es Abstract. The existence of variations in the organizational environment makes the configuration of business process models a complex activity, even for experienced business analysts. The increasing adoption of busi- ness processes models by software engineers as a input for requirements analysis strengthens the importance of adressing this issue. The challenge is to configure business processes to fit the organization better. We pro- pose an approach that combines variability analysis and non-functional requirements to drive the configuration of a business process. Applying this approach we can analyze variability in the model in order to as- sess the impact of the choices on the process quality constraints - the non-functional requirements. Moreover, it provides a rationale for the selection of a specific configuration. Key words: Business Process Models, Business Process Configuration, Variability, Non-Functional Requirements 1 Introduction With the increasing interest of the software engineering community in using business process models as a source of requirements, raised the importance of representing variability on these models. Variability, on business process models, consists of defining alternative paths of execution in a workflow [1]. In this way, the process can be personalized for a specific context, e.g., for a foreign subsidiary of a corporation. There are several approaches for representing variability in a business process model, like Schnieders and Puhlmann [2], Montero et al. [3] and De la Rosa et al. [4]. However, the problem of choosing the most suitable alternative - the so-called process configuration - is not solved yet. In the industry, the configuration still is performed in an ad hoc basis, guided only by the analyst’s experience. Some techniques have been proposed in academia, like the usage of questionnaires [4] and domain analysis [5], but these techniques are more concerned with the elicitation of variability than with the configuration itself.