Self-Adaptive Systems Requirements Modelling: four related approaches comparison

Abstract

When developing Self Adaptive Systems (SAS), their highly adaptiveness has to be taken into account as early as the requirements elicitation. Because such systems modify their behaviour at run-time in response to changing environmental conditions, Non Functional Requirements (NFR's) play an impor-tant role. One has to identify as early as possible the requirements that are adaptable. Because of the inherent uncertainty in these systems, goal based approaches can help in the development of their requirements. In order to cope with this purpose, we have defined a combined approach based on several requirements modelling techniques. In this paper we use a common case study and well defined comparison criteria to illustrate the way those techniques can benefit from each other. This submission is a synthesis and hence make some reference of more specific requirements models submissions. I. INTRODUCTION This paper aims at studying the differences and potential combination of 4 different requirements' modelling tech-niques: KAOS [8], a goal-based approach (detailed in sec-tion II-B), SYSML [5], a general purpose system modelling notation (detailed in section II-C), SYSML/KAOS a combi-nation of KAOS and SYSML (detailed in section II-D) and RELAX [6] a dedicated language for adaptive systems. We have taken the opportunity of the CMA@RE workshop in order to compare those techniques that were chosen to provide the best approach to deal with the specificities of Self-Adaptive Systems requirements. In section II we describe the context in which this study has been conducted, as well as the minimum of background needed to understand the models; in section III we provide the different requirements models we have developed for the bCMS case study; in section IV we provide our inputs to the defined comparison criteria; and we analyse the results of the parallel efforts; and finally in section V we conclude this study.