An experimentation framework for validating architectural properties as proxies for the ilities

Abstract

Desired system ilities, understood as lifecycle properties, generally drive the value of a system’s architecture. However, ilities are hard to predict during system architecture development before the architecture has been instantiated with a system design. In this paper, an experimentation framework that can be employed to validate the adequacy of using architectural properties as proxies for ilities is presented. The experimentation framework builds upon a theoretical, generic conceptualization of ilities, which helps to formally justify the difficulties of and limitations in predicting how architectural choices influence system ilities. Application of the experimental framework is showcased with a notional scenario, where the framework is used to evaluate how different system topologies influence the integrability of a system.