A model-driven approach for engineering trust and reputation into software services

Abstract

The ever-increasing complex, dynamic and distributed nature of current systems demands model-driven techniques that allow working with abstractions and self-adaptive software in order to cope with unforeseeable changes. Models@run.time is a promising model-driven approach that supports the runtime adaptation of distributed, heterogeneous systems. Yet, frameworks that accommodate this paradigm have limited support to address security concerns, hindering their usage in real scenarios. We address this challenge by enhancing models@run.time with the notions of trust and reputation. Trust improves decision-making processes under risk and uncertainty and constitutes a distributed and flexible mechanism that does not entail heavyweight administration. This paper presents a trust and reputation framework that is integrated into a distributed component-model that implements the models@run.time paradigm, thus allowing the system to include trust in their reasoning process. The framework is illustrated in a chat application by implementing several state-of-the-art trust and reputation models. We show that the framework entails negligible computational overhead and that it requires a minimal amount of work for developers.