Conceptual Modelling of Service-Oriented Software Systems

Abstract

Conceptual modelling has originated from the areas of software engineering, databases and knowledge representation, and Heinrich C. Mayr, to whom this article is dedicated, has been involved in this area from the very beginnings. While in these areas a high degree of maturity has been achieved, conceptual modelling still lacks this maturity in other areas such as service-oriented systems despite the demand from novel application areas such as cloud computing. In this article we discuss the axiomatic BDCM2 framework capturing behaviour, description, contracting, monitoring and mediation. We argue that the framework gives an abstract answer to the ontological question what service-oriented systems are. On these grounds we address the intrinsically connected modelling question how to capture cloud-enabled service-oriented systems. We outline a conceptual modelling approach that is grounded in a distributed middleware coordinating the client access to multiple clouds through a concept of mediator. For this we exploit abstract machines with interconnected layers for normal operation, monitoring and adaptation. We illustrate the model by the use case of a robotic care system showing that the general model can be fruitfully exploited for failure alerts, failure anticipation and prevention, and safety hazards detection, which links the research to recent interests of Heinrich in conceptual modelling for ambient assistance systems.