Model-driven semantic integration of service-oriented applications

The integration of enterprise applications is an extremely complex problem since the most applications have not been designed to work with other applications. That is, they have different information models, do not share common state, and do not consult each other when updating their states. Unfortunately, domain experts, who have the required knowledge to resolve the mismatches between the applications, are typically not IT experts. Their expertise is only used in the very early stages of integration projects, namely, in the requirements elicitation stage. This makes the gap between business integration requirements and the software implementation of the integration solution wide, which, in turn, additionally complicates the integration process and often leads to the realization of solutions that work incorrectly. \ud \ud The work presented in this thesis makes contributions in the area of Enterprise Application Integration. We propose a method for the semantic integration of service-oriented applications. The key feature of the method is that semantically-rich service models, at different levels of abstraction, are employed to build flexible integration solutions. The method allows system integrators (both domain experts and software developers) to address only a limited set of concerns in a series of design steps. In addition, the proposed method supports handling of changes in the implementation technology and integration requirements. Finally, integration solutions produced by the method can be formally verified for correctness using automatic reasoners