Semantic Trading: Tackling Interoperability Problems During System Integration

The development of new enterprise systems is a process of dynamically selecting the appropriate components, and composing them into a new configuration. During this process severe interoperability problems arise since different systems were developed using different languages and protocols, on different platforms, under different models. A community of traders with the support of a composition service could tackle interoperability problems by selecting "compatible" components and plugging them together. Currently, a trader with the support of an object request broker (ORB) could solve interoperability problems by locating all the compatible components (compatibility is translated into type conformance, which guarantees through interface matching that no unexpected input or output requirements or operation termination conditions will emerge) and allowing the ORB to plug them together by forwarding messages (method calls). But, limiting the matching process to the interface of the component means that information, such as message ordering or behaviour semantics, cannot be captured. So, although current trading is sufficient for solving interoperability problems at the signature level is unable to tackle problems at the protocol or the semantic level. What is needed is a new kind of semantically rich trading, which we call semantic trading, that shifts focus from the syntax (interface) to the behaviour (semantics) of the component.