A synthetic approach to building a canonical model of subject areas in the integration bus

This paper is dedicated to the implementation considerations of a canonical model of subject areas in the integration bus and to the definition of data mapping corresponding to this model. The proposed approach to transforming data, when transferring them between individual applications or services of the system, is to convert input messages into output messages using an intermediate canonical representation of the data via rules that map the source model to the canonical one, and the canonical model to the target. Since the canonical model of a subject area serves a technical important task and is not intended for human use, it may be generated automatically instead of manually and designed `from scratch', as a `union' of models used in the various application parts of the system. Subject area models of the application parts being integrated may be written using different formalisms, and yet another formalism may be used for the canonical model, so that a mechanism is required to automatically capture various concepts expressed in various formal systems. In the present paper we focus on developing such a mechanism, based on the automatic generation of a (somewhat simplified) representation of the most important kinds of entities in the canonical model.