Oleg Garipov

Sergey Zykov

Lately in the industry of an information technology the question on integration of heterogeneous sources of the data is even more often brought up. In difference to homogeneous systems, heterogeneous data source systems can be described by various data models. Integration of heterogeneous data source means creation of the general scheme of the description of the given all subsystems for the unified access to the data, with preservation of semantics of this data. Our article is attempt of consideration of the various models used to describe the general scheme of databases.

Ideas of conceptual modeling are not a new phenomenon in the modern world. Sources of researches in the given direction have arisen in 70-s’ years 20ого centuries and were connected with software development. Concepts were used to describe the data abstractions with respect to idea of encapsulation data from users.

In our work we analyze few most popular approaches which can used to represent heterogeneous data. Our special attention was given to ER model, ORM model, RM/T model, functional approach, semantic networks and ontology. We identified benefits and weaknesses of these models. We propose few statements which “ideal” model must be satisfied. We argue that existing models are not satisfied to all of these principles.

It is necessary to develop new models which would be a composition of existing approaches. This approach must be founded on few statements: based on types of object; able to tracking dynamic of the universe of discus; software implementation independence; knowledge models compatible; semantic interoperability; user-friendly notation; integrity, consistency, redundancy data verification mechanisms; reused components, support different data viewpoint.

Download presentation