Schemas Integration - Information Systems Reengineering, Integration and Normalization - page 277

7.3 Schema Integration for ORDBs

The RDB system has been dominant in the industry for the last two decades. OODB
application is recognized as a postrelational technology that can improve productiv-
ity. Hence, most companies need to enhance their existing RDB systems to support
new OO applications as and when needed. The trend of the current industry is to im-
plement an ORDB system using a relational engine with OO features. This section
proposes a methodology to integrate existing ORDB based on user requirements.
A frame metadata is used to enforce constraints for solving semantic conflicts aris-
ing from schema integration. The metadata is an object-relational metadata that
can specify static data semantics as well as dynamic data operation based on four
relational tables.
In order to have coherence between new OO database applications and the exist-
ing database systems, the leading database manufacturers gradually modify their
RDB system to support OO features. It results in the so-called object-relational
database management system (ORDBMS) in the current market. Most of these OR-
DBMS are powered by a RDB engine with extensions to OO interface and features.
When designing database using these systems, user employ either relational view
with some OO features, or use OO view under a relational core. We propose a
practitioner approach to integrate this kind of ORDBMS. A simplified schema inte-
gration technique is applied to the source database schemas, either in relational or
OO structure, based on the user requirements. The frame model metadata is used
to capture these semantic constraints and other abstractions result from the integra-
tion. The resultant system is an integrated schema of the object RDB system (Fong
et al. 2000 ).

7.3.1 Frame Model Metadata

The frame model metadata follows an OO paradigm, based on frame. All concep-
tual entities are modeled as objects and group in object types called classes. The
frame model metadata is implemented with a knowledge representation schema that
represents the taxonomy inheritance structure (i.e., abstract relationship), properties
of objects (i.e., classes and attributes), and the relationship between those objects in
a standardized form. The details can be referred to Fig. 8.2 in Chap. 8.
Schema integration provides a global view of multiple schemas. Our approach
uses a bottom-up approach to integrate an existing database into a global database
by pairs. The main objective is to provide an integrated schema based on user re-
quirements with no loss of information. The general algorithm is as follows: