DISTRIBUTED DATABASE SYSTEMS - Database Design and Development: An Essential Guide for IT Professionals - page 550

Distributed: homogeneous Distributed: heterogeneous Centralized

DB DB

DB

DB

DBMS

DB

DBMS

DBMS

DB

DB

DBMS

DB DB

DB

DBMS

DB

DBMS DBMS

DBMS

Many computer systems
One type of DBMS
One conceptual schema

Many computer systems
Many types of DBMSs
Many dissimilar databases

One computer system
One DBMS
One database

Figure 18-4 Two types of distributed database systems.

A configuration may be heterogeneous in several ways. The global schema must
reconcile the variations. Heterogeneity can be manifested as follows:

Different data models. You may have relational databases at a few sites, and some
hierarchical or network databases at other sites. In organizations with old legacy
systems still in full operation, this configuration may be a practical method for
implementing distributed database systems. However, you will have to deal with
variations in data representations.

Different constraints. Different models implement data integrity constraints in dif-
ferent ways. For example, the referential integrity constraint in relational models
governs relationships between entities in the relational model.

Different naming conventions. Each database may name the same data item in dif-
ferent ways, assigning different data types.

Different data content. The same data element may have different attributes at dif-
ferent sites.

Different data languages. Each site may have its own data definition and data
manipulation languages.

DDBMS

The DDBMS constitutes the software that manages the distributed database as a
logical whole. Although parts of the data are stored at different sites and managed
as local data through the local database management system, DDBMS envelops all
of the parts of data at different sites and provides global data access.