Who is AIS? News about this site Reviews of CASE products Reading - White Papers CASE technology training  & consulting CASE technology training  & consulting
Who is AIS? Users' Groups & Feedback Opinions - We have strong ones! Links to Other Pages of Interest Training & Consulting
AIS CASE Home Page Site Table of Contents Private Pages E-mail to AIS  

Conceptual ~ Logical ~

03 Dec 1997

Physical Models


  The terms "conceptual". "logical", and "physical" are frequently used in data modeling to differentiate levels of abstraction versus detail in the model. Although there is no general agreement, let alone accepted authority, which defines these terms, nevertheless data modelers generally understand the approximate scope of each.

ceptual E-R Model

A conceptual entity-relationship model shows how the business world sees information. It suppresses non-critical details in order to emphasize business rules and user objects. It typically includes only significant entities which have business meaning, along with their relationships. Many-to-many relationships are acceptable to represent entity associations.

A conceptual model might discover that there is a need to house information about each person in an organization. While considerable thought is given to discovering and describing the relevant properties of each person, the designers accept implicitly that each person is distinct and unique.

A conceptual model may include a few significant attributes to augment the definition and visualization of entities. No effort need be made to inventory the full attribute population of such a model. A conceptual model may have some identifying concepts or candidate keys noted but it explicitly does not include a complete scheme of identity, since identifiers are logical choices made from a deeper context.


Logical E-R Model

A logical entity-relationship model is provable in the mathematics of data science. Given the current predominance of relational databases, logical models generally conform to relational theory. Thus a logical model contains only fully normalized entities. Some of these may represent logical domains rather than potential physical tables.

For a logical data model to be normalized, it must include the full population of attributes to be implemented and those attributes must be defined in terms of their domains or logical data types (e.g., character, number, date, picture, etc.).

A logical data model requires a complete scheme of identifiers or candidate keys for unique identification of each occurrence in every entity. Since there are choices of identifiers for many entities, the logical model indicates the current selection of identity. Propagation of identifiers as foreign keys may be explicit or implied.

Since relational storage cannot support many-to-many concepts, a logical data model resolves all many-to-many relationships into associative entities which may acquire independent identifiers and possibly other attributes as well.


Physical Database Schema

A physical data model is a single logical model instantiated in a specific database management product (e.g., Sybase, Oracle, Informix, etc.) in a specific installation. The physical data model specifies implementation details which may be features of a particular product or version, as well as configuration choices for that database instance. These include index construction, alternate key declarations, modes of referential integrity (declarative or procedural), constraints, views, and physical storage objects such as tablespaces.

In Summary

The conceptual model is concerned with the real world view and understanding of data; the logical model is a generalized formal structure in the rules of information science; the physical model specifies how this will be executed in a particular DBMS instance.

Various data modeling methodologies and products provide these layers of abstraction in different ways. Some address only the physical implementation; some model only the logical structure; others may provide elements of all three but not necessarily in three separate views. In each case it helps the data modeler to understand the level of abstraction to which a particular feature or task belongs.

Copyright 1997 Applied Information Science