Design Layer Modeling

To support the application development process, you can create separate logical and physical data models. Typically, a design layer consists of a logical or a physical data model designed for a specific purpose. For example, the first design layer may be a logical model, which captures business requirements. In a second design layer, a physical model, which represents the database implementation, may be derived from the logical model. In each design layer, design decisions may be implemented that are relevant to the purpose of the data model. Often, in a logical data model, design decisions are made in accordance with the rules of normalization, which attempt to minimize the redundancy of data. In a physical data model, a design decision may be implemented to denormalize a model in favor of improved performance. Usually, when you work with design layers it is important to link related models, keep them synchronized, and maintain a historical record of the design decisions in each layer. Design layer methodology is supported in many ways including wizard-based tools for deriving models, transformations, and historical tracking of design decisions at the model and object levels.

The choice of models is as follows:

The Logical Model: You can create an independent logical model, with unique notation and display options.
The Physical Model: You can create an independent physical model with unique notation and display options.
The Logical/Physical Model: You can create a new model with a combined logical and physical components.