Physical Database Design

Abstract

The two stages of database design are logical and physical design. Logical database design, also referred to as data modeling, is done first. It is the process of studying the workings of a business organization, constructing a set of tables to store the business data, and understanding the constraints on the data, the dependencies between the tables, and the business rules concerning the data. The logical database design process is conducted without reference to any specific database technology such as Oracle or Microsoft SQL Server, and a simple example can be found in Oracle Database 11g Java Developer’s Guide. Tools such as Oracle Designer, CA ERwin Data Modeler, Embarcadero ER/Studio, and Toad Data Modeler can be used for the purpose. Physical database design follows logical database design. First, the logical model is mapped to the proprietary features of the chosen database technology. Security requirements, integrity requirements, and business rules are also implemented. Finally, we consider performance: the ability of the database engine to handle work requests efficiently. This typically involves the creation of indexes on the data, and Oracle Database provides a wealth of indexing mechanisms to choose from.