Database design is the process of moving from real-world business models and requirements to a database model that meets these requirements.
Normalization in a relational database, is an approach to structuring information in order to avoid redundancy and inconsistency and to promote efficient maintenance, storage, and updating. Several “rules” or levels of normalization are accepted, each a refinement of the preceding one.
Of these, three forms are commonly used: first normal, second normal, and third normal. First normal forms, the least structured, are groups of records in which each field (column) contains unique and nonrepeating information. Second and third normal forms break down first normal forms, separating them into different tables by defining successively finer interrelationships between fields.
For relational databases such as Adaptive Server, the standard design creates tables in Third Normal Form.
When you translate an Entity-Relationship model in Third Normal Form (3NF) to a relational model:
Relations become tables.
Attributes become columns.
Relationships become data references (primary and foreign key references).