A computer database is arranged by tables, which generally refer to a listing of records that share some commonality. In a standard hierarchical database model, tables are associated with one another based on a parent and child relationship, where one parent table can have multiple children, but a child table cannot have multiple parents. For example, the "Employees" table — the parent table — might be further subdivided into two child tables: "Current Workers" and "Past Workers." A network model database offers an alternative to this parent and child limitation. In a network model database, any single table can have both multiple child tables and multiple parent tables, resulting in a more natural graphical structure between the entities in the database.
The primary advantage of a network model database is that it allows for a more realistic portrayal of the relationship between tables. Trying to squeeze table arrangements into a one-parent model can be restrictive, especially when two or more tables could equally be considered parent categories. Continuing the past example, while the "Employees" table was used as the parent for both the "Current Workers" and "Past Workers" tables, an "Independent Contractors" table would serve equally well as the parent for either, or even both. Through the network model setup, the end-user can structure the database so that both "Employees" and "Independent Contractors" are parents of the "Current Workers" and "Past Workers" tables.
Tables in a network model database can be represented graphically, by listing each table name individually, starting with the highest-level parent tables at the top of the graph and working down to the most dependent child tables at the bottom. Each parent-child relationship can be represented by drawing a line between the related tables. This allows database designers to more readily understand the connection between entities.
Although the network model allows for a more realistic representation of the relationships between data entities, it is largely obsolete in modern database design. Escalating hardware performance for computers allowed larger organizations to evolve to what is called a relational database model, which allows the end-users to directly input data using key values instead of focusing strictly on the parent-child relationship. For example, a list of employees in a relational database model might have the employee's name listed as the key value. That links the employee list to all other tables containing the employee's name, allowing all information about that employee to be pulled up at once, regardless of what table holds it.