Databases have a wide range of applications in business systems. They help automate repetitive document customization tasks, they enable telephone operators to get rapid access to customer details and they speed up the consolidation of financial information. Databases categorize types of data and then establish relationships between those categories, the three main management systems for databases are relational, hierarchical and network. Each has its own merits. There are advantages and disadvantages to the network database model in comparison to the other two data management systems.
Information is grouped into entities or records and each entity has attributes, which correspond to column headings. For example, the entity "Customer" would have Business Name and Telephone Number as two attributes. Every entity has one defining attribute that uniquely identifies each record in the table. This is called the key, which is an index. Tables are joined together by these key attributes to expand the range of data available to each query.
The relational database management system dominates business application. Data is "normalized" and then stored in tables. Normalized data fits into a structure that eradicates repetition and redundancy of data. It is the joins between the tables of normalized data that creates a relation between the attributes and thus gives this database management system its name. Hierarchical databases are less common. They link entities together, again by key attributes, but organized as parent-child relationships. This creates a tree structure of data and is suitable for geographical information, or archived data used for "data mining."
A network database is similar to a hierarchical database. However, whereas a hierarchical database has only one-to-many relationships between entities, a network database management system allows many-to-many relationships. This is the defining characteristic of network DBMSs. In a hierarchical database, one parent entity may have many children entities, but each child may have only one parent. This is the one-to-many relationship. The many-to-many relationship in a network DBMS allows one parent entity to have many children entities, and one child entity to have many parent entities.
The main benefit of the network DBMS model is the many-to-many relationship. The hierarchical database does not account from many human organizational events that require an entity to have many upward links to other entities. For example, in a medical database, a doctor may be assigned to many wards and take care of many patients, while the patient is one ward that does not link through to the relationship to the doctor entity and so the patient entity also needs a link to both the doctor entity and the ward entity. So a series of cross-relationships quickly develop.
Relational databases have structures that resolve many-to-many relationships between instances, or records, of an entity, but network databases do not. For example, in the network database example, the patient may have many doctors, and a doctor has many patients, but the network DBMS could only establish a one doctor to many patient relationship or one patient to many doctor relationships, not both.