Types of Database Models

It is possible to broadly distinguish among three generations of architectural database models. Listed in order of increasing sophistication they are the primitive, classic and semantic data
models.

The primitive data model is characterised by record-structures grouped in file structures. The main functions available are read and write operations over records. The classic data model comprises the hierarchical, network and relational models, with the hierarchical and network models being the precursors to the relational model. Systems based on the relational data model
are currently the most popular in commerce and industry. Products such as ORACLE , SQLServer  and  Access are based on this data model. The semantic data model attempts to provide a more
expressive way of representing information. This is achieved by moving away from the fundamental record orientation of the previous two systems. The currently much popularised
object-oriented model can be regarded as adhering to a semantic database model

In recent years, a lot of resources have been invested into re-engineering database management systems (DBMS) that are based on the classic data model into formats compatible with the semantic data model. Such an approach has the advantage of benefiting from many of the advancements of new technology without requiring  old  system to be discarded.

The Relational Data Model

The relational data model was largely devised by one man, E. F. Codd. Codd’s 1970 paper on database architecture entitled A relational model for large shared data banks laid the foundation
for the model and from 1968 to 1988 Codd published more than 30 papers on the subject.

In a relational database management system (RDBMS), data is stored in simple two dimensional (database) tables. Each table is usually stored as a separate file. Each file is composed of records and each record is made up of a number of fields. Each field contains one data value. The pointer structures of the network model and the hierarchy of the hierarchical model are replaced by data redundancy in the form of identification codes called keys, which are fields containing unique values that identify each record. The data redundancy is usually minimised by a process known as normalisation.

The main advantages of the relational database model, and probably the main reasons for its
widespread popularity, are its simplicity, flexibility and  logic. The relational data model has
one data structure, the table, and only eight simple operators (namely: restrict, project, join,
product, union, intersection, difference and division) bundled together as relational algebra.
Addition or removal of data is easy; it involves the addition or removal of one or more records
or a whole table. Queries across different tables are supported by joining different tables
through their common fields (i.e. relational join). However, if the relationships between tables
are complex and a number of joins are needed, operations can become slow. Hence, good
database design is important.