Synopsis: Second part of
Kaled Kandil’s discussion on databases, focusing on implementations that form the core of many modern installations.
Network
An adaptation on the hierarchical model, network databases still use trees to nest information, but data points can connect across to other branches of the tree. This is accomplished through two complementary principles: records and sets. Records are individual pieces of data, whereas sets describe relationships between different records. Any record may therefore serve as a node in any other tree structure. Usually, network databases are parsed by maintaining a position in the data, and traversing from one record to another according to the properties of the set.
Relational
Introduced by E. F. Codd in 1970, relational databases sort data into a series of tables, with relationships between the tables. Each table resembles the tables in a flat database, with rows and columns of information related in the same way each time.
The sequence of rows and columns in each table is inconsequential, but the type of data in each column is always the same. In addition, each field or record must contain only one piece of data. Tables may include a column of keys that connect information across other tables for quick access. Alternately, fields may be given the same name across multiple tables in order to allow queries to be made based on these values.
Object-Oriented
The object-oriented model for databases arose from the need to find more efficient ways of programming interfaces with complex relational databases. In previous implementations, the data from the relational database would need to be extracted and “flattened” into a two-dimensional, linear relationship before being accessed by the computer program.
By extending object-oriented programming functionality into the relational database framework, information can be accessed and employed much more efficiently, as no translation is required. This improves the performance of the system considerably.
ec2
