3.3. Network systems

A typical representative is the Integrated Database Management System (IDMS) of Cullinet Software, Inc., designed for use on IBM main-class machines running most operating systems. The system architecture is based on the Data Base Task Group (DBTG) proposals of the Programming Languages ​​Committee of the Conference on Data Systems Languages ​​(CODASYL), the organization responsible for defining the Kobol programming language. The DBTG report was published in 1971, and in the 1970s several systems appeared, including IDMS.

3.3.1. Network data structures

The network approach to data organization is an expansion of the hierarchical one. In hierarchical structures, the child record must have exactly one ancestor; a descendant can have any number of ancestors in a network data structure.

A network database consists of a set of records and a set of links between these records, and more precisely, from a set of instances of each type from a set of record types specified in the database schema and a set of instances of each type from a given set of connection types.

The type of relationship is determined for two types of record: ancestor and descendant. A link type instance consists of one instance of the parent record type and an ordered set of child record type instances. For this type of connection L with the ancestor's record type P and the child record type C, the following two conditions must be met:

• Each instance of type P is an ancestor in only one instance of L;
• Each instance of C is a descendant of not more than one instance of L.

There are no special restrictions on the formation of communication types; the following situations are possible, for example:

1. The type of child record in one type of L1 link can be the type of ancestor record in another type of link L2 (as in a hierarchy).
2. This type of record P can be the type of ancestor record in any number of types of communication.
3. This type of record P can be a type of child record in any number of types of communication.
4. There may be any number of link types with the same ancestor record type and the same descendant record type; and if L1 and L2 are two types of communication with the same type of ancestor P record and the same type of descendant C record, then the rules on which the relationship is formed may differ in different links.
5. Record types X and Y can be an ancestor and descendant in one connection and a descendant and ancestor in another.
6. Ancestor and descendant can be of the same record type.

A simple example of a network DB scheme:

3.3.2. Data manipulation

An approximate set of operations may be as follows:

• Find a specific record in the set of records of the same type (engineer Sidorov);
• Move from ancestor to the first descendant by some connection (to the first employee of department 310);
• Go to the next descendant in some connection (from Sidorov to Ivanov);
• Move from descendant to ancestor by some connection (find the department of Sidorov);
• Create a new entry;
• Destroy the record;
• Modify record;
• Include in communication;
• Exclude from communication;
• Rearranged to another connection, etc.

3.3.3. Integrity constraints

In principle, their maintenance is not required, but sometimes they require integrity by reference (as in a hierarchical model).