When you want to design a database, you have to choose one of the many different data models. NoSQL ones are great for various atypical data problems. In this topic, you will learn about a Network Database. Also, we will study the problems it solves, its advantages and disadvantages.
How is a network database organized?
As you can assume from the name, the network data model is based on a network or, in another way, on an oriented graph. The nodes of the graph contain the main information, the relations are implemented by the connections between the vertices. Relationships are carried out using pointers or links. All nodes are arranged in a clear hierarchy, located by layers pre-defined by the developer. There is a member-owner concept in such a graph, that is very similar to the parent-child tree concept. It is easier to imagine such a graph as a tree with members (children) having many parents.
This is well illustrated with the picture below:
So, what's the purpose?
The Network data model is very similar to the Hierarchical one – a clear hierarchy, data structure, and so on. Actually, the network database model was a progression from the Hierarchical database and was designed to solve some of that model's problems, specifically the impossibility of creating a many-to-many connection.
Other advantages of the model include:
- Efficient access to data because of the clear structure
- Presence of good single standardization
- Simplicity and nativity of the model
- Flexiblity when designing the database
However, the model is not without its disadvantages:
- It's absolutely not flexible for the users – they are limited by the connections between nodes that the architect provided at the development stage.
- It is difficult to change and add information – sometimes it is necessary to change the entire structure to enter a new connection between nodes.
- Integrity has to be maintained manually.
- There exist query restrictions, i.e. they are executed only from top to bottom.
Thus, the model is good as an improved hierarchical one. It is suitable if your data has a clear structure that will not change in the future. Or if you need high-speed access to information and you do not plan to add a lot of new information or manipulate it often. For example, for modeling the biological cellular structure – the data has a clear scheme and is unlikely to change much.
Practical application
The model is not very actively used nowadays. But it is the optimal solution in certain scientific and applied fields.
The subway network of any major city is probably best represented by a network data model. In such a network, all stops on the system form the nodes of the network, and a link is a connection between two stops. The structure and data change quite rarely.
Biochemical processes can be modeled as a network to implement reactions in the living organism. For example, the process of DNA or RNA replication, as it has a clear algorithm and structure of impact.
Databases supporting the network model were widely used on IBM 360/370 series computing systems. For instance, IDMS is one of that databases, released in 1973. Despite some interesting features and advantages, only a small number of network databases have survived to this day, for example, the Raima (formerly dbVista). Raima is now used for transactional applications as a lightweight cross-platform solution.
Conclusion
Nowadays, Network Databases are not widely popular. However, they still are an effective solution for modeling structured information and they are free from some disadvantages of their prototype – a Hierarchical data model.