Database classification: options, data models and main characteristics

The database is the set of data that needs to be streamlined, and the database management system (DBMS) is responsible for its administration, thus determining the structure, order, access rights and dependencies. For this purpose, we use our own compiler and a suitable model that defines the architecture of the database system. On the basis of architecture, a database is classified.

History of creation

Databases (DB) are a logically structured system for electronic administration, which is performed using a database management system (DBMS), adding it to the repository. Most databases can be opened, edited and consulted only using specific applications. By these principles, the classification of the database. In the 1960s, the concept of an electronic information base began to be developed as a separate layer of software between the OS and the application program.

The idea of ​​an electronic database system has become one of the most relevant innovations in computer development. The first models that were developed were hierarchical and network databases. IBM in the seventies revolutionized this sector with the development of a relational database model. The most successful products at that time were the Oracle SQL database query language and the successors of IBM, SQL / DS, and DB2.

The concept of the database and classification of the database

Today, database systems are important in many fields of science, technology, and user applications. Any type of software developed for companies is based on reliable databases with a large number of options and tools for system administrators. Data security is also becoming increasingly important, passwords, personal data and even electronic currencies are stored and encrypted in electronic databases.

The modern financial system is nothing more than a database network in which most of the money amounts exist only in the form of electronic information units, the protection of which with the help of safe databases is one of the main tasks of financial institutions.

The concept of the database and classification of the database

Depending on the variability of the database, its type is classified according to the classification of the database as static or dynamic.

Static DB functions:

  1. Allow only read data, excluding modification.
  2. They are used for biographies and historical facts or scenarios that can be accessed for research, without the need to change the content.
  3. They are safe and easy to use when connected to a network.

Dynamic DB functions:

  1. They have the concept of self-government.
  2. May be associated with dynamic networks.
  3. This structural association allows you to store and update database information.
  4. Uses HTML as the language of communication between the network and the dynamic database.
  5. The most used languages ​​for creating dynamic networks associated with BBDD: Perl, CGI, PHP, JSP and ASP.

The main DBMSs that work with dynamic web pages are PostgresQL, MySQL, Oracle, and Microsoft SQL.

In order to understand what are the options for the classification of databases used in the scientific and educational environment, consider:

  • bibliographic;
  • documentary;
  • specialized;
  • reference books.

Functionality of bibliographic databases:

  1. Associated with old records that contain information about the location of a book or document.
  2. Do not contain the full text, only a link.
  3. Thanks to formats such as PDF, it allows you to access original articles to which there are links.
  4. With the development of technology, links from other media are included.

Features of specialized databases:

  1. They contain accurate information and are focused on a specific topic.
  2. Used in the academic and scientific environment.
  3. For some cases, they are not considered as valid BBDDs: for example, a telephone directory, a contact list of a company or an international company.

Electronic Processing Models

In order to study in detail the question of what are the options for classifying a database, one cannot bypass the topic of models. Hierarchical databases were the first developed in the 60s in the works of Hollerith, they depended on the type of information storage 1N / NN in the form of an inverted tree.

Relations are of type 1N when a parent node can have several child subnodes, but a child node cannot belong to several parent subnodes. Their disadvantage is that data redundancy is not well represented.

The network database model proposed by CODASYL is its first management system (IMS), it appeared in 1968 for NASA's Apollo program. She solved some problems of the previous hierarchical model, which are almost never used in the modern IT-process.

In order to understand the modern model, it is necessary to consider what kind of relations between the parent and child nodes exist in the classification of the database. Today, relationships like NN are used when a child subnode is allowed to belong to several parent nodes. Together with the hierarchical model, it forms the first generation of the database.

Advantages of the model: they offer excellent stability, good performance and better processing redundancy. The disadvantage of the model is the complexity of the system, which requires knowledge of programming.

Features of transactional databases:

  1. The only goal is sending and receiving data at high speed.
  2. They focus on quality analysis and production data.
  3. A unique purpose is the collection and recovery of data at the highest possible speed, so redundancy and duplication of information is not a problem, as with other databases.
  4. Allow connection to relational databases.
  5. Operations are atomic, in this type it is only possible that they are performed in full (integrity) or not performed at all.

Major differences in databases

Documentary - return content, work with cognitive and conceptual documents, belong to the intellectual and academic environment. They have document and content managers such as CDS / ISIS, Filemaker, Knosys or Imagic Text for terminological control. They are easily accessible using standardized query languages ​​and have a database classification according to the type of data model.

Relational are based on establishing relationships between data sets, organized in the form of tables that correspond to some basic requirements. They have a fixed number of fields. Each attribute has a name and many possible values. Each entry is unique and identified using a key. They implement the SQL query language and are based on a model developed by Edgar Codd in the 70s.

Golden rule

Object-oriented databases return physical files or program code, they appeared at the end of the twentieth century. Used in industrial production and design. They work with an object oriented language such as C ++ or Python. Comply with the golden rule: persistence, secondary storage manager, concurrency, recovery, and the request object.

DBMS management systems

DBMS management systems

Database Management System (DBMS) is a term for describing the functions and requirements of transactions in a database management system , in short it is ACID (ASID) from atomicity, consistency, isolation and durability. These four parameters cover the most important ACID compliant DBMS requirements:

  1. Atomicity denotes the "all or nothing" property of database managers so that the request is valid, the transaction was executed correctly and implemented in the correct order of procedures.
  2. Consistency, or coherence, when the database transaction remains stable, requiring constant monitoring of all operations.
  3. Isolation is a condition and a guarantee that transactions do not interfere with each other, which is usually achieved by blocking certain functions that isolate the data involved in the transaction.
  4. Durability means that in the DBMS all data is stored in the long term even after the conclusion of the transaction, as well as in the event of a system crash if the DBMS falls. For this condition, transaction records are required that record all ongoing processes.

Classification of functions and requirements

The database stores information and links it into a logical unit together with the metadata necessary for processing. This is a very useful tool for managing large files with a simple request, with a permission system that determines which users or programs have the right to access.

Classification of functions and requirements

DB classification:



Store data

The database stores texts, documents, passwords. In electronic format, access to data can be obtained through consultation.

To change the data

Depending on which permissions are available, most databases allow you to edit data protection filters.

Clear data

Entries in most database classification options can be completely deleted without leaving spaces. In some cases, deleted data can be recovered, but in others it is deleted permanently.

Metadata management

Typically, information is stored with metadata or meta tags that maintain order in the database and make the search function possible. Metadata is also often used to regulate permissions.

Data security

Databases must be protected to prevent unauthorized persons from accessing the information that they store.

Data integrity

Data integrity means that they must comply with certain rules to ensure their correctness and determine the business logic of the data bank.

Multi-user function

Database applications provide access from different devices. Permission allocation and data security are elementary in multi-user use.

Optimize Queries

Technically, the database should be able to handle queries in the best way to guarantee good performance.

Triggers and Stored Procedures

These two procedures are mini-applications stored in the DBMS. Triggers and stored procedures are typical relational database processes.

System transparency

The transparency of the system is relevant, especially in distributed database classification models.

Hierarchical model

Hierarchical model

The differences between the most common database models are the result of the technical evolution of electronic data transfer, which not only pursued the goals of efficiency and manageability, but also expanded the capabilities of the most famous manufacturers. This is the oldest model, which today significantly exceeds the relational, although in recent years there has been an increase in its popularity.

XML uses this system to store information. Some insurance companies and banks access hierarchical databases in the oldest applications. The most famous is the IBM IMS / DB base.

In the hierarchical model for classifying database data, there are strict and unambiguous dependencies. Each entry has only one use case (Parent-Child Relationships, PCR), with the exception of the root, which makes up the tree diagram. Although each child node can have only one parent, “parents” can have as many child nodes as they want.

Given a strict hierarchical ordering, levels that do not have a direct connection do not interact with each other, so connecting two different trees is not easy. At the same time, hierarchical database structures are extremely flexible and understandable. Records with “children” are called records, and those without are called leaves, and are usually documents in the records for leaves in the database classification. Queries on the hierarchical database reach the leaves, starting at the root and going through various records.

Graphically Oriented DMS

The network model developed almost simultaneously with the relational model, although over time it was defeated by competitors. Unlike the hierarchical model, the entries here do not reveal strict parent-child relationships, but each one can have several precedents, which gives it the network structure of its name. There is also a unique and unchanging path to access the record.

There is no fixed hierarchy in the network database model, and therefore there are several paths leading to the same destination. The record located in the center of the image can be theoretically accessible from five others, and having access to it, you can access five other records.

In the network model, dependencies can also be defined - the register located above. It is not directly connected with the register in the extreme right position, therefore, to achieve it, it must go through the register in the center, which it can accept or reject. You can contact the top left. In the network model, entries are added or deleted without affecting the global structure.

Today, this model is used on large computers. Other areas still rely on a hierarchical model or refer to a relational model that is much more flexible and easy to use. Some well-known models of network databases are UDS Siemens and DMS Sperry Univac. Over time, both manufacturers also developed interesting mixed forms between the network model and the relational model. Due to its reticular structure, a graphically oriented database is considered a modern evolution of the network model.

Graphically Oriented DMS Sperry Univac

Storage Scalability

In a document-oriented database model, documents are the main unit of information storage. These units are the ones that structure the data and should not be confused with the documents of word processing programs. Here, data is stored in so-called key-value pairs.

Since neither the structure nor the number of pairs has been determined, the documents that make up the document-oriented database can be very different from each other. Each document in itself is a closed unit, and establishing relationships between documents is not easy.

In recent years, thanks to the success of NoSQL, documentary databases have experienced a big boom, especially due to their good scalability. An example of a database system of this type is MongoDB. In a document-centric database model, data is stored in separate documents, and not in tables, as in the relational model.

These systems are especially interesting for web applications because they allow you to save full HTML forms. It must be emphasized that among the various document-based systems there are notable differences, from syntax to internal structure, so not all document-oriented databases are suitable for this scenario. It is precisely because of these differences that several database systems exist that target Lotus Notes, Amazon SimpleDB, MongoDB, CouchDB, Riak, ThruDB, and OrientDB reputational documents.

Advantages and disadvantages

Advantages and disadvantages

Proper database management systems help you get better access to your data and optimize its management. In turn, point access helps end users quickly and efficiently exchange data as part of the organization’s tasks.

Database model

Year of creation





Very fast read access, clear structure, technically simple.

Fixed structure in the tree, which does not allow communication between trees.


Early 1970s

Supports several ways to access records, without a strict hierarchy.

Poor review with large databases.



Simple, flexible creation and editing, easily extensible, quick commissioning, simple extension, quick launch, very dynamic context.

Unmanageable with large amounts of data, poor segmentation, artificial key attributes, an external programming interface, poorly reflects the properties and behavior of objects.

Object Oriented

Late 1980s

The best support for object-oriented programming languages, storage of multimedia content. Supports object-oriented programming languages, allows you to store multimedia content.

Lower performance with large amounts of data, few compatible interfaces.

Document Oriented


The relevant data is stored centrally in independent documents, free structure, multimedia concepts, refers to the classification of database entities.

Organizational work is relatively high, often requiring programming skills.

Areas of use

Areas of use

You may not be aware of this, but databases are everywhere. Regardless of whether he knows anything about them or not, their influence on everyday life is very great. From weather applications to online movies, databases are responsible for many of the services that people use every day, and in order not to get confused by the increased amount of information, they use data classification in the database.

DBMS application areas:

  1. Banking - for customer information, accounts and loans, as well as banking operations.
  2. Airlines - for reservations and schedule information. Airlines were among the first to use databases in a geographically distributed manner: terminals located around the world accessed a central database system via telephone lines and other data networks.
  3. Universities - for student information, course registration, and grades.
  4. Credit card transactions - for purchases with credit cards and the formation of monthly statements.
  5. Telecommunications - for keeping records of calls made, compiling monthly bills, maintaining balance on prepaid phone cards and storing information about communication networks.
  6. Finance - to store information on stocks, sales and purchases of financial instruments, such as stocks and bonds.
  7. Sales - information about the customer, product and purchase.
  8. Production - to manage the supply chain and to track the production of goods in factories, stocks of goods in warehouses, shops and orders for goods.
  9. Human resources - to obtain information about employees, wages, payroll taxes and benefits, as well as to receive salaries.

Future trends

Future trends

In the future worldview of databases, the World Wide Web (WWW, or, in abbreviated form, the web) will continue to be an important aspect as a means of publishing documents and as a means of exchanging information. WWW provides one of the most diverse and complex environments in the field of interaction.

Recently, technologies and standards have emerged to make the network a scalable and manageable infrastructure. One such technology is XML, which is transformed into the Internet, into a database system, in the style of traditional DB handlers, which gives much better results than search engines. The challenge is to integrate this functionality into XML and make the most of the strategic information that a user can find on the Internet.

New trends include proactive and predictive performance analysis, database load testing, the use of NOSQL - mongodb and cassandra and BigData (Hadoop) in enterprise and cloud environments.


All Articles