Data Architecture
Introduction to the world of data architecture
Data forms one of several areas within an enterprise architecture. The data architecture consists of models that are based on guidelines, rules and standards. The aim is to clearly regulate which data is collected and how it is stored and used. Due to the increasingly stringent data protection regulations, it is practically essential for companies today to deal with their data architecture and organise it in such a way that data is managed in compliance with regulations and at the same time generates business benefits. Modelling also enables a good overview of the data flows to be maintained, compliance to be met in a traceable manner at all times and optimum benefits to be derived from the data stocks.
Maintaining an overview of data streams with data architecture
Data modelling is an important part of data architecture. The main objective is to clearly define and specify the objects to be managed in an information system in order to obtain an overview of the data view of the information system. The results of data modelling are data models that lead to usable databases or datasets. These models have a much longer lifespan than functions and processes (software) and therefore represent a solid element of the data architecture.
Data modelling is also used to depict certain facts. For example, data from a company division or business processes can be recorded and documented with their interrelationships. As part of this use, standardised terms are often defined to facilitate further communication.
A data architecture aims to define a model of possible interactions for all data systems. Data integration, for example, must be subject to clear architectural standards as it requires interactions between data systems. The data architecture also describes data structures that are used by a company. Last but not least, it also provides criteria for controlling the data flows in the system.
The data architect is usually responsible for defining the target state and must also ensure that enhancements are carried out in line with the original design. To do this, he breaks down the process into the three traditional architecture levels: Conceptual (business units), Logical (relationships between the data) and Physical (data mechanisms).
In a more comprehensive view, data architecture includes a complete analysis of the relationships between the functions of an organisation, the available technologies and the data types.
Model-based data architecture offers several advantages that can be of great benefit to companies in the effective management and utilisation of their data assets. Here are some of the key benefits:
Comprehensive data visualisation through a holistic view and representation of the relationship between data components
Alignment with business objectives by harmonising the data architecture with the business objectives
Improved data quality and consistency through standardised data models and easier data management
Improved data integration and interoperability through modelling of interfaces and coherent representation of different data sets
Simplified data analysis and insights through clear structuring and illustrative visualisation of complex data relationships
Scalability and agility improve the adaptability to new business requirements
Cost cutting and increased efficiency through optimised data infrastructure and improved management over the entire data life cycle
Here you can find practical use cases from our customers:
KBC Group uses Sparx Enterprise Architect for architecture modelling
Nictiz builds a reference architecture for hospitals with Enterprise Architect
FAQs on Data modelling
What is Data modelling?
Data modelling is an essential part of software development and can therefore also be carried out with a UML tool (e.g. Enterprise Architect), just like the entire software and system development. It takes place over different project phases. The results are data models that ultimately lead to usable databases after several modelling stages.
Data models usually have a longer lifespan than software, which has to be constantly adapted to new circumstances. Data modelling is also used for application development in order to depict certain facts. For example, a business process (or the entire company) is modelled and documented with all dependencies and interrelationships.
What is the data modelling process like?
In data modelling, data objects are represented schematically in order to depict a specific context in a system.
Here are some steps of the process:
- Understanding the requirements: What data and what relationships should be captured?
- Design: The data flow into and out of the database is depicted in a diagram.
- Normalisation: To avoid redundancies, the data is divided into tables and set in relation to each other
- Selection of the database management system (DBMS)
- Implementation: The data model is transferred to the DBMS
- Documentation: The data model also serves as documentation
- Optimisation: The data model is continuously improved to enhance performance
How can companies benefit from data modelling?
A clearly structured data model enables companies to utilise their data effectively and remain competitive:
- Efficient data processing: By structuring data in tables and defining relationships, queries can be carried out more quickly
- Fewer redundancies: The normalisation step saves storage space and ensures consistent data
- Data integrity: The data model contains clear rules for data integrity
- Scalability: Scalable data models ensure easy expansion of the database
- One data source: The data model is the central source for all systems in the company
- Decision-making: Sound decisions are based on consistent and reliable information.
- Compliance and security: Security guidelines and data protection rules can be easily implemented in the data model
- Cost savings: Data modelling leads to efficient data processing, minimisation of redundancies and better scalability
