New data platforms have marked a significant shift in how businesses manage and interpret vital information. However, beyond technological considerations, we want to focus on domain modelling, an innovative approach that differs from traditional practices. Furthermore, this approach promises greater agility and alignment with the company’s business needs.
In other words, when starting a New Data Platform project, you should avoid recreating a Monolithic Data Model. This will hinder the agility and benefits of your new platform.
Beyond Data Mesh: A Pragmatic Approach to Data ETL
Firstly, let’s address what we will not cover: the Data Mesh. Although structured for large companies, its complexity and multiple interpretations make it less accessible. (We are big fans of Data Mesh, but it’s not accessible to many companies. It requires significant data maturity across teams for deployment).
Instead, let’s focus on outdated practices that you should avoid. Domain modelling stands out from the three-zone vertical models of old BI, which were IT-centric and limited by hardware and technology constraints. This resulted in rigid and inflexible data models.
Autonomous Data Domains: A New Paradigm for Data ETL
Imagine autonomous data domains, each existing independently and interacting with others, like the services of a digital platform. These domains communicate via Data Contracts, similar to APIs in the digital world. This flexibility allows each company to define its own data domains tailored to its specific needs.
Examples of Data Domains in Data ETL
We often associate data domains with the company’s business, but this is not always necessary. The goals of these areas can vary significantly:
- Technical (like an SAP domain for SAP data)
- Professions (Procurement, Finance, etc.)
- Functional (Customer, Product, etc.)
- Logic (Application, Sandbox, Analytics, Marketing, etc.)
Consequently, each domain can evolve, split, merge, or even have a limited lifespan, providing unprecedented flexibility.
Conclusion
Today’s data platforms are not what they used to be. Domain modelling is a step towards more agile, responsive, and inherently more powerful data management. You may notice that there is only one step to go from a domain approach to a Data Product approach.
Devoteam helps you modernise your data.
With a team of 1,000+ data consultants with over 960 certifications across leading cloud platforms like AWS, Google Cloud, Microsoft, DataBricks and Snowflake, Devoteam helps you build a modern, scalable data architecture.