Posted in:

What are Data Mesh and its 4 Principles in a Nutshell?

© by Shutterstock

Designing a Central Data Warehousing or a Data Lake is the prime function of various large-scale organizations. These organizations deal with millions of daily transactions that need to be analyzed for forecasting, reporting, or ML projects.

But what is data mesh, and where does it come in here?

Data Mesh in a Nutshell

Data with its associated operations and infrastructures are distributed.

In many cases, data is produced by a specific service or team, for consumption by one or more services and teams. The production data is highly specific for the source services, in terms of format, structure, when and how it gets available to external services, and the specific permissions needed to access it – this list is long and goes on. The team which owns/produces the data is naturally familiar with all these idiosyncrasies – however, others are not.

What is Data Mesh and its Current Implications?

The world has gone through a metamorphosis and turned effectively digital.


The pandemic has only accelerated the global trend. Successful organizations in every industry and sector of the economy from banking to retail to life sciences – are looking for better ways to capitalize on and harness the abundance of data.

But there are some challenges to reaching the full throttle of data value, including;

  1. Only 32 percent of IT leaders realize the tangible value of data;
  2. With 77 percent integrating up to 5 types of data in their data pipelines;
  3. Out of which, only 3 percent of data meets quality standards;
  4. More than 65 percent of enterprises use at least ten varying data engineering tools;

Most new applications are now being designed with domain-driven design. These applications are intact with data that is specific to the application – this brings on a new challenge for the database engineering teams to come up with an organized solution – serving the purpose for all aspects.

Are the 4 Principles of Data Mesh a solution?

The Data Mesh offers the same functionalities for serving the domain-driven purpose of decentralization. For designing a Data Mesh – the 4 principles are followed when it comes to different responsibilities for different teams of an organization.

These principles are guidelines for running successful data mesh projects in any enterprise; these include;

Data Ownership by Domain

Coming up with the domain-driven decentralization approach, with data mesh, the data is divided around a specific domain – similar to the microservices. The same will apply to Data Domain, too – with the Data Domain team being there to keep tabs on liveliness. The data domain team can utilize the data and build data products that are put to good use by other data domain teams.  

Data as a Product

Inside a data mesh, the data classifies as a product that can be published by one and consumed by another data domain team. The teams also engage in the product – being completely responsible for the data, quality, cohesiveness, representation, and more. The domain team also engages with data mesh enabling the team for data product entitlement.

Self-Driven Data Approach

The data in a data mesh is available anywhere in an organization under the self-driven approach platform. This helps in producing new reports or data products in a shorter time – while also being able to be propagated for subsequent data products.


However, this does come with governance issues, as controlling data can only be done with a governance policy in place.

Federated Governance

The governance gets handled with different data and security policies – and these come with varying contracts by data domain teams – subjective to data publishing and consumption. Governance can become an issue for the data if the policies are not defined clearly and correctly.

Data mesh Maximizing Data Value

  • Error-Prone
  • Time Taking
  • Unsustainable
  • Un-Scalable

– if data management sounds like this at your organization – a data mesh model is what you need.

What is data mesh – to sum up, the data mesh is designed to decentralize the IT team’s load lifting and data platforms. The data mesh model efficiently transfers the onus of data management to independent business domains within an organization. Domain data ownership requires truly knowledgeable, and expert personnel to control data. Instead of boiling the entire data lake, multiple business domain teams can focus on cleaning data, and its trustworthiness, while supporting business agility. This enables organizations that are building data management architectures with self-services to easily give the right data access to consumers whenever they need it.

Data mesh is truly a star game-changer for enterprises, offering a framework that removes bottlenecks by empowering business domains. With data mesh models in place, the independent business domains can quickly curate and produce data at an enterprise scale.

Stay tuned with Memphis for more details on each of the 4 principles of data mesh.