Understanding Data Mesh

Data is growing at a much faster rate than ever. Businesses have already realized the potential of using data for better operations and decisions. With this increasing data, there should be a change in how to manage that data effectively so that business analytics can use data effectively.

Data Mesh is the approach that focuses on getting value from the Data effectively. Treating data as a product is at the core of the Data mesh strategy; let me repeat, not data products but Data as a product.

Let’s see what the are principles of the Data mesh.

• Domain Driven Data ownership: Distributing ownership of data from one centralized team to the people who are most qualified and appropriate to handle it — often, the business domains where the data comes from
• Data as a Product: Treat Data as a product. Data should fulfill the requirements of the domain but also provide high-quality data to other domains via APIs.
• Self-Service Infrastructure: The data platform team should enable domain teams to consume and create data products seamlessly; for this purpose, Empower teams with a self-serve data infrastructure
• Federated Governance: As the name suggests, this is regarding the governance to create a data ecosystem with a commitment to the organizational rules and industry regulations.

I know that The data mesh is not a technical solution or even a subset of technologies — it’s an organizational paradigm for addressing and operationalizing data and getting real value. However, considering above mentioned principles, I believe data teams require the following capabilities to achieve data mesh architecture.

• Domain Driven Data Ownership

To divide data into domains, data needs to be accessed from different sources seamlessly. Data modeling is also equally crucial for domain creation. All the dividing and forming of new domains is impossible when you don’t know reference data about the data

• Data as a product

Data as a product is product thinking around data for better analytics. It relies on creating value from the data. Data Access, Data Quality, Data Enrichment, and Data Delivery are the capabilities required to successfully create value out of the data. Data Products should have the features mentioned above so that every product can be easily identified and shared.

Self Service infrastructure and Federated governance principles can be easily fulfilled by using cloud-native / cloud-first technology, which will empower data teams to create/manage the infrastructure with ease and required governance rules.

In the next blog post, we will talk more about the practical questions around the adoption of data mesh