Data Mesh
Definition
A decentralized data architecture approach where domain teams own, share, and manage their data as products, enhancing collaboration and agility.
Use Cases
- Netflix: Scaling analytics and data sharing across many autonomous engineering teams — Netflix has publicly described a decentralized data platform approach where teams publish and consume datasets via a central data platform with strong metadata, governance, and self-service tooling (often cited as aligned with Data Mesh principles). (Improved team autonomy and faster delivery of data capabilities by reducing reliance on a single centralized data engineering bottleneck.)
- Zalando: Enabling domain-oriented data products across a large e-commerce organization — Zalando has publicly discussed adopting Data Mesh concepts, organizing around domain-owned data products supported by shared platform capabilities (e.g., standardized interfaces, discoverability, and governance). (Better scalability of data ownership and clearer accountability for data quality and usability within domains.)
- Intuit: Sharing trusted data across business units while maintaining domain ownership — Intuit has shared public talks about moving toward domain-oriented data and platform patterns that align with Data Mesh, emphasizing self-serve data platforms, governance, and reusable data assets. (Faster cross-team data access and improved reuse of data assets through clearer product-style ownership.)
Frequently Asked Questions
- What's the difference between Data Mesh and a data lake?
- A data lake is a storage and processing pattern (often one big repository). Data Mesh is an operating model: it assigns ownership of data to domain teams and treats datasets as "data products" with clear contracts, documentation, and quality standards. You can build a Data Mesh on top of a data lake, a data warehouse, or both.
- When should I use Data Mesh?
- Consider Data Mesh when your organization has many domains, many data producers/consumers, and a centralized data team can’t keep up with demand. It works best when domains can take on ownership (people, skills, and accountability) and when you can provide a strong self-service platform (catalog, access controls, pipelines, observability) plus federated governance.
- How much does Data Mesh cost?
- There is no fixed price because Data Mesh is not a product. Costs usually come from (1) cloud compute/storage for pipelines, lakes/warehouses, and streaming, (2) tooling for catalog, governance, data quality, and observability, and (3) organizational investment—domain teams need time and skills to build and maintain data products. Costs can rise if data is duplicated across domains or if governance and standards are weak, leading to rework.
Category: data
Difficulty: advanced
Related Terms
See Also