Data mesh
Domain teams own and publish their analytical data as products, instead of one central data team cleaning everything in a shared lake.
Zhamak Dehghani described it on Martin Fowler’s site with four principles: domain ownership, data as a product, a self-serve data platform and federated governance. It applies the bounded context idea to analytics: the team that knows what a “booking” means publishes the booking data set, documents it and answers for its quality.
It is an organizational change more than a tool, and it pays off at the size where one central data team is the bottleneck. Smaller companies take the useful part: a named owner per data set (code ownership for data), published schemas, and feeds built with change data capture rather than reports reading production tables directly.
Sources
- Zhamak Dehghani: Data Mesh Principles and Logical Architecture martinfowler.com
Related terms
- Bounded context: A part of a system where each business term has one meaning and one model, with explicit translation at its edges.
- Code ownership (CODEOWNERS): Every part of a codebase has a named team responsible for it, often enforced by a CODEOWNERS file that requests that team’s review on every change.
- CDC (change data capture): Reading committed changes from a database’s own transaction log and streaming them as events to other systems.
- AI-ready data: Company data an AI system can use safely: accessible by software, documented, current, with an owner and with permissions that carry over to the AI.
