Data Mesh

Data mesh is a decentralized data architecture where domain teams own and serve their own data products rather than a centralized data team building monolithic pipelines. Each domain (property listings, pricing, user accounts) treats its data as a product with a defined API, SLA, and documentation - following a 'data as a product' principle.

The four principles of data mesh are: domain ownership (teams own their data), data as a product (data meets quality and usability standards), self-serve data infrastructure (a platform team provides tooling but not pipelines), and federated governance (standards and policies applied uniformly without centralizing all control).

Data mesh addresses the scaling limitations of centralized data lakes and warehouses, where a single data engineering team becomes a bottleneck for all analytical needs. By distributing data ownership to domain experts, data mesh creates higher-quality, more trustworthy data products - at the cost of requiring more engineering capability distributed across teams.