Data Mesh Architecture

Data mesh is the organizational architecture that applies microservices principles to data. Instead of a centralized data team that owns the data platform, each domain team owns its data as a product — discoverable, addressable, self-describing, and governed. The centralized team provides the platform infrastructure; the domain teams provide the data products. This is the architecture for organizations where the central data team has become the bottleneck between data producers and data consumers.

1. Domain Data Ownership Each domain team owns its data end-to-end: ingestion, transformation, quality, and serving. The order team owns order data. The customer team owns customer data. Ownership means accountability — the team that produces the data is responsible for its quality, freshness, and discoverability. This eliminates the centralized bottleneck where every data request routes through a single team.
2. Data as a Product Domain data is treated as a product with consumers. Each data product has an SLA (freshness, availability, quality), documentation (schema, semantics, known limitations), and discoverability (registered in the catalog with metadata). The product mindset shifts data from a byproduct of operations to a first-class deliverable with quality standards.
3. Self-Serve Platform The centralized platform team provides infrastructure that domain teams use to build data products: storage, compute, cataloging, access control, and monitoring. The platform is a product that serves internal teams. The domain teams use the platform to build their data products without reinventing infrastructure. The platform abstracts complexity; the domains provide context.