Data Mesh and Governed Data Democratisation

How federated governance supports data mesh

Data mesh is a socio-technical approach for managing analytical data at scale. Its four widely cited principles are domain-oriented decentralised data ownership and architecture, data as a product, self-serve data infrastructure as a platform, and federated computational governance.

Federated Data Governance and data mesh are related, but they are not the same thing. An organisation can adopt federated governance without implementing a full data mesh.

Domain-oriented ownership

Business- or governance-aligned domains and accountable roles place decisions closer to the data. In Microsoft Purview, Governance Domains, domain roles, Data Product ownership and data estate mappings can support this governance view.

Data as a product

Priority data is packaged around consumer use cases with clear ownership, context, quality and access information. Purview can support Data Products, Owners, glossary terms, Data Quality, lineage and classifications where supported.

Self-serve data platform

Consumers and domain teams should be able to use data and platform capabilities without unnecessary central bottlenecks. Unified Catalog can support discovery, metadata and governed request experiences, but the broader self-serve data platform sits outside Purview.

Federated computational governance

Enterprise policies should be applied consistently while domains retain appropriate autonomy, with controls automated where platforms support it. Purview offers governance mechanisms such as health controls, inherited policies, classifications and Data Quality rules, but these are only part of full computational governance.

Benefits when these principles work together

• Data ownership can scale with the number of domains and Data Products.
• Consumers can get a more consistent experience even when data remains distributed across many platforms and teams.
• Domain teams can improve products without losing enterprise interoperability.
• Central governance can focus on standards, assurance and cross-domain concerns instead of curating every asset itself.
• Reusable governed Data Products can create a stronger foundation for analytics, AI and cross-domain insight.

Governed data democratisation

Data democratisation means putting useful data into the hands of more people who need it to solve business problems and make informed decisions. It does not mean unrestricted access, bypassing security or removing role-appropriate privacy, regulatory and access controls.

• Discoverability: consumers can search for governed Data Products instead of relying on personal contacts or specialist gatekeepers.
• Understandability: business descriptions, glossary terms, metadata and ownership make data easier to interpret.
• Trust: quality, lineage, classification, limitations and stewardship provide evidence for responsible use where supported.
• Governed access: consumers can understand the access route and request appropriate access in context.
• Reuse: existing governed products become visible before new extracts, reports or pipelines are created.
• Reduced shadow activity: a usable governed route gives people less reason to create uncontrolled copies or work around official processes.
• Supports wider data literacy: governed self-service can help people develop a better understanding of ownership, quality and appropriate data use.

Governed self-service, not open access

The goal is to remove unnecessary friction while preserving need-to-know, least-privilege, privacy, security and regulatory controls. Data should be easier to discover and understand without making inappropriate access easier.

Framework reference

The Data SkyLab Studio framework uses the four data mesh principles described by Zhamak Dehghani in Data Mesh Principles and Logical Architecture.