Federated Data Architecture
summary
UniCredit Bank Serbia adopts a modified Data Mesh: a source-first model that builds one source per source-aligned data product and composes consumer-facing products from those pieces. A 1.5‑year pilot commits to deliver 30 data products in 6 months, publish them on a Power BI data marketplace with owners and lineage, and tier access as basic/gold/platinum. Governance runs in a SaaS tool with about 800 automated controls, aiming to cut 30–40% of ad‑hoc report requests via self‑service.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 0:47Federated data architecture principles
- 3:13Building data products from sources
- 6:02Six-month delivery cycles
- 8:08Defining data product ownership
- 10:36Assigning owners to domains
- 13:13Financial analysis data products example
- 15:42Final architecture and BI goals
- 17:52Data quality via ownership
- 18:55Three-tier governance levels
- 20:22Q&A: Business benefits
- 23:43Q&A: Granularity and alignment
- 26:10Q&A: Golden source and validation
- 29:20Q&A: Governance controls
related talks
30mData Mesh – Architecture And Insights From The Self-service Platform Boosting Financial Services
23mData Mesh Architecture in Banking
48mBuilding a data mesh
01h 02mShift Data Quality Left with Data Mesh Principles
43mUntangling Data Mesh: A paradigm shift in Big Data
31mData Mesh In Practice – How To Set Up A Data Driven Organization
43mWhen design meets data: Using Data Mesh for a consumer and business-driven data strategy
23mApplying Data Mesh to Your Organization