Four Years of Data Mesh in Practice
summary
Claims four years of production experience and a concrete reference implementation: a CLI that provisions a data product in about 25–30 minutes, multicloud support (AWS EMR + S3 or Azure Databricks + blob) with Apache Atlas / Azure Purview for cataloging, Open Policy Agent checks, per-domain accounts and sidecar control ports. Reports repeatable wins—multi‑year data access waits reduced to weeks—while calling out hard limits: governance, culture, data reliability and tooling gaps. Advocates spec‑driven deployments and devops foundations for federated automated governance.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 4:00What data mesh is
- 13:51Data mesh architecture
- 18:07Lessons learned
- 29:08Organizational change
- 31:13Governance challenges
- 36:53Cloud strategies
- 40:01Data mesh vs fabric
- 41:05Q&A
related talks
42mFour Years of Data Mesh in Practice
46mA Data Mesh Reference Implementation on AWS
39mData Mesh in the wild
47mData Mesh in Kubernetes
54mData Mesh in Practice
51mWhat is a Data mesh?
40mData Mesh, Data Fabric and DataOps; what should you choose?
30mFrom Data Lake Entanglement to Data Mesh Decoupling: Scaling a Self-Service Data Platform