Data Mesh - a hot air balloon?
summary
Argument lays out concrete claims and tools: Data Mesh (coined 2019) reorganizes ownership into domain teams, defines four principles (domain ownership, data as a product, self‑serve platform, federated governance), and expects platform pieces from Snowflake, Databricks, Azure hub‑spoke or Microsoft Fabric. Examples: an internal Snowflake marketplace is built and used to share products; spinning up cloud landing zones hits real cost and governance friction. Hard numbers and verdicts: adoption cost is high and 80% won't need full mesh; hybrid or staged implementation often works better.
AI-written summary · about
chapters (20)
- 0:00Data Mesh concepts
- 1:10Data analytics evolution
- 6:30Zhamak Dehghani and Data Mesh
- 7:58Problems Data Mesh solves
- 9:28Four principles of Data Mesh
- 14:47Data product usability
- 15:29Self-serve data platform
- 16:12Federated governance
- 20:10Data Mesh overhyped or revolutionary
- 22:17People over technology
- 23:42Data Mesh patterns
- 26:31Multi-cloud Data Mesh example
- 28:59Microsoft Fabric misconceptions
- 31:05Organizational challenges
- 33:56Producer challenges
- 34:38Consumer challenges
- 37:06Data product contracts
- 38:54Suitability for everyone
- 40:38Hype and expectations
- 42:52Q&A
related talks
38mData Mesh, Now
39mData Mesh: Why, What and How
43mUntangling Data Mesh: A paradigm shift in Big Data
36mData Mesh and Its Superpowers
20mEmbracing Data Mesh: A Modern Strategy for Scalable Data Management
51mData Mesh – Successful with People, Process, and Technology
50mData Mesh: the new paradigm for data management
13mData Mesh is a Big Idea: Take Your Team From Vision To Reality