About This Talk
At the Sibur Digital Community’s DWHard Meetup in 2021, I shared my perspective on Data Mesh — the decentralized data architecture paradigm that was generating significant buzz in the industry. The talk separated the hype from the substance, exploring what Data Mesh actually means in practice and when it makes sense to adopt.
The Problem
Data Mesh emerged as a response to a real problem: centralized data teams become bottlenecks as organizations scale. Every request goes through the same team, priorities conflict, and domain expertise gets lost in translation. Data Mesh proposes a radical alternative: give data ownership to domain teams.
But the gap between the theoretical framework and practical implementation is wide. Most organizations that attempt Data Mesh underestimate the organizational change required and overestimate their readiness.
Key Ideas
What Data Mesh Actually Is — Not just “decentralized ETL.” Data Mesh is an organizational and architectural paradigm built on four principles: domain-oriented ownership, data as a product, self-serve data infrastructure, and federated computational governance. All four must work together.
When It Makes Sense — Data Mesh is not for every organization. Prerequisites: sufficient organizational scale (usually 100+ data consumers), clear domain boundaries, engineering maturity within domain teams, and executive sponsorship for the organizational change.
The Hard Parts — Governance in a decentralized model is genuinely difficult. How do you ensure quality standards across autonomous teams? How do you prevent duplication? How do you handle cross-domain data products?
Practical First Steps — Rather than a big-bang transformation: start with one domain, define what “data as a product” means concretely (SLAs, documentation, discoverability), and build the self-serve infrastructure gradually.
Why It Matters
Whether or not you adopt the full paradigm, the principles — domain ownership, product thinking, self-serve infrastructure — are valuable for any data organization. Understanding the trade-offs helps you take the right ideas without the unnecessary risk.
