Should we transition from a centralized Data Lake to a decentralized Data Mesh architecture?

All answers to this question.

Moving to a Data Mesh is as much a cultural shift as it is a technical one. In my experience at a FinTech firm, we only succeeded when we treated "Data as a Product." This means each domain team must have their own data architects and engineers. We used a global schema registry to prevent the "silos" you’re worried about. The biggest hurdle wasn't the technology, but getting business units to accept accountability for their own data quality. If your domains aren't ready to own the lifecycle, your mesh will quickly turn into a mess.

   Answered 2025-05-18 by Jennifer Hayes

How are you planning to handle the "self-serve" infrastructure layer? Without a unified platform for domains to deploy their data products, you'll just end up with ten different ways of doing ETL.

   Answered 2025-06-10 by Robert Miller

• Robert, we are looking at using a standardized Terraform-based platform on AWS. The idea is to provide "blueprints" for S3 buckets and Glue jobs so every domain follows the same security and tagging standards. It’s the only way to keep the federated governance from becoming a total manual nightmare for the central IT team.

     Commented 2025-06-15 by David Clark

Start small with one "lighthouse" project. Pick a domain that already has high data maturity to prove the Mesh concept before rolling it out company-wide.

   Answered 2025-07-02 by Barbara Wilson

• Completely agree, Barbara. We started with the Marketing domain because they were the most frustrated with the central team, and their success made the rest of the company eager to join.

     Commented 2025-07-05 by Sarah Thompson