1. 8
  1. 4

    This feels like a bad justification for new technical architecture decisions:

    Many organizations have invested in a central data lake and a data team with the expectation to drive their business based on data. However, after a few initial quick wins, they notice that the central data team often becomes a bottleneck. The team cannot handle all the analytical questions of management and product owners quickly enough.

    It’s a bad move to put all your data in one place and then task a single team with the role of answering questions about it. You should put all your data in one place and then make querying it easy for any team to do, no mesh needed. Maybe I’m spoiled at my present dayjob but we have a very big data warehouse for literally everything with analytical tools available to everyone in the company who can write a bit of SQL, currently tens of thousands of people. When I come in contact with dedicated data teams I can usually assume that they won’t make their time available to my team and we still have to self-serve.

    Am I just restating the data mesh vision here? If so, why don’t I recognize our approach in this article?

    1. 1

      Hi migurski! I am one of the authors of datamesh-architecture.com. Thanks for describing the data approach in your company.

      You might have data mesh. Two important concepts of data mesh are not mentioned in your description:

      1. Domain Ownership: Data is only valid within one domain/bounded context (see Strategic Domain-driven Design) and needs to be curated and described that other domains can get an understanding what data actually means. The domain providing data is responsible for the selection and quality of the provided data products. Which team ingests data into your platform?

      2. Federated Governance: Do you have rules for the domain teams, how data is provided and how keep them interoperable? Do you have a board that agrees on these rules?

    2. 1

      Is this basically a buzzword for “don’t build a data warehouse”? As someone who has built more than one data warehouse I think this might be the buzzword I need.

      1. 1

        Well, it depends… Data Mesh can rely on data warehouse and data lake technologies. The important thing is the organizational shift of ownership from a central data team toward domain teams.

        1. 1

          So it’s basically the same as the “democratization of data” buzzword but useful for leading people away from data warehouses?

          1. 1

            Yes, data mesh is a way to achieve democratization of data.