Cheers, Sven,

Thank you for the answer.

I saw opinion that data warehouse could be treated as a separate node, but that sounds odd to me.

Let's split data warehouse and data lake because it is different things.

I would agree the data lake could be a node of a data mesh where people throw everything not knowing upfront if they will use it later, not necessarily quality assured, and might be useful for data scientist for data exploration or for some other use cases.

But if you have data warehouse as one node, and at the neighbourhood consumer-oriented datasets with orders, customer, and marketing datasets (examples from your article) that would make Data Warehouse to be called Data Mart with only subset of data, not the integrated and consolidated DW by definition. And that can lead to issues/challenges I described in my article.

Could you please also answer a question, if you practice data mesh approach:

* How many domains are currently in your solution?

* How many data engineers are dedicated to each of them? Do you have a data owner in each of the domain?

Thank you,


Big Data Architect & Data Warehouse Expert at SoftServe Inc.