Data Mesh explained simply

Airports and borders don’t care about me at all. It doesn’t matter whether I make it into the USA within 20 minutes, an hour, or three.

But that’s not the case for everyone. Some people are really important, like a CEO of a big company or a diplomat. Then there are the people who are somewhat important and fly often, the usual business travelers.

What’s the worst thing that could happen to those two groups of people? To be stuck in a long queue behind me, lots of travelers like me.

The solution? I like to call it decentralization. But that’s just a fancy word…

We let those people go through the global entry or even fly them to a separate airport; we remove the central bottleneck so that I cannot cause problems for the frequent and somewhat important people or for the infrequent, really important people.

With the data mesh, we do the very same thing. If we see congestion, we remove it by decentralizing whatever the central bottleneck is. But we don’t throw it away; we simply take our VIPs into a separate queue to move them along faster.

That’s why there’s a completely separate team dealing with all the data and everything around the recommendation engine at Netflix; they don’t go through the central data warehouse; they got their own queue.