FoundationDB's lesson: A fast key-value store is not enough

> 2015

How did this play out in the last 9 years? Anyone able to assess?

Clearly the points in the article are still meaningful for understanding the tradeoffs in database architecture. But it’s also not clear to me that the take “build a custom storage layer instead of trying to generalize” won either. It’s a big cookie to eat for even a medium sized company. And it’s been a very active space, so it’s not for a lack of trying.

As a curious bystander, it seems the gold medalist in databases, – despite its crippling web scale shortcomings – is Postgres, even more so today than a decade ago. It seems at least nobody has slapped some consensus/distributed tech onto it and built an obviously superior product. Similarly, the noSQL slash weakly consistent horizontally “web-scalable” relational databases seem to have grown in a different direction - towards analytics, telemetry, fraud detection - ie nobody wants to put their core source-of-truth business data there if they can avoid it.

I’m probably a terrible person to “check the pulse” on the industry though. Maybe someone with real experience can chime in?