Single source of truth is different in distributed systems, well the costs are different.
Using the language of 'software architecture: thee hard parts' this forces your entire system into a single 'architectural quanta', basically it is a monolith.
There are situations where monoliths are acceptable or even the least worst option.
The 'Fallacies of distributed computing' covers most of these.
It can work better if you use a hierarchical model, but OLTP is almost exclusively the relational model.
ACID transactions become pretty brittle, software becomes set in stone etc...
There are use cases, but the value proposition is typically much weaker
Using the language of 'software architecture: thee hard parts' this forces your entire system into a single 'architectural quanta', basically it is a monolith.
There are situations where monoliths are acceptable or even the least worst option.
The 'Fallacies of distributed computing' covers most of these.
It can work better if you use a hierarchical model, but OLTP is almost exclusively the relational model.
ACID transactions become pretty brittle, software becomes set in stone etc...
There are use cases, but the value proposition is typically much weaker