Sharding (huge data and local distribution, even worldwide) and HA by retaining serializable transactions. Possibly easier to operate.
The downsides are:
- slower - Postgres (if it can handle the amount of data, which is very much on proper hardware and partitioning of > 1B row tables) is much faster, esp. for joins
The downsides are:
- slower - Postgres (if it can handle the amount of data, which is very much on proper hardware and partitioning of > 1B row tables) is much faster, esp. for joins
- features
- ecosystem (see the countless extensions)
- cost of course