PlanetScale's move is intriguing, but sharding data across nodes is where the real challenge lies. Are the trade-offs similar to Vitess with MySQL, or does Postgres's feature richness complicate things further?
When data grows beyond a single machine, the hard work begins. Distributed systems sacrifice features like complex joins, extensions, and strong consistency. Maybe they should get Jepsen to poke at their setup. Who can show us what we lose compared to standard Postgres?
replies(1):