PostgreSQL's Distributed Evolution: What changed and what stayed the same in the Cloud

Moving PostgreSQL to distributed, cloud-native architectures exposes assumptions that work well on single nodes but break across multiple nodes. This session examines the technical details through concrete examples — explain plan differences, cardinality estimation across nodes, transaction coordination approaches, and consistency model trade-offs — drawing on experience from Aurora DSQL and patterns emerging in other distributed PostgreSQL implementations. We'll look at what changes in distributed deployments: data dictionary implementations, ACID guarantee approaches, and where performance implications appear. This is a technical examination of open problems and opportunities for community collaboration, focusing on what worked, what didn't, and what remains unsolved as PostgreSQL is adapted to (and adopted in) cloud-native environments.