Schedule - PGConf.EU 2023
PostgreSQL Distributed: Architectures & Best practices
While PostgreSQL is seen as a single server database system, most of the time we actually run it in a distributed set up. You might already be using a hot standby, network-attached storage, read replicas, postgres_fdw, or even a more explicitly distributed set up that uses sharding, Citus, a distributed fork of PostgreSQL, or decoupled storage & compute. This talk will discuss the properties and trade-offs of different distributed PostgreSQL architectures.
Distributed database systems inherently involve trade-offs regarding response time & throughput, scalability, durability, availability, consistency, and operational simplicity. Navigating those trade-offs can be a challenge. We will walk through various examples of typical distributed set ups and contrast different alternatives. Performance and scalability will be especially important themes.
At the end of the talk, you should have a better understanding of different distributed PostgreSQL architectures, the trade-offs involved, when to apply which architecture, and some best practices to follow.