Learn Live: Model data for Azure Cosmos DB for PostgreSQL
summary
Azure Cosmos DB for PostgreSQL groups Postgres servers into a coordinator plus worker nodes to deliver a scalable distributed relational database; rows are sharded by the hash of a chosen distribution column (default 32 shards). Three table types exist: local (coordinator-only), reference (replicated to every worker) and distributed (sharded across workers); reference tables are created with CREATE REFERENCE TABLE and distributed tables with CREATE DISTRIBUTED TABLE or ALTER; distribution column must be part of primary/unique keys and foreign keys to enforce constraints; time partitioning and functions like CREATE_TIME_PARTITIONS / DROP_ALL_TIME_PARTITIONS manage retention.
AI-written summary · about
chapters (11)
- 0:00Start
- 2:51Scenario: Scaling Woodgrove Bank's contactless payment app
- 6:19Distributing data across multiple servers
- 19:42Reference tables.
- 22:23Reference table versus distributed table
- 28:42Trouble with implicit colocation
- 30:03Foreign key implications in a distributed environment
- 32:33Compare from a single ERD to a distributed ERD
- 37:33Partition time series data
- 38:17What is time-partitioning?
- 41:27Time partitions in a distributed environment
related talks
13mHow to do Data Modeling for Distributed Postgres in Azure Cosmos DB
09mMulti-Region Replication & DR for Distributed Postgres in Azure Cosmos DB
52mBuilding Distributed OLTP with Azure Cosmos DB for PostgreSQL
27mMulti-tenant Saas apps made simple on Azure Cosmos DB for PostgreSQL
44mMulti-tenant SaaS apps made simple on Azure Cosmos DB for PostgreSQL
06mKeynote: Azure Cosmos DB for PostgreSQL
08mFoundations of Distributed PostgreSQL with Citus
18mBuilding Performant Data Architecture for Multi-Tenant SaaS Apps