Azure Cosmos DB Modeling Methodology
summary
Methodology gives concrete rules for migrating a 300+ table SQL Server e‑commerce schema into Cosmos DB SQL API using DDD aggregates and query stats: each aggregate → one container, partition key = anchor id (primary key), non‑root entities become separate JSON document types in the same container, value objects either embedded or separate documents depending on cardinality and churn, use a shared dictionary container for lookup aggregates, reference IDs for inter‑aggregate links, and duplicate hot query data (orders into accounts) to avoid cross‑partition scans.
AI-written summary · about
chapters (11)
- 0:00Intro
- 0:58Importance of modeling in NoSQL
- 2:02Case study: legacy e-commerce platform migration challenge Scaling issues related to support bigger number of Relational Tables simultaneous users from around the globe
- 4:41Methodology inspiration
- 7:04The methodology: executive summary
- 11:51Case study: application flow
- 15:04Case study: DDD aggregates
- 16:23Data realities & queries statistics
- 17:49Mapping rules. Part 3
- 20:22Case study: logical data model
- 23:32Case study: physical optimization
related talks
01h 00mAzure Cosmos DB : A brief Note for Developers
01h 03mData Modeling and Partitioning in Azure Cosmos DB
01h 01mFrom SQL to Azure Cosmos DB
01h 02mFrom SQL to Azure Cosmos DB
13mHow to do Data Modeling for Distributed Postgres in Azure Cosmos DB
50mAzure Cosmos DB - Technical Deep Dive
48mJSON Data Modeling
55mSQL Server vs Azure Cosmos DB