DynamoDB in an EDA World
summary
DynamoDB solves the producer dual-write problem by turning writes into change events via DynamoDB Streams (Lambda reads are free), avoiding transactional outbox complexity; teams publish cleaned events to an event bus rather than letting many consumers read the stream directly (Dynamo limits concurrent stream consumers to two). The service targets consistent performance at scale—hundreds of customers at 0.5M req/s, billions/hr, tables >200 TB—offers on-demand billing for 10–20x daily variability, and enforces idempotency with condition expressions and transactions (10-minute idempotency token window).
AI-written summary · about
chapters (8)
- 0:00Intro
- 1:02What's unique about DynamoDB?
- 5:28Agenda
- 6:59What are event-driven applications?
- 15:04DynamoDB in an EDA producer
- 26:39DynamoDB in an EDA consumer
- 35:22Takeaways
- 36:33Outro
related talks
30mBuild for Scale, Performance With Feature-packed Amazon DynamoDB
47mEvent-driven Serverless Architecture with DynamoDB
20mAmazon DynamoDB: A Scalable, Predictably Performant, and Fully Managed NoSQL...
33mDynamoDB Essentials and Integration for Java Developers
33mBuild for scale with Amazon DynamoDB developer features
01h 08mTransactions and Scalability in Cloud Databases—Can’t We Have Both?
24mDistributed Transactions at Scale in Amazon DynamoDB
43mDYNAMO: the paper that changed the database world