DynamoDB, Neptune and OpenSearch: A Unified Approach to Collaborative Editing and Search #JSConfAM25
summary
Describes a production architecture combining DynamoDB for high-throughput key-value reads/writes, Amazon Neptune for graph traversals to compute permission inheritance across deep trees, and OpenSearch for listing/indexing and retrieval‑augmented generation (RAG) embeddings. Presents a migration pipeline and iterative POC that reduced a planned nine‑month rework to about a month and achieved a zero‑downtime migration supporting roughly 500k meetings/day and over a billion rows.
AI-written summary · about
chapters (11)
- 0:00Unified data systems intro
- 2:13Collaborative editing challenges
- 7:35Database trade-offs
- 10:02CAP theorem and design
- 12:12Neptune intro
- 15:28High-throughput design
- 23:51Choosing databases
- 27:43Multi-db architecture
- 33:29Scaling in production
- 36:23Team timeline
- 39:15Evolving systems simply
related talks
37mBuilding a Search Engine to Improve User Engagement
43mFuture-proofing your applications with AWS databases
09mIntroducing Opensearch in existing multisource enterprise App
51mImmutable data stores and CORS for safety, flexibility and profit
35mInnovating at Scale: Lessons from Pegasystems' OpenSearch Journey
09mUnconference: OpenSearch is not a database!
39mHarnessing Amazon’s DynamoDB with Spring
15mWhen is it time to modernize your AWS database platform?