Productizing AsyncAPI for Data Replication
summary
Oracle integrates AsyncAPI into GoldenGate to publish millisecond-level DML and DDL streams from hundreds of databases to brokers or lakehouses, supporting exact-once and at-least/at-most delivery and shipping JSON by default; the platform already runs in ~10,000 customers and a ~$1B business, provides YAML channel docs and generated client stubs, maintains a ledger for historical replay, and ships a secure websocket prototype with plans for parquet/delta/iceberg and AsyncAPI studio embedding.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 1:08Change data capture and replication
- 3:35GoldenGate overview and history
- 6:25GoldenGate use cases
- 8:12Kafka and transactional consistency
- 9:59AsyncAPI data streams in GoldenGate
- 13:53Data binding patterns and trade-offs
- 17:21Consumer-friendly AsyncAPI
- 18:03Legacy app integration with AsyncAPI
- 19:50AsyncAPI in GoldenGate end-to-end
- 21:53QoS and filtering in AsyncAPI
- 22:59WebSockets and AsyncAPI studio
- 24:03AsyncAPI message schemas
- 25:48Consumer start positions
- 26:52CloudEvents integration
- 27:55Business object semantics
- 28:57Q&A
related talks
30mOCI GoldenGate: Elastic, real-time data integration
27mDistributed AI Processing with GoldenGate 23ai
30mReal-Time Data Ingestions for Data Fabric and AI Applications
13mReal-time Data for AI and RAG with GoldenGate 23ai
13mOracle Globally Distributed Database
42mAsyncAPI Recipes for EDA Gourmet
31mSuperfast event stream processing with a converged database
23mBroadening the API Landscape: AsyncAPI and CloudEvents in the Enterprise