Databaseless Data Processing - High-Performance for Cloud-Native Apps and AI

on YouTube ↗53m

" summary

EclipseStore replaces Hibernate/JPA by mirroring real Java object graphs in memory and appending binary snapshots to object storage (AWS S3 or files) so queries run with Java Streams in microseconds rather than milliseconds; the demo claims up to 1000x faster and an estimated 96% cost savings by avoiding cloud databases. The implementation uses EclipseStore with a custom binary engine Eclipse Serializer, supports lazy loading, JVM, native image and Kubernetes deployments, and reports multi‑year production use.

AI-written summary · about

" chapters (12)

1. 0:00Introduction and session change
2. 2:16Performance demo comparison
3. 8:01Problems with ORM and databases
4. 13:48Why so many databases exist
5. 22:46Cloud costs and complexity
6. 27:43Java’s advantages for data processing
7. 31:15Eclipse Store introduction
8. 35:55Framework history and adoption
9. 40:13How Eclipse Store works
10. 45:32Eclipse Serializer core
11. 49:52Lazy loading and object graph
12. 51:42Resources and next steps

" related talks

• 53mHigh Performance Java Cloud Native AppsMarkus Kett · GeeCON 2023
• 55mDatabaseless High-Performance Persistence with EclipseStore...Markus Kett · GeeCON Prague 2023
• 43mEclipseStore - Ultra-Fast Java Cloud-Native Persistence for Microservices and ServerlessMarkus Kett · EclipseCon 2023
• 22mEclipseStore – The new, ultra-fast Java persistence for the cloudMarkus Kett · GeeCON 2023
• 41mThe Evolution of Data Processing with JavaMarkus Kett · JCON Europe 2023
• 52mGet Started Coding with EclipseStore...Markus Kett, Christian Kuemmel · GeeCON Prague 2023
• 53mJava In-Memory Database-Apps with EclipseStoreFlorian Habermann · JCON WORLD ONLINE 2023
• 32mJava's Ignored Potential: Supercharge Database Apps with Java In-Memory ProcessMarkus Kett · JCON EUROPE 2025