Distributed Object Graphs with MicroStream
summary
MicroStream replaces JPA/Hibernate by persisting live Java object graphs as binary snapshots (files or cloud blob stores) and restores them into memory for microsecond queries, with a demo showing a Helidon+MicroStream service reading a 250 GB bookstore dataset up to 100–1000x faster than a PostgreSQL + JPA stack. A cluster uses a single writer node that publishes stores to Kafka and readers consume and merge updates; core is open source, cluster tooling is closed, and customers report up to 90% CPU savings and stores appended binaries with Legacy type mapping for class evolution.
AI-written summary · about
chapters (10)
- 0:00Introduction and welcome
- 1:28Data challenges in modern apps
- 4:41What is Helidon?
- 10:30Performance demo with MicroStream
- 12:16Why traditional databases slow down apps
- 21:15How MicroStream works under the hood
- 26:18Simplified architecture with MicroStream
- 31:45Distributed object graphs with MicroStream
- 37:03Challenges and limitations
- 43:09Q&A and closing
related talks
50mGetting started with MicroStream - High-Performance Java-Native Persistence
15mFaster Data Access with MicroStream
50mUltra fast Java In Memory Database Apps and Microservices with MicroStream
54mEclipse MicroStream - Ultra-Fast Java Cloud-Native Persistence for Microservice
01h 00mEclipse MicroStream - Ultra-fast Java cloud-native persistence for microservices
48mIn-Memory-Computing - The Big Picture
44mJava In-Memory Database-Apps with MicroStream
50mJava In-Memory Database Apps with MicroStream