Distributed Eventsourcing with Akka
summary
Explains how to build distributed, stateful identities by combining event sourcing with Akka Persistence and Cluster Sharding: persistent actors append events to a pluggable journal (Cassandra, JDBC, etc.), use snapshotting to avoid replaying millions or billions of events, and auto-activate/passivate entities so a 30M-user scale can be sharded across nodes. Shows command → event flow, at-least-once delivery, and version-based conflict checks; proves resilience and elasticity but warns about availability vs consistency tradeoffs and the need for domain-driven snapshotting and tuning.
AI-written summary · about
chapters (10)
- 0:00Introduction and speaker background
- 1:07Why distributed systems matter
- 10:05CRUD equals pain in distributed systems
- 14:36What is event sourcing
- 19:37Commands vs events explained
- 27:26Akka Persistence architecture
- 35:19Identity versus state in distributed systems
- 38:31Cluster sharding for scalability
- 42:28Consistency in distributed systems
- 48:23Q&A and closing remarks
related talks
29mEVENT SOURCING IN AKKA PERSISTENCE
54mHow Events Are Reshaping Modern Systems
17mAkka Persistence
43mFunctional and Reactive Event Sourced Applications
49mDDD, Event Sourcing and Actors
38mMulti-Cloud Replication: From Chaos to Control
51mUp up and Out: Scaling Software with Akka
47mData in Motion: Streaming Static Data Efficiently in Akka Persistence