Google IO extended 2017 - Caching Entities using Actors in Distributed Systems
0 likes253 views
The document discusses the implementation of caching entities using actors in a distributed system, highlighting concepts like the actor model, sharding entities, and fault tolerance. It explains the importance of Google's Protocol Buffers for serializing structured data and presents a demo of sharding counters. References to further resources and tools related to actor supervision and best practices are also included.
Google IO extended 2017 - Caching Entities using Actors in Distributed Systems
- 1. Caching Entities using Actors in a Distributed System Aldrin Seychell Technical Architect @aseychell
- 6. Actors = Threads Lightweight Processes
- 11. Akka ● Actor Model on the JVM ● Abstract scheduling of actors ● Abstract details of message-passing ● Location transparency through ActorRef ● Elasticity of Java applications ● Remote Actors ● Fault tolerance with exceptions @aseychell
- 14. Virtual Card 1 Virtual Card 2 Virtual Card 3 Router Take €10 VC1 Take €25 VC1 Take €15 VC2 Put €50 VC3
- 15. Entity exists on only ONE node
- 17. Shards place an upper-bound on the load-balancing problem
- 20. Tuning Parameters - Number of shards - Sharding Function - Load Balancing - Recreating Actors on new nodes - Remember-Entities? - Resolving Split-Brain Situations @aseychell
- 21. Google Protocol Buffers (aka Protobuf) “Protocol buffers are a language-neutral, platform-neutral, extensible mechanism for serializing structured data.” https://developers.google.com/protocol-buffers/ message Counter { int64 id = 1; int64 value = 2; } message CounterOp { enum Type { INCREMENT = 0; DECREMENT = 1; } int64 id = 1; Type type = 2; } 10101111010101011... @aseychell
- 22. Demo - Sharding Counters https://github.com/aseychell/demo-akka-clustersharding @aseychell
- 23. Counter 1 Counter 2 Counter 3 Router INCREMENT 1 INCREMENT 1 DECREMENT 2 GET 1 GET 2 GET 3