Up up and Out: Scaling Software with Akka
summary
Claims that Akka unifies scaling UP and OUT by making distribution a deployment artifact and exposing Distributed by Design features like location transparency, adaptive automatic load‑balancing, cluster rebalancing, replication and partitioning. Uses actor primitives, routers, supervision, a cluster extension with seed nodes, and a file‑backed durable mailbox. Actors are extremely small 300 bytes per actor so systems can spawn millions of actors. Akka 2 adds cluster routers and a console and is already running in production across telecom, e‑commerce, gaming and finance.
AI-written summary · about
chapters (41)
- 0:00Introduction
- 1:40Production users
- 2:55What is Akka
- 4:25Concurrency
- 5:15Workflow
- 8:30Deployment artifact
- 12:40Actor Model
- 14:10Public API
- 15:55Actors
- 18:10Actor system
- 19:45Actor form hierarchies
- 22:35Actor alias
- 23:15Routers
- 24:20Become
- 25:25Why
- 26:15Finite State Machines
- 27:55spawning empty workers
- 28:30dynamic actors
- 29:00pass
- 29:20failure recovery
- 30:05failure management
- 30:40error handling
- 31:30failure
- 32:05supervision
- 33:00error curve
- 34:15failure mechanism
- 35:35extra system span
- 36:35react failure
- 37:55supervisor strategy
- 38:20callbacks
- 38:45remote access
- 40:05user space
- 40:30cluster
- 42:10masterless
- 42:25clustered routers
- 43:35typeset console
- 44:20conclusion
- 45:35team blog
- 45:55questions
- 46:15API
- 47:20Guarantees
related talks
37mAkka Present and Future
48mThe Zen of Akka
01h 32mFrom Spring + Java to Spring + Akka - A Journey of Discovery
32mAdopting Actors... an epic tale of loss and learning
53mDistributed Eventsourcing with Akka
50mThe Zen of Akka
58mDeath to Latency: Building Reactive, Cloud Native Apps with Akka.NET
46mAkka Streams & Reactive Streams in action