How to build real-world applications with Orleans
summary
Orleans presents grains as virtual actors that hold state, auto activate and deactivate, and persist via declarative providers so objects run across a cluster without manual sharding. Proven at scale in Halo and Azure with tens of millions of users and hundreds of thousands requests/sec. Patterns covered include smart caches (in-memory grain cache removing cache invalidation), event sourcing with append only events and aggregates, registries, streaming and timers, and state machines. Examples show a Skype telemetry service getting 4x throughput by eliminating storage write contention. Cross-cluster, .NET Core and Docker support are underway.
AI-written summary · about
chapters (19)
- 0:00Intro
- 5:16Stateless Life is Sad
- 7:20Gist of Orleans
- 16:58Lifecycle of Grain
- 19:39Pattern: Object Storage Grain
- 20:57Pattern: Registry
- 23:36Traditional Approach to Caching
- 25:06Orleans Approach
- 26:30Types of Caches
- 30:39Pattern: Smart Cache
- 31:41Pattern: Event-Sourcing Smart Cache
- 41:21Pattern: Aggregate - Grain State
- 45:00Composite Pattern: Object Store
- 49:45Pattern: Object Graph - Multiplayer Game
- 50:46Pattern: Object Graph - IoT Control Structure
- 51:57Orleans Streaming
- 54:07Pattern: Streaming Analytics
- 55:03Pattern: State Machine
- 55:31Cross-Cutting Benefits of Orleans
related talks
59mIntroduction to project Orleans for your distributed applications
01h 02mPub crawling in Orleans: exploring the Actor model
37mState of Affairs or Affairs of State
53mOrleans under the hood
46mState or events? Which shall I keep?
53mHow do we build high‑load microservices
17mRapidly Build Distributed Applications using Orchestration
44mGetting started with Orchestration and Choreography in Distributed Applications