"Actors and Factories in Rust"
summary
Meta teams add FB Thrift support and build a C++ to Rust Tokio + Folly shim so C++ stays on sockets and Rust handles deserialization and business logic. They implement the Raport actor framework with factories that keep a pool of warm workers, supervisor restarts, adjustable queue depth and worker count, load‑shedding when queues fill, and graceful request draining. The result is bounded concurrency that prevents Tokio from spawning unbounded tasks and solves overloads in production.
AI-written summary · about
chapters (10)
- 0:00Intro
- 1:22How Rust came to Meta
- 9:40Sean Lawlor
- 12:17What are Factories
- 15:45Factories in Thrift
- 16:54Concurrency and Queue Depth
- 17:47Factory concurrency
- 18:33Request draining
- 19:09Factory life cycle
- 20:22Conclusion
related talks
56mThe RUST Language for High‑Load Network Services — Quick Start
15mTowards Safe HPC: Productivity and Performance via Rust Interfaces for a Distribu...
27mThe State of Rust Web Frameworks: Pros, Cons & Use Cases
35mBuilding Actor Systems in Rust
22mUsing Rust in Metal Fabrication
16mEmpowering Fearless Data Engineering
29mLessons Learned: Porting a Streaming Data Pipeline from Scala to Rust.
17mRust in Production