m07-concurrency

To share state safely across async tasks in Rust, wrap your data in Arc<Mutex<T>> or Arc<RwLock<T>>. This allows multiple Tokio tasks to access and modify shared data concurrently without causing data races.

Use async for I/O-bound workloads and task orchestration, while threads suit CPU-bound multi-threaded data processing. Choosing the right Rust concurrency model prevents bottlenecks and ensures efficient resource utilization.

Send and Sync errors occur when types cross thread or task boundaries unsafely. Guided concurrency design helps trace down these violations and apply correct synchronization primitives to satisfy Rust's safety guarantees.

Avoid deadlocks in Rust async by never holding Mutex guards across .await points. Use structured concurrency patterns and channels for task communication to prevent locks from blocking the async runtime.

Use Tokio channels like mpsc or broadcast to orchestrate tasks in Rust services. Channels provide thread-safe message passing, decoupling producers and consumers for safe async data flow without shared state.

RwLock is suitable for high-concurrency Rust services with many readers and few writers. It allows multiple concurrent read locks, improving throughput compared to Mutex when write contention is low.