A Rust implementation of the Snowflake ID generator, which produces unique 64-bit IDs. This implementation ensures thread safety and high performance, suitable for distributed systems.
Each Snowflake ID consists of three parts:
- Timestamp: 41 bits
- Node ID: 10 bits
- Sequence Number: 12 bits
- Timestamp (41 bits): Milliseconds since a custom epoch.
- Node ID (10 bits): Unique identifier for the node generating the IDs (0-1023).
- Sequence Number (12 bits): Incremental counter within the same millisecond.
- Thread-safe: Can be used safely across multiple threads.
- Custom Epoch: Allows setting a custom epoch.
- High Performance: Generates a large number of IDs per second.
[dependencies]
snowflake-rs-impl="*"
use snowflake::Snowflake;
fn main() {
// Create a new Snowflake instance with node ID 1 and default epoch
let snowflake = Snowflake::new(1, None).unwrap();
// Generate a new ID
let id = snowflake.generate().unwrap();
println!("Generated ID: {}", id);
}
use snowflake::Snowflake;
fn main() {
// Custom epoch (2023-01-01T00:00:00Z in milliseconds since Unix epoch)
let custom_epoch = 1672531200000;
let snowflake = Snowflake::new(1, Some(custom_epoch)).unwrap();
// Generate a new ID
let id = snowflake.generate().unwrap();
println!("Generated ID: {}", id);
}
This library includes tests to verify the correct functionality of the Snowflake ID generator.
cargo test
- Single-threaded ID generation: Measures IDs generated per second in a single thread.
- Multi-threaded ID generation: Measures IDs generated per second using multiple threads.
- Uniqueness: Ensures that all generated IDs are unique.
- Node ID range validation: Verifies that creating a Snowflake with an invalid node ID returns an error.
cargo bench
Approximately 4,100,000 IDs per second