SCRU64: Sortable, Clock-based, Realm-specifically Unique identifier

SCRU64 ID offers compact, time-ordered unique identifiers generated by distributed nodes. SCRU64 has the following features:

• 63-bit non-negative integer storable as signed/unsigned 64-bit integer
• Sortable by generation time (as integer and as text)
• 12-digit case-insensitive textual representation (Base36)
• ~38-bit Unix epoch-based timestamp that ensures useful life until year 4261
• Variable-length node/machine ID and counter fields that share 24 bits

// pass node ID through environment variable
// (e.g., SCRU64_NODE_SPEC=42/8 command ...)

// generate a new identifier object
let x = scru64::new_sync();
println!("{}", x); // e.g., "0u2r85hm2pt3"
println!("{}", x.to_u64()); // as a 64-bit unsigned integer

// generate a textual representation directly
println!("{}", scru64::new_string_sync()); // e.g., "0u2r85hm2pt4"

See SCRU64 Specification for details.

SCRU64's uniqueness is realm-specific, i.e., dependent on the centralized assignment of node ID to each generator. If you need decentralized, globally unique time-ordered identifiers, consider SCRU128.

Crate features

Default features:

• std integrates the library with, among others, the system clock to draw current timestamps. Without std, this crate provides limited functionality available under no_std environments.
• global_gen (implies std) enables the new(), new_string(), new_sync(), and new_string_sync() primary entry point functions as well as the process-wide global generator under the hood.

Optional features:

• serde enables serialization/deserialization via serde.

License

Licensed under the Apache License, Version 2.0.

See also

• docs.rs/scru64