//! Cache backend trait — pluggable caching abstraction.
//!
//! # v2.x — SOTA Abstraction
//!
//! All cache implementations implement this trait. Swap backends
//! without changing any application code.
use Bytes;
use HashMap;
/// Core cache operations. All backends implement this.
///
/// # Performance contract
///
/// - `get()` must return in <5ms p99 for in-memory backends
/// - `put()` is fire-and-forget (no guarantee of durability)
/// - `stats()` is informational, not a consistency guarantee