NSURLCache

NSURLCache is a caching mechanism provided by the Foundation framework that stores responses to URL load requests. It uses a two-tier storage system, combining memory cache for fast access with a disk cache for persistence, governed by configurable memory and disk capacities. Cached responses help reduce network traffic and speed up repeated requests by serving data locally when appropriate.

The cache stores data as NSCachedURLResponse objects, which pair an NSURLResponse with the corresponding data. Core properties include memoryCapacity, diskCapacity, and diskPath, which determine how much data can be kept in memory and on disk and where the on-disk cache is stored. An instance of NSURLCache can be accessed via the shared URL cache, and it can be configured by adjusting its capacities or by creating a custom cache with an init method such as initWithMemoryCapacity:diskCapacity:diskPath:.

Key API methods include cachedResponseForRequest:, storeCachedResponse:forRequest:, removeCachedResponseForRequest:, and removeAllCachedResponses. In Objective-C these are negative- or positive-formed methods on NSURLCache, while in Swift they translate to cachedResponse(for:), storeCachedResponse(_:for:), and related helpers. These methods enable explicit management of cached entries, though the URL loading system generally handles caching automatically based on HTTP and URL protocol rules.

Usage and behavior: The URL loading system consults the URL cache for appropriate requests, respecting cache-control headers and other HTTP caching directives. Developers can assign a URLCache to a URLSessionConfiguration via its urlCache property to enable per-session caching, or rely on the shared cache for application-wide behavior. Cache eviction occurs automatically under memory pressure or when capacity limits are reached, and the disk cache persists across launches.

Limitations: The cache stores only cacheable responses and is not a general-purpose data store. It works with HTTP and other cacheable URL schemes as governed by the underlying protocols and caching policies.