Which HTTP protocol header controls the cache behavior of retrieved objects?

The Cache-Control header is essential for managing the caching behavior of resources retrieved via HTTP. It instructs both browsers and intermediate caches on how to handle the stored data. By specifying various directives such as no-cache, no-store, max-age, or must-revalidate, the Cache-Control header allows developers and system administrators to control aspects like how long content should be cached, whether it can be stored, and when it should be validated.

For instance, using max-age=3600 would indicate that the resource can be cached for up to one hour, while no-store would dictate that the resource should never be stored in cache at all. This level of control is crucial for optimizing web performance and ensuring that users receive the most up-to-date content without unnecessary delay or resource usage.

Other headers present in the choices have different roles; for example, the Connection header primarily manages keep-alive connections and controls routing of network traffic, the Expect header is used to express expectations regarding the server's handling of the request, and the User-Agent header provides information about the client's browser and operating system. These headers do not pertain to caching and thus do not influence how retrieved objects are cached or stored.