HTTP Caching

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.

Section 2 of [HTTP] defines conformance criteria and contains considerations regarding error handling.

This specification uses the Augmented Backus-Naur Form (ABNF) notation of [RFC5234], extended with the notation for case-sensitivity in strings defined in [RFC7405].

It also uses a list extension, defined in Section 5.6.1 of [HTTP], that allows for compact definition of comma-separated lists using a "#" operator (similar to how the "*" operator indicates repetition). Appendix A shows the collected grammar with all list operators expanded to standard ABNF notation.

• The Connection header field and fields whose names are listed in it are required by Section 7.6.1 of [HTTP] to be removed before forwarding the message. This MAY be implemented by doing so before storage.
• Likewise, some fields' semantics require them to be removed before forwarding the message, and this MAY be implemented by doing so before storage; see Section 7.6.1 of [HTTP] for some examples.
• The no-cache (Section 5.2.2.4) and private (Section 5.2.2.7) cache directives can have arguments that prevent storage of header fields by all caches and shared caches, respectively.
• Header fields that are specific to the proxy that a cache uses when forwarding a request MUST NOT be stored, unless the cache incorporates the identity of the proxy into the cache key. Effectively, this is limited to Proxy-Authenticate (Section 11.7.1 of [HTTP]), Proxy-Authentication-Info (Section 11.7.3 of [HTTP]), and Proxy-Authorization (Section 11.7.2 of [HTTP]).

Caches MAY either store trailer fields separate from header fields or discard them. Caches MUST NOT combine trailer fields with header fields.

Caches are required to update a stored response's header fields from another (typically newer) response in several situations; for example, see Sections 3.4, 4.3.4, and 4.3.5.

• Header fields excepted from storage in Section 3.1,
• Header fields that the cache's stored response depends upon, as described below,
• Header fields that are automatically processed and removed by the recipient, as described below, and
• The Content-Length header field.

In some cases, caches (especially in user agents) store the results of processing the received response, rather than the response itself, and updating header fields that affect that processing can result in inconsistent behavior and security issues. Caches in this situation MAY omit these header fields from updating stored responses on an exceptional basis but SHOULD limit such omission to those fields necessary to assure integrity of the stored response.

For example, a browser might decode the content coding of a response while it is being received, creating a disconnect between the data it has stored and the response's original metadata. Updating that stored metadata with a different Content-Encoding header field would be problematic. Likewise, a browser might store a post-parse HTML tree rather than the content received in the response; updating the Content-Type header field would not be workable in this case because any assumptions about the format made in parsing would now be invalid.

Furthermore, some fields are automatically processed and removed by the HTTP implementation, such as the Content-Range header field. Implementations MAY automatically omit such header fields from updates, even when the processing does not actually occur.

Note that the Content-* prefix is not a signal that a header field is omitted from update; it is a convention for MIME header fields, not HTTP.

If the request method is GET, the response status code is 200 (OK), and the entire response header section has been received, a cache MAY store a response that is not complete (Section 6.1 of [HTTP]) provided that the stored response is recorded as being incomplete. Likewise, a 206 (Partial Content) response MAY be stored as if it were an incomplete 200 (OK) response. However, a cache MUST NOT store incomplete or partial-content responses if it does not support the Range and Content-Range header fields or if it does not understand the range units used in those fields.

A cache MAY complete a stored incomplete response by making a subsequent range request (Section 14.2 of [HTTP]) and combining the successful response with the stored response, as defined in Section 3.4. A cache MUST NOT use an incomplete response to answer requests unless the response has been made complete, or the request is partial and specifies a range wholly within the incomplete response. A cache MUST NOT send a partial response to a client without explicitly marking it using the 206 (Partial Content) status code.

A response might transfer only a partial representation if the connection closed prematurely or if the request used one or more Range specifiers (Section 14.2 of [HTTP]). After several such transfers, a cache might have received several ranges of the same representation. A cache MAY combine these ranges into a single stored response, and reuse that response to satisfy later requests, if they all share the same strong validator and the cache complies with the client requirements in Section 15.3.7.3 of [HTTP].

When combining the new response with one or more stored responses, a cache MUST update the stored response header fields using the header fields provided in the new response, as per Section 3.2.

A shared cache MUST NOT use a cached response to a request with an Authorization header field (Section 11.6.2 of [HTTP]) to satisfy any subsequent request unless the response contains a Cache-Control field with a response directive (Section 5.2.2) that allows it to be stored by a shared cache, and the cache conforms to the requirements of that directive for that response.

In this specification, the following response directives have such an effect: must-revalidate (Section 5.2.2.2), public (Section 5.2.2.9), and s-maxage (Section 5.2.2.10).

When a cache receives a request that can be satisfied by a stored response and that stored response contains a Vary header field (Section 12.5.5 of [HTTP]), the cache MUST NOT use that stored response without revalidation unless all the presented request header fields nominated by that Vary field value match those fields in the original request (i.e., the request that caused the cached response to be stored).

• adding or removing whitespace, where allowed in the header field's syntax
• combining multiple header field lines with the same field name (see Section 5.2 of [HTTP])
• normalizing both header field values in a way that is known to have identical semantics, according to the header field's specification (e.g., reordering field values when order is not significant; case-normalization, where values are defined to be case-insensitive)

If (after any normalization that might take place) a header field is absent from a request, it can only match another request if it is also absent there.

A stored response with a Vary header field value containing a member "*" always fails to match.

If multiple stored responses match, the cache will need to choose one to use. When a nominated request header field has a known mechanism for ranking preference (e.g., qvalues on Accept and similar request header fields), that mechanism MAY be used to choose a preferred response. If such a mechanism is not available, or leads to equally preferred responses, the most recent response (as determined by the Date header field) is chosen, as per Section 4.

Some resources mistakenly omit the Vary header field from their default response (i.e., the one sent when the request does not express any preferences), with the effect of choosing it for subsequent requests to that resource even when more preferable responses are available. When a cache has multiple stored responses for a target URI and one or more omits the Vary header field, the cache SHOULD choose the most recent (see Section 4.2.3) stored response with a valid Vary field value.

If no stored response matches, the cache cannot satisfy the presented request. Typically, the request is forwarded to the origin server, potentially with preconditions added to describe what responses the cache has already stored (Section 4.3).

A fresh response is one whose age has not yet exceeded its freshness lifetime. Conversely, a stale response is one where it has.

A response's freshness lifetime is the length of time between its generation by the origin server and its expiration time. An explicit expiration time is the time at which the origin server intends that a stored response can no longer be used by a cache without further validation, whereas a heuristic expiration time is assigned by a cache when no explicit expiration time is available.

A response's age is the time that has passed since it was generated by, or successfully validated with, the origin server.

When a response is fresh, it can be used to satisfy subsequent requests without contacting the origin server, thereby improving efficiency.

The primary mechanism for determining freshness is for an origin server to provide an explicit expiration time in the future, using either the Expires header field (Section 5.3) or the max-age response directive (Section 5.2.2.1). Generally, origin servers will assign future explicit expiration times to responses in the belief that the representation is not likely to change in a semantically significant way before the expiration time is reached.

If an origin server wishes to force a cache to validate every request, it can assign an explicit expiration time in the past to indicate that the response is already stale. Compliant caches will normally validate a stale cached response before reusing it for subsequent requests (see Section 4.2.4).

Since origin servers do not always provide explicit expiration times, caches are also allowed to use a heuristic to determine an expiration time under certain circumstances (see Section 4.2.2).

   response_is_fresh = (freshness_lifetime > current_age)

freshness_lifetime is defined in Section 4.2.1; current_age is defined in Section 4.2.3.

Clients can send the max-age or min-fresh request directives (Section 5.2.1) to suggest limits on the freshness calculations for the corresponding response. However, caches are not required to honor them.

• Although all date formats are specified to be case-sensitive, a cache recipient SHOULD match the field value case-insensitively.
• If a cache recipient's internal implementation of time has less resolution than the value of an HTTP-date, the recipient MUST internally represent a parsed Expires date as the nearest time equal to or earlier than the received value.
• A cache recipient MUST NOT allow local time zones to influence the calculation or comparison of an age or expiration time.
• A cache recipient SHOULD consider a date with a zone abbreviation other than "GMT" to be invalid for calculating expiration.

Note that freshness applies only to cache operation; it cannot be used to force a user agent to refresh its display or reload a resource. See Section 6 for an explanation of the difference between caches and history mechanisms.

When a cache has one or more stored responses for a requested URI, but cannot serve any of them (e.g., because they are not fresh, or one cannot be chosen; see Section 4.1), it can use the conditional request mechanism (Section 13 of [HTTP]) in the forwarded request to give the next inbound server an opportunity to choose a valid stored response to use, updating the stored metadata in the process, or to replace the stored response(s) with a new response. This process is known as validating or revalidating the stored response.

Because unsafe request methods (Section 9.2.1 of [HTTP]) such as PUT, POST, or DELETE have the potential for changing state on the origin server, intervening caches are required to invalidate stored responses to keep their contents up to date.

A cache MUST invalidate the target URI (Section 7.1 of [HTTP]) when it receives a non-error status code in response to an unsafe request method (including methods whose safety is unknown).

A cache MAY invalidate other URIs when it receives a non-error status code in response to an unsafe request method (including methods whose safety is unknown). In particular, the URI(s) in the Location and Content-Location response header fields (if present) are candidates for invalidation; other URIs might be discovered through mechanisms not specified in this document. However, a cache MUST NOT trigger an invalidation under these conditions if the origin (Section 4.3.1 of [HTTP]) of the URI to be invalidated differs from that of the target URI (Section 7.1 of [HTTP]). This helps prevent denial-of-service attacks.

Invalidate means that the cache will either remove all stored responses whose target URI matches the given URI or mark them as "invalid" and in need of a mandatory validation before they can be sent in response to a subsequent request.

A "non-error response" is one with a 2xx (Successful) or 3xx (Redirection) status code.

Note that this does not guarantee that all appropriate responses are invalidated globally; a state-changing request would only invalidate responses in the caches it travels through.

The "Age" response header field conveys the sender's estimate of the time since the response was generated or successfully validated at the origin server. Age values are calculated as specified in Section 4.2.3.

  Age = delta-seconds

The Age field value is a non-negative integer, representing time in seconds (see Section 1.2.2).

Although it is defined as a singleton header field, a cache encountering a message with a list-based Age field value SHOULD use the first member of the field value, discarding subsequent ones.

If the field value (after discarding additional members, as per above) is invalid (e.g., it contains something other than a non-negative integer), a cache SHOULD ignore the field.

The presence of an Age header field implies that the response was not generated or validated by the origin server for this request. However, lack of an Age header field does not imply the origin was contacted.

The "Cache-Control" header field is used to list directives for caches along the request/response chain. Cache directives are unidirectional, in that the presence of a directive in a request does not imply that the same directive is present or copied in the response.

See Section 5.2.3 for information about how Cache-Control directives defined elsewhere are handled.

A proxy, whether or not it implements a cache, MUST pass cache directives through in forwarded messages, regardless of their significance to that application, since the directives might apply to all recipients along the request/response chain. It is not possible to target a directive to a specific cache.

Cache directives are identified by a token, to be compared case-insensitively, and have an optional argument that can use both token and quoted-string syntax. For the directives defined below that define arguments, recipients ought to accept both forms, even if a specific form is required for generation.

  Cache-Control   = #cache-directive

  cache-directive = token [ "=" ( token / quoted-string ) ]

For the cache directives defined below, no argument is defined (nor allowed) unless stated otherwise.

The "Expires" response header field gives the date/time after which the response is considered stale. See Section 4.2 for further discussion of the freshness model.

The presence of an Expires header field does not imply that the original resource will change or cease to exist at, before, or after that time.

The Expires field value is an HTTP-date timestamp, as defined in Section 5.6.7 of [HTTP]. See also Section 4.2 for parsing requirements specific to caches.

  Expires = HTTP-date

For example

Expires: Thu, 01 Dec 1994 16:00:00 GMT

A cache recipient MUST interpret invalid date formats, especially the value "0", as representing a time in the past (i.e., "already expired").

If a response includes a Cache-Control header field with the max-age directive (Section 5.2.2.1), a recipient MUST ignore the Expires header field. Likewise, if a response includes the s-maxage directive (Section 5.2.2.10), a shared cache recipient MUST ignore the Expires header field. In both these cases, the value in Expires is only intended for recipients that have not yet implemented the Cache-Control header field.

An origin server without a clock (Section 5.6.7 of [HTTP]) MUST NOT generate an Expires header field unless its value represents a fixed time in the past (always expired) or its value has been associated with the resource by a system with a clock.

Historically, HTTP required the Expires field value to be no more than a year in the future. While longer freshness lifetimes are no longer prohibited, extremely large values have been demonstrated to cause problems (e.g., clock overflows due to use of 32-bit integers for time values), and many caches will evict a response far sooner than that.

The "Pragma" request header field was defined for HTTP/1.0 caches, so that clients could specify a "no-cache" request (as Cache-Control was not defined until HTTP/1.1).

However, support for Cache-Control is now widespread. As a result, this specification deprecates Pragma.

The "Warning" header field was used to carry additional information about the status or transformation of a message that might not be reflected in the status code. This specification obsoletes it, as it is not widely generated or surfaced to users. The information it carried can be gleaned from examining other header fields, such as Age.

Storing malicious content in a cache can extend the reach of an attacker to affect multiple users. Such "cache poisoning" attacks happen when an attacker uses implementation flaws, elevated privileges, or other techniques to insert a response into a cache. This is especially effective when shared caches are used to distribute malicious content to many clients.

One common attack vector for cache poisoning is to exploit differences in message parsing on proxies and in user agents; see Section 6.3 of [HTTP/1.1] for the relevant requirements regarding HTTP/1.1.

Because one of the primary uses of a cache is to optimize performance, its use can "leak" information about which resources have been previously requested.

For example, if a user visits a site and their browser caches some of its responses and then navigates to a second site, that site can attempt to load responses it knows exist on the first site. If they load quickly, it can be assumed that the user has visited that site, or even a specific page on it.

Such "timing attacks" can be mitigated by adding more information to the cache key, such as the identity of the referring site (to prevent the attack described above). This is sometimes called "double keying".

Implementation and deployment flaws (often led to by the misunderstanding of cache operation) might lead to the caching of sensitive information (e.g., authentication credentials) that is thought to be private, exposing it to unauthorized parties.

Note that the Set-Cookie response header field [COOKIE] does not inhibit caching; a cacheable response with a Set-Cookie header field can be (and often is) used to satisfy subsequent requests to caches. Servers that wish to control caching of these responses are encouraged to emit appropriate Cache-Control response header fields.

IANA has updated the "Hypertext Transfer Protocol (HTTP) Cache Directive Registry" at https://www.iana.org/assignments/http-cache-directives with the registration procedure per Section 5.2.4 and the cache directive names summarized in the table below.

The Warning header field (and the warn codes that it uses) has been obsoleted for HTTP per [RFC9111].

9.1. Normative References

[HTTP]

Fielding, R., Ed., Nottingham, M., Ed., and J. Reschke, Ed., “HTTP Semantics”, STD 97, RFC 9110, DOI 10.17487/RFC9110, June 2022.

[RFC2119]

Bradner, S., “Key words for use in RFCs to Indicate Requirement Levels”, BCP 14, RFC 2119, DOI 10.17487/RFC2119, March 1997, <https://www.rfc-editor.org/info/rfc2119>.

[RFC5234]

Crocker, D., Ed. and P. Overell, “Augmented BNF for Syntax Specifications: ABNF”, STD 68, RFC 5234, DOI 10.17487/RFC5234, January 2008, <https://www.rfc-editor.org/info/rfc5234>.

[RFC7405]

Kyzivat, P., “Case-Sensitive String Support in ABNF”, RFC 7405, DOI 10.17487/RFC7405, December 2014, <https://www.rfc-editor.org/info/rfc7405>.

[RFC8174]

Leiba, B., “Ambiguity of Uppercase vs Lowercase in RFC 2119 Key Words”, BCP 14, RFC 8174, DOI 10.17487/RFC8174, May 2017, <https://www.rfc-editor.org/info/rfc8174>.

9.2. Informative References

[COOKIE]

Barth, A., “HTTP State Management Mechanism”, RFC 6265, DOI 10.17487/RFC6265, April 2011, <https://www.rfc-editor.org/info/rfc6265>.

[HTTP/1.1]

Fielding, R., Ed., Nottingham, M., Ed., and J. Reschke, Ed., “HTTP/1.1”, STD 99, RFC 9112, DOI 10.17487/RFC9112, June 2022.

[RFC2616]

Fielding, R., Gettys, J., Mogul, J., Frystyk, H., Masinter, L., Leach, P., and T. Berners-Lee, “Hypertext Transfer Protocol -- HTTP/1.1”, RFC 2616, DOI 10.17487/RFC2616, June 1999, <https://www.rfc-editor.org/info/rfc2616>.

[RFC5861]

Nottingham, M., “HTTP Cache-Control Extensions for Stale Content”, RFC 5861, DOI 10.17487/RFC5861, May 2010, <https://www.rfc-editor.org/info/rfc5861>.

[RFC7234]

Fielding, R., Ed., Nottingham, M., Ed., and J. Reschke, Ed., “Hypertext Transfer Protocol (HTTP/1.1): Caching”, RFC 7234, DOI 10.17487/RFC7234, June 2014, <https://www.rfc-editor.org/info/rfc7234>.

[RFC8126]

Cotton, M., Leiba, B., and T. Narten, “Guidelines for Writing an IANA Considerations Section in RFCs”, BCP 26, RFC 8126, DOI 10.17487/RFC8126, June 2017, <https://www.rfc-editor.org/info/rfc8126>.