draft-ietf-httpbis-cache-latest.xml

<?xml version="1.0" encoding="utf-8"?>
<?xml-stylesheet type='text/xsl' href='lib/myxml2rfc.xslt'?>
<!DOCTYPE rfc [
  <!ENTITY MAY "<bcp14>MAY</bcp14>">
  <!ENTITY MUST "<bcp14>MUST</bcp14>">
  <!ENTITY MUST-NOT "<bcp14>MUST NOT</bcp14>">
  <!ENTITY NOT-RECOMMENDED "<bcp14>NOT RECOMMENDED</bcp14>">
  <!ENTITY OPTIONAL "<bcp14>OPTIONAL</bcp14>">
  <!ENTITY RECOMMENDED "<bcp14>RECOMMENDED</bcp14>">
  <!ENTITY REQUIRED "<bcp14>REQUIRED</bcp14>">
  <!ENTITY SHALL "<bcp14>SHALL</bcp14>">
  <!ENTITY SHALL-NOT "<bcp14>SHALL NOT</bcp14>">
  <!ENTITY SHOULD "<bcp14>SHOULD</bcp14>">
  <!ENTITY SHOULD-NOT "<bcp14>SHOULD NOT</bcp14>">
  <!ENTITY ID-VERSION "latest">
  <!ENTITY mdash "&#8212;">
  <!ENTITY uuml "&#252;">
  <!ENTITY Note "<x:h xmlns:x='http://purl.org/net/xml2rfc/ext'>Note:</x:h>">
]>
<?rfc compact="yes"?>
<?rfc subcompact="no" ?>
<?rfc linkmailto="no" ?>
<?rfc editing="no" ?>
<?rfc comments="yes"?>
<?rfc inline="yes"?>
<?rfc rfcedstyle="yes"?>
<?rfc-ext allow-markup-in-artwork="yes" ?>
<?rfc-ext map-dfn="none" ?>
<?rfc-ext html-pretty-print="prettyprint https://cdn.rawgit.com/google/code-prettify/master/loader/run_prettify.js"?>
<?rfc-ext include-references-in-index="yes" ?>
<?rfc-ext xml2rfc-backend="202007"?>
<?github-issue-label cache?>

<rfc category="std" docName="draft-ietf-httpbis-cache-&ID-VERSION;" ipr="pre5378Trust200902"
  obsoletes="7234" x:maturity-level="internet" xmlns:x="http://purl.org/net/xml2rfc/ext" xmlns:rdf='http://www.w3.org/1999/02/22-rdf-syntax-ns#'
  version="3" tocDepth="4" sortRefs="true">
<x:feedback template="mailto:ietf-http-wg@w3.org?subject={docname},%20%22{section}%22&amp;body=&lt;{ref}&gt;:"/>
<front>

  <title>HTTP Caching</title>

  <author fullname="Roy T. Fielding" initials="R." surname="Fielding" role="editor">
    <organization>Adobe</organization>
    <address>
      <postal>
        <postalLine>345 Park Ave</postalLine>
        <postalLine>San Jose, CA 95110</postalLine>
        <country>United States of America</country>
      </postal>
      <email>fielding@gbiv.com</email>
      <uri>https://roy.gbiv.com/</uri>
    </address>
  </author>

  <author fullname="Mark Nottingham" initials="M." surname="Nottingham" role="editor">
    <organization>Fastly</organization>
    <address>
       <postal>
         <postalLine>Prahran</postalLine>
         <country>Australia</country>
       </postal>
      <email>mnot@mnot.net</email>
      <uri>https://www.mnot.net/</uri>
    </address>
  </author>

  <author fullname="Julian Reschke" initials="J." surname="Reschke" role="editor">
    <organization abbrev="greenbytes">greenbytes GmbH</organization>
    <address>
      <postal>
        <postalLine>Hafenweg 16</postalLine>
        <postalLine>48155 M&uuml;nster</postalLine>
        <country>Germany</country>
      </postal>
      <email>julian.reschke@greenbytes.de</email>
      <uri>https://greenbytes.de/tech/webdav/</uri>
    </address>
  </author>

  <date/>

  <area>Applications and Real-Time</area>
  <workgroup>HTTP</workgroup>

  <keyword>Hypertext Transfer Protocol</keyword>
  <keyword>HTTP</keyword>
  <keyword>HTTP Caching</keyword>

<abstract>
<t>
   The Hypertext Transfer Protocol (HTTP) is a stateless application-level
   protocol for distributed, collaborative, hypertext information systems.
   This document defines HTTP caches and the associated header fields that
   control cache behavior or indicate cacheable response messages.
</t>
<t>
   This document obsoletes RFC 7234.
</t>
</abstract>

<note title="Editorial Note" removeInRFC="true">
  <t>
    Discussion of this draft takes place on the HTTP working group
    mailing list (ietf-http-wg@w3.org), which is archived at
    <eref target="https://lists.w3.org/Archives/Public/ietf-http-wg/"/>.
  </t>
  <t>
    Working Group information can be found at <eref target="https://httpwg.org/"/>;
    source code and issues list for this draft can be found at
    <eref target="https://github.com/httpwg/http-core"/>.
  </t>
  <t>
    The changes in this draft are summarized in <xref target="changes.since.19"/>.
  </t>
</note>

   </front>
   <middle>

<section title="Introduction" anchor="caching">
<t>
   The Hypertext Transfer Protocol (HTTP) is a stateless application-level
   request/response protocol that uses extensible semantics and
   self-descriptive messages for flexible interaction with network-based
   hypertext information systems. It is typically used for distributed information systems, where
   the use of response caches can improve performance. This document
   defines aspects of HTTP related to caching and reusing response
   messages.
</t>

<iref item="cache" />
<t>
   An HTTP <x:dfn>cache</x:dfn> is a local store of response messages and the
   subsystem that controls storage, retrieval, and deletion of messages in it.
   A cache stores cacheable responses to reduce the response time and
   network bandwidth consumption on future equivalent requests. Any client or
   server &MAY; use a cache, though not when acting as a tunnel (<xref target="intermediaries"/>).
</t>
<iref item="shared cache" />
<iref item="private cache" />
<t anchor="shared.and.private.caches">
   A <x:dfn>shared cache</x:dfn> is a cache that stores responses for reuse
   by more than one user; shared caches are usually (but not always) deployed
   as a part of an intermediary. A <x:dfn>private cache</x:dfn>, in contrast,
   is dedicated to a single user; often, they are deployed as a component of
   a user agent.
</t>
<t>
   The goal of HTTP caching is significantly improving performance
   by reusing a prior response message to satisfy a current request.
   A cache considers a stored response "fresh", as defined in
   <xref target="expiration.model" />, if it can be reused without
   "validation" (checking with the origin server to see if the cached response
   remains valid for this request).  A fresh response can therefore
   reduce both latency and network overhead each time the cache reuses it.
   When a cached response is not fresh, it might still be reusable if validation
   can freshen it (<xref target="validation.model" />) or if the
   origin is unavailable (<xref target="serving.stale.responses" />).
</t>
<t>
   This document obsoletes <xref target="RFC7234" format="none">RFC 7234</xref>,
   with the changes being summarized in <xref target="changes.from.rfc.7234"/>.
</t>

<section title="Requirements Notation" anchor="requirements.notation">
<t>
   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 <xref target="RFC2119"/> <xref target="RFC8174"/> when, and only when, they
   appear in all capitals, as shown here.
</t>
<t>
   <xref target="HTTP" x:rel="#conformance"/> defines conformance criteria and contains considerations regarding error handling.
</t>
</section>

<section title="Syntax Notation" anchor="notation">
<iref primary="true" item="Grammar" subitem="DIGIT"/>
<t>
   This specification uses the Augmented Backus-Naur Form (ABNF) notation of
   <xref target="RFC5234"/>, extended with the notation for case-sensitivity
   in strings defined in <xref target="RFC7405"/>.
</t>
<t>
   It also uses a list extension, defined in <xref target="abnf.extension"/>,
   that allows for compact definition of comma-separated lists using a "#"
   operator (similar to how the "*" operator indicates repetition). <xref
   target="collected.abnf"/> shows the collected grammar with all list
   operators expanded to standard ABNF notation.
</t>

<section title="Imported Rules" anchor="abnf.imported">
<t anchor="core.rules">
  <x:anchor-alias value="DIGIT"/>
   The following core rule is included by
   reference, as defined in <xref target="RFC5234" x:fmt="," x:sec="B.1"/>:
   DIGIT (decimal 0-9).
</t>
<t anchor="imported.rules">
  <x:anchor-alias value="HTTP-date"/>
  <x:anchor-alias value="OWS"/>
  <x:anchor-alias value="field-name"/>
  <x:anchor-alias value="quoted-string"/>
  <x:anchor-alias value="token"/>
  <xref target="HTTP"/> defines the following rules:
</t>
<sourcecode type="abnf7230">
  <x:ref>HTTP-date</x:ref>     = &lt;HTTP-date, see <xref target="http.date"/>&gt;
  <x:ref>OWS</x:ref>           = &lt;OWS, see <xref target="whitespace"/>&gt;
  <x:ref>field-name</x:ref>    = &lt;field-name, see <xref target="fields.names"/>&gt;
  <x:ref>quoted-string</x:ref> = &lt;quoted-string, see <xref target="quoted.strings"/>&gt;
  <x:ref>token</x:ref>         = &lt;token, see <xref target="tokens"/>&gt;
</sourcecode>
</section>

<section title="Delta Seconds" anchor="delta-seconds">
<t>
   The delta-seconds rule specifies a non-negative integer, representing time
   in seconds.
</t>
<sourcecode type="abnf7230"><iref item="Grammar" primary="true" subitem="delta-seconds" />
  <x:ref>delta-seconds</x:ref>  = 1*<x:ref>DIGIT</x:ref>
</sourcecode>
<t>
   A recipient parsing a delta-seconds value and converting it to binary form
   ought to use an arithmetic type of at least 31 bits of non-negative integer
   range.
   If a cache receives a delta-seconds value greater than the greatest integer
   it can represent, or if any of its subsequent calculations overflows,
   the cache &MUST; consider the value to be 2147483648
   (2<sup>31</sup>) or the greatest positive integer it can conveniently
   represent.
</t>
<aside>
   <t>
       &Note; The value 2147483648 is here for historical reasons,
       represents infinity (over 68 years), and does not need to be stored in
       binary form; an implementation could produce it as a string if
       any overflow occurs, even if the calculations are performed with an
       arithmetic type incapable of directly representing that number.
       What matters here is that an overflow be detected and not treated as a
       negative value in later calculations.
   </t>
</aside>
</section>
</section>
</section>

<section title="Overview of Cache Operation" anchor="caching.overview">
<iref item="cache key" />
<t>
   Proper cache operation preserves the semantics of HTTP transfers
   while reducing the transmission of information already held in the
   cache. See <xref target="HTTP" x:fmt="of" x:rel="#terminology"/>
   for the general terminology and core concepts of HTTP.
</t>
<t>
   Although caching is an entirely &OPTIONAL; feature of HTTP, it can be
   assumed that reusing a cached response is desirable and that such reuse
   is the default behavior when no requirement or local configuration
   prevents it. Therefore, HTTP cache requirements are focused
   on preventing a cache from either storing a non-reusable response or
   reusing a stored response inappropriately, rather than mandating that
   caches always store and reuse particular responses.
</t>
<iref item="cache key" />
<t>
   The <x:dfn>cache key</x:dfn> is the information a cache uses to choose a response and
   is composed from, at a minimum, the request method and target
   URI used to retrieve the stored response; the method determines under which
   circumstances that response can be used to satisfy a subsequent request. However, many
   HTTP caches in common use today only cache GET responses and therefore only
   use the URI as the cache key.
</t>
<t>
   A cache might store multiple responses for a request target that is 
   subject to content negotiation. Caches differentiate these responses
   by incorporating some of the original request's header fields
   into the cache key as well, using information in the <x:ref>Vary</x:ref>
   response header field, as per <xref target="caching.negotiated.responses"/>.
</t>
<t>
  Caches might incorporate additional material into the cache key.
  For example, user agent caches might include the referring site's identity,
  thereby "double keying" the cache to avoid some privacy risks (see <xref
  target="security.timing"/>).
</t>
<t>
   Most commonly, caches store the successful result of a retrieval
   request: i.e., a <x:ref>200 (OK)</x:ref> response to a GET request, which
   contains a representation of the target resource
   (<xref target="GET"/>). However, it is also possible to store
   redirects, negative results (e.g., <x:ref>404 (Not Found)</x:ref>),
   incomplete results (e.g., <x:ref>206 (Partial Content)</x:ref>), and
   responses to methods other than GET if the method's definition allows such
   caching and defines something suitable for use as a cache key.
</t>
<t>
   A cache is <x:dfn>disconnected</x:dfn> when it cannot contact the origin
   server or otherwise find a forward path for a request. A
   disconnected cache can serve stale responses in some circumstances (<xref
   target="serving.stale.responses"/>).
</t>

</section>

<section title="Storing Responses in Caches" anchor="response.cacheability">
<t>
   A cache &MUST-NOT; store a response to a request unless:
</t>
<ul>
  <li><t>the request method is understood by the cache;</t></li>
  <li><t>the response status code is final (see
         <xref target="status.codes"/>);</t></li>
  <li><t>if the response status code is 206 or 304, or the must-understand cache directive (see <xref
         target="cache-response-directive.must-understand"/>) is present: the cache understands the response status code;</t></li>
  <li><t>the no-store cache directive is not present in the response
         (see <xref target="cache-response-directive.no-store"/>);</t></li>
  <li><t>if the cache is shared: the private response directive is either not
         present or allows a shared cache to store a modified response;
         see <xref target="cache-response-directive.private" />);</t></li>
  <li><t>if the cache is shared: the <x:ref>Authorization</x:ref> header field
         is not present in the request
         (see <xref target="field.authorization"/>) or a
         response directive is present that explicitly allows shared caching
         (see <xref target="caching.authenticated.responses" />);
         and</t></li>
  <li><t>the response contains at least one of the following:</t>
     <ul>
        <li>a public response directive
            (see <xref target="cache-response-directive.public"/>);</li>
        <li>a private response directive, if the cache is not shared
            (see <xref target="cache-response-directive.private"/>);</li>
        <li>an <x:ref>Expires</x:ref> header field
            (see <xref target="field.expires"/>);</li>
        <li>a max-age response directive
            (see <xref target="cache-response-directive.max-age" />);</li>
        <li>if the cache is shared: an s-maxage response directive
            (see <xref target="cache-response-directive.s-maxage" />);</li>
        <li>a cache extension that allows it to be cached
            (see <xref target="cache.control.extensions" />); or</li>
        <li>a status code that is defined as heuristically cacheable
            (see <xref target="heuristic.freshness" />).</li>
     </ul>
  </li>
</ul>
<t>
   Note that a cache extension can override any of the requirements
   listed; see <xref target="cache.control.extensions" />.
</t>
<t>
   In this context, a cache has "understood" a request method or a response
   status code if it recognizes it and implements all specified
   caching-related behavior.
</t>
<t>
   Note that, in normal operation, some caches will not store a response that
   has neither a cache validator nor an explicit expiration time, as such
   responses are not usually useful to store. However, caches are not
   prohibited from storing such responses.
</t>

<section title="Storing Header and Trailer Fields" anchor="storing.fields">
<t>
   Caches &MUST; include all received response header fields &mdash; including
   unrecognized ones &mdash; when storing a response; this assures that new HTTP
   header fields can be successfully deployed. However, the following exceptions
   are made:
</t>
<ul>
   <li>The Connection header field and fields whose names are listed in it are
   required by <xref target="field.connection"/> to be removed before
   forwarding the message. This &MAY; be implemented by doing so
   before storage.</li>
   <li>Likewise, some fields' semantics require them to be removed
   before forwarding the message, and this &MAY; be implemented by doing so
   before storage; see <xref target="field.connection"/> for some
   examples.</li>
   <li>The no-cache (<xref target="cache-response-directive.no-cache"/>) and
   private (<xref target="cache-response-directive.private"/>) cache
   directives can have arguments that prevent storage of header fields by all
   caches and shared caches, respectively.</li>
   <li>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
   (<xref target="field.proxy-authenticate"/>), Proxy-Authentication-Info
   (<xref target="field.proxy-authentication-info"/>), and Proxy-Authorization
   (<xref target="field.proxy-authorization"/>).</li>
</ul>
<t>
   Caches &MAY; either store trailer fields separate from header fields or
   discard them. Caches &MUST-NOT; combine trailer fields with header fields.
</t>
</section>

<section title="Updating Stored Header Fields" anchor="update">
<t>
   Caches are required to update a stored response's header fields from another
   (typically newer) response in several situations; for example, see Sections <xref
   target="combining.responses" format="counter"/>, <xref target="freshening.responses" format="counter"/>, and
   <xref target="head.effects" format="counter"/>.
</t>
<t>
   When doing so, the cache &MUST; add each header field in the provided response
   to the stored response, replacing field values that are already present,
   with the following exceptions:
</t>
<ul>
   <li>Header fields excepted from storage in <xref target="storing.fields"/>,</li>
   <li>Header fields that the cache's stored response depends upon, as described below,</li>
   <li>Header fields that are automatically processed and removed by the recipient, as described below, and</li>
   <li>The <x:ref>Content-Length</x:ref> header field.</li>
</ul>
<t>
   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.
</t>
<t>
   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 <x:ref>Content-Encoding</x:ref>
   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 <x:ref>Content-Type</x:ref> header field would not be workable
   in this case because any assumptions about the format made in parsing would
   now be invalid.
</t>
<t>
   Furthermore, some fields are automatically processed and removed by the
   HTTP implementation, such as the <x:ref>Content-Range</x:ref> header field.
   Implementations &MAY; automatically omit such header fields from updates,
   even when the processing does not actually occur.
</t>
<t>
   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.
</t>
</section>

<section title="Storing Incomplete Responses" anchor="incomplete.responses">
<t>
   If the request method is GET, the response status code is <x:ref>200
   (OK)</x:ref>, and the entire response header section has been received, a
   cache &MAY; store a response that is not complete (<xref target="message.framing"/>) provided that the stored response
   is recorded as being incomplete. Likewise, a <x:ref>206 (Partial
   Content)</x:ref> response &MAY; be stored as if it were an incomplete
   <x:ref>200 (OK)</x:ref> response. However, a cache &MUST-NOT; store
   incomplete or partial-content responses if it does not support the
   <x:ref>Range</x:ref> and <x:ref>Content-Range</x:ref> header fields or if
   it does not understand the range units used in those fields.
</t>
<t>
   A cache &MAY; complete a stored incomplete response by making a subsequent
   range request (<xref target="field.range"/>) and combining the successful response with the
   stored response, as defined in <xref target="combining.responses"/>. 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 <x:ref>206 (Partial Content)</x:ref> status code.
</t>
</section>

<section title="Combining Partial Content" anchor="combining.responses">
<t>
   A response might transfer only a partial representation if the
   connection closed prematurely or if the request used one or more Range
   specifiers (<xref target="field.range"/>).  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 <xref target="combining.byte.ranges"/>.
</t>
<t>
   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 <xref target="update"/>.
</t>
</section>

<section title="Storing Responses to Authenticated Requests" anchor="caching.authenticated.responses">
<t>
   A shared cache &MUST-NOT; use a cached response to a request with an
   <x:ref>Authorization</x:ref> header field (<xref target="field.authorization"/>) to
   satisfy any subsequent request unless the response contains a
   <x:ref>Cache-Control</x:ref> field with a response directive
   (<xref target="cache-response-directive"/>) that allows it to be stored by
   a shared cache, and the cache conforms to the requirements of that
   directive for that response.
</t>
<t>
   In this specification, the following response directives have such an effect:
   must-revalidate (<xref target="cache-response-directive.must-revalidate"/>),
   public (<xref target="cache-response-directive.public"/>), and
   s-maxage (<xref target="cache-response-directive.s-maxage"/>).
</t>
</section>

</section>


<section title="Constructing Responses from Caches" anchor="constructing.responses.from.caches">
<t>
   When presented with a request, a cache &MUST-NOT; reuse a stored response
   unless:
</t>
<ul>
   <li><t>the presented target URI (<xref target="target.resource"/>) and
   that of the stored response match, and</t></li>
   <li><t>the request method associated with the stored response allows it to
   be used for the presented request, and</t></li>
   <li><t>request header fields nominated by the stored response (if any)
   match those presented (see <xref target="caching.negotiated.responses"
   />), and</t></li>
   <li><t>the stored response does not contain the no-cache directive
   (<xref target="cache-response-directive.no-cache"/>), unless it is
   successfully validated (<xref target="validation.model"/>), and</t></li>
   <li><t>the stored response is one of the following:</t>
     <ul>
        <li>fresh (see <xref target="expiration.model" />), or</li>
        <li>allowed to be served stale (see <xref
        target="serving.stale.responses" />), or</li>
        <li>successfully validated (see <xref target="validation.model"
        />).</li>
     </ul>
   </li>
</ul>
<t>
   Note that a cache extension can override any of the requirements
   listed; see <xref target="cache.control.extensions" />.
</t>
<t>
   When a stored response is used to satisfy a request without validation, a
   cache &MUST; generate an <x:ref>Age</x:ref> header field (<xref
   target="field.age"/>), replacing any present in the response with a value
   equal to the stored response's current_age; see <xref
   target="age.calculations" />.
</t>
<t>
   A cache &MUST; write through requests with methods that are unsafe
   (<xref target="safe.methods"/>) to the origin server; i.e., a cache is not allowed to
   generate a reply to such a request before having forwarded the request and
   having received a corresponding response.
</t>
<t>
   Also, note that unsafe requests might invalidate already-stored responses;
   see <xref target="invalidation" />.
</t>
<iref item="collapsed requests" />
<t>
   A cache can use a response that is stored or storable to satisfy
   multiple requests, provided that it is allowed to reuse that response
   for the requests in question. This enables a cache to <x:dfn>collapse
   requests</x:dfn> &mdash; or combine multiple incoming requests into a single forward
   request upon a cache miss &mdash; thereby reducing load on the origin server
   and network. Note, however, that if the cache cannot use the returned
   response for some or all of the collapsed requests, it will need to
   forward the requests in order to satisfy them, potentially introducing
   additional latency.
</t>
<t>
   When more than one suitable response is stored, a cache &MUST; use the
   most recent one (as determined by the <x:ref>Date</x:ref> header
   field). It can also forward the request with "Cache-Control: max-age=0" or
   "Cache-Control: no-cache" to disambiguate which response to use.
</t>
<t>
   A cache without a clock (<xref target="http.date"/>) &MUST; revalidate
   stored responses upon every use.
</t>

<section title="Calculating Cache Keys with the Vary Header Field" anchor="caching.negotiated.responses">
<t>
   When a cache receives a request that can be satisfied by a stored response
   and that stored response contains a <x:ref>Vary</x:ref> header field
   (<xref target="field.vary"/>),
   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).
</t>
<t>
   The header fields from two requests are defined to match if
   and only if those in the first request can be transformed to those in the
   second request by applying any of the following:
</t>
<ul>
  <li>
     adding or removing whitespace, where allowed in the header field's
     syntax
  </li>
  <li>
     combining multiple header field lines with the same field name
     (see <xref target="field.lines"/>)
  </li>
  <li>
     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)
  </li>
</ul>
<t>
   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.
</t>
<t>
   A stored response with a <x:ref>Vary</x:ref> header field value containing
   a member "*" always fails to match.
</t>
<t>
   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 <x:ref>Accept</x:ref> 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 <x:ref>Date</x:ref> header field) is
   chosen, as
   per <xref target="constructing.responses.from.caches"/>.
</t>
<t>
   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 <xref target="age.calculations"/>) stored response with a valid Vary
   field value.
</t>
<t>
   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 (<xref target="validation.model"/>).
</t>
</section>

<section title="Freshness" anchor="expiration.model">
<iref item="fresh" />
<iref item="stale" />
<t>
   A <x:dfn>fresh</x:dfn> response is one whose age has not yet exceeded its
   freshness lifetime. Conversely, a <x:dfn>stale</x:dfn>
   response is one where it has.
</t>
<iref item="freshness lifetime" />
<iref item="explicit expiration time" />
<iref item="heuristic expiration time" />
<t>
   A response's <x:dfn>freshness lifetime</x:dfn> is the length of time
   between its generation by the origin server and its expiration time. An
   <x:dfn>explicit expiration time</x:dfn> 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 <x:dfn>heuristic expiration
   time</x:dfn> is assigned by a cache when no explicit expiration time is
   available.
</t>
<iref item="age" />
<t>
   A response's <x:dfn>age</x:dfn> is the time that has passed since it was
   generated by, or successfully validated with, the origin server.
</t>
<t>
   When a response is fresh, it can be used to satisfy
   subsequent requests without contacting the origin server, thereby improving
   efficiency.
</t>
<t>
   The primary mechanism for determining freshness is for an origin server to
   provide an explicit expiration time in the future, using either the
   <x:ref>Expires</x:ref> header field (<xref target="field.expires" />) or
   the max-age response directive (<xref
   target="cache-response-directive.max-age" />). 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.
</t>
<t>
   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 <xref
   target="serving.stale.responses" />).
</t>
<t>
   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 <xref target="heuristic.freshness"/>).
</t>
<t>
  The calculation to determine if a response is fresh is:
</t>
<sourcecode type="pseudocode">
   response_is_fresh = (freshness_lifetime &gt; current_age)
</sourcecode>
<t>
   freshness_lifetime is defined in <xref
   target="calculating.freshness.lifetime" />; current_age is defined in
   <xref target="age.calculations" />.
</t>
<t>
   Clients can send the max-age or min-fresh request directives (<xref
   target="cache-request-directive" />) to suggest limits on the freshness
   calculations for the corresponding response. However, caches are not
   required to honor them.
</t>
<t>
   When calculating freshness, to avoid common problems in date parsing:
</t>
<ul>
   <li>Although all date formats are specified to be case-sensitive,
      a cache recipient &SHOULD; match the field value
      case-insensitively.</li>
   <li>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 <x:ref>Expires</x:ref> date as the
      nearest time equal to or earlier than the received value.</li>
   <li>A cache recipient &MUST-NOT; allow local time zones to influence the
      calculation or comparison of an age or expiration time.</li>
   <li>A cache recipient &SHOULD; consider a date with a zone abbreviation
      other than "GMT" to be invalid for calculating expiration.</li>
</ul>
<t>
   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 <xref
   target="history.lists" /> for an explanation of the difference between
   caches and history mechanisms.
</t>

<section title="Calculating Freshness Lifetime" anchor="calculating.freshness.lifetime">
<t>
   A cache can calculate the freshness lifetime (denoted as
   freshness_lifetime) of a response by evaluating the following rules and using the first match:
</t>
<ul>
   <li>If the cache is shared and the s-maxage response directive
   (<xref target="cache-response-directive.s-maxage" />) is present, use its value,
   or</li>
   <li>If the max-age response directive (<xref
   target="cache-response-directive.max-age" />) is present, use its value, or</li>
   <li>If the <x:ref>Expires</x:ref> response header field
   (<xref target="field.expires" />) is present, use its value minus the
    value of the <x:ref>Date</x:ref> response header field
    (using the time the message was received if it is not present, as per <xref target="field.date"/>), or</li>
   <li>Otherwise, no explicit expiration time is present in the response. A
   heuristic freshness lifetime might be applicable; see <xref
   target="heuristic.freshness" />.</li>
</ul>
<t>
   Note that this calculation is intended to reduce clock skew by using the
   clock information provided by the origin server whenever possible.
</t>
<t>
   When there is more than one value present for a given directive (e.g., two
   <x:ref>Expires</x:ref> header field lines or multiple Cache-Control: max-age
   directives), either the first occurrence should be used or the response should
   be considered stale. If directives conflict (e.g.,
   both max-age and no-cache are present), the most restrictive directive should
   be honored. Caches are encouraged to consider responses that have
   invalid freshness information (e.g., a max-age directive with non-integer content) to
   be stale.
</t>
</section>

<section title="Calculating Heuristic Freshness" anchor="heuristic.freshness">
  <iref item="heuristically cacheable"/>
<t>
   Since origin servers do not always provide explicit expiration times, a
   cache &MAY; assign a heuristic expiration time when an explicit time is not
   specified, employing algorithms that use other field values (such as
   the <x:ref>Last-Modified</x:ref> time) to estimate a plausible expiration
   time. This specification does not provide specific algorithms, but it does
   impose worst-case constraints on their results.
</t>
<t>
   A cache &MUST-NOT; use heuristics to determine freshness when an explicit
   expiration time is present in the stored response. Because of the
   requirements in <xref target="response.cacheability"/>,
   heuristics can only be used on responses without explicit
   freshness whose status codes are defined as <x:dfn>heuristically cacheable</x:dfn> (e.g., see
   <xref target="overview.of.status.codes"/>) and on responses without
   explicit freshness that have been marked as explicitly cacheable (e.g.,
   with a public response directive).
</t>
<t>
   Note that in previous specifications, heuristically cacheable response status
   codes were called "cacheable by default".
</t>
<t>
   If the response has a <x:ref>Last-Modified</x:ref> header field
   (<xref target="field.last-modified"/>), caches are encouraged to use a heuristic
   expiration value that is no more than some fraction of the interval since
   that time. A typical setting of this fraction might be 10%.
</t>
<aside>
   <t>
      &Note;
      A previous version of the HTTP specification
      (<xref target="RFC2616" x:fmt="of" x:sec="13.9"/>) prohibited caches
      from calculating heuristic freshness for URIs with query components
      (i.e., those containing "?"). In practice, this has not been widely
      implemented. Therefore, origin servers are encouraged to send explicit
      directives (e.g., Cache-Control: no-cache) if they wish to prevent
      caching.
   </t>
</aside>
</section>

<section title="Calculating Age" anchor="age.calculations">
<t>
   The <x:ref>Age</x:ref> header field is used to convey an estimated
   age of the response message when obtained from a cache. The Age field value
   is the cache's estimate of the number of seconds since the origin server generated
   or validated the response. The Age value is therefore
   the sum of the time that the response has been resident in each of the
   caches along the path from the origin server, plus the time it
   has been in transit along network paths.
</t>
<t>
   Age calculation uses the following data:
</t>
<dl newline="true">
  <dt><x:dfn>age_value</x:dfn></dt>
  <dd>
     The term "age_value" denotes the value of the <x:ref>Age</x:ref>
     header field (<xref target="field.age"/>), in a form appropriate for
     arithmetic operation; or 0, if not available.
  </dd>
  <dt><x:dfn>date_value</x:dfn></dt>
  <dd>
     The term "date_value" denotes the value of
     the Date header field, in a form appropriate for arithmetic
     operations. See <xref target="field.date"/> for the definition of the Date header
     field and for requirements regarding responses without it.
  </dd>
  <dt><x:dfn>now</x:dfn></dt>
  <dd>
     The term "now" means the current value of this implementation's clock
     (<xref target="http.date"/>).
  </dd>
  <dt><x:dfn>request_time</x:dfn></dt>
  <dd>
     The value of the clock at the time of the request that
     resulted in the stored response.
  </dd>
  <dt><x:dfn>response_time</x:dfn></dt>
  <dd>
     The value of the clock at the time the response
     was received.
  </dd>
</dl>
<t>
   A response's age can be calculated in two entirely independent ways:
</t>
<ol>
   <li>the "apparent_age": response_time minus date_value, if the
     implementation's
   clock is reasonably well synchronized to the origin server's clock. If
   the result is negative, the result is replaced by zero.</li>
   <li>the "corrected_age_value", if all of the caches along the response
   path implement HTTP/1.1 or greater. A cache &MUST; interpret this value
   relative to the time the request was initiated, not the time that the
   response was received.</li>
</ol>
<sourcecode type="pseudocode">
  apparent_age = max(0, response_time - date_value);

  response_delay = response_time - request_time;
  corrected_age_value = age_value + response_delay;
</sourcecode>
<t>
   The corrected_age_value &MAY; be used as the corrected_initial_age. In
   circumstances where very old cache implementations that might not correctly
   insert <x:ref>Age</x:ref> are present, corrected_initial_age can be calculated
   more conservatively as
</t>
<sourcecode type="pseudocode">
  corrected_initial_age = max(apparent_age, corrected_age_value);
</sourcecode>
<t>
   The current_age of a stored response can then be calculated by adding the
   time (in seconds) since the stored response was last validated by
   the origin server to the corrected_initial_age.
</t>
<sourcecode type="pseudocode">
  resident_time = now - response_time;
  current_age = corrected_initial_age + resident_time;
</sourcecode>
</section>

<section title="Serving Stale Responses" anchor="serving.stale.responses">
<t>
   A "stale" response is one that either has explicit expiry information or is
   allowed to have heuristic expiry calculated, but is not fresh according to
   the calculations in <xref target="expiration.model" />.
</t>
<t>
   A cache &MUST-NOT; generate a stale response if it is prohibited by an
   explicit in-protocol directive (e.g., by a no-cache response
   directive, a must-revalidate response directive, or an applicable
   s-maxage or proxy-revalidate response directive; see <xref
   target="cache-response-directive"/>).
</t>
<t>
   A cache &MUST-NOT; generate a stale response unless it is disconnected
   or doing so is explicitly permitted by the client or origin server
   (e.g., by the max-stale request directive in <xref
   target="cache-request-directive" />, extension directives such as those
   defined in <xref target="RFC5861"/>, or configuration in accordance
   with an out-of-band contract).
</t>
</section>
</section>

<section title="Validation" anchor="validation.model">
<t>
   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 <xref target="caching.negotiated.responses"/>), it can use
   the conditional request mechanism (<xref target="conditional.requests"/>) 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
   <x:dfn>validating</x:dfn> or <x:dfn>revalidating</x:dfn> the stored
   response.
</t>

<section title="Sending a Validation Request" anchor="validation.sent"><iref item="validator" />
<t>
   When generating a conditional request for validation, a cache either starts with
   a request it is attempting to satisfy or &mdash; if it is initiating
   the request independently &mdash; synthesizes a request using a stored
   response by copying the method, target URI, and request header fields
   identified by the Vary header field (<xref
   target="caching.negotiated.responses"/>).
</t>
<t>
   It then updates that request with one or more precondition header fields.
   These contain validator metadata sourced from a stored response(s) that has
   the same URI. Typically, this will include only the stored response(s) that
   has the same cache key, although a cache is allowed to validate
   a response that it cannot choose with the request header fields it is sending
   (see <xref target="caching.negotiated.responses"/>).
</t>
<t>
   The precondition header fields are then compared by recipients to
   determine whether any stored response is equivalent to a current
   representation of the resource.
</t>
<t>
   One such validator is the timestamp given in a <x:ref>Last-Modified</x:ref>
   header field (<xref target="field.last-modified"/>), which can be used in an
   <x:ref>If-Modified-Since</x:ref> header field for response validation, or
   in an <x:ref>If-Unmodified-Since</x:ref> or <x:ref>If-Range</x:ref> header
   field for representation selection (i.e., the client is referring
   specifically to a previously obtained representation with that timestamp).
</t>
<t>
   Another validator is the entity tag given in an <x:ref>ETag</x:ref>
   field (<xref target="field.etag"/>). One or more entity tags, indicating one or more
   stored responses, can be used in an <x:ref>If-None-Match</x:ref> header
   field for response validation, or in an <x:ref>If-Match</x:ref> or
   <x:ref>If-Range</x:ref> header field for representation selection (i.e.,
   the client is referring specifically to one or more previously obtained
   representations with the listed entity tags).
</t>
<t>
   When generating a conditional request for validation, a cache:
</t>
<ul>
   <li>&MUST; send the relevant entity tags
     (using <x:ref>If-Match</x:ref>, <x:ref>If-None-Match</x:ref>, or
     <x:ref>If-Range</x:ref>) if the entity tags were provided in the
     stored response(s) being validated.</li>
   <li>&SHOULD; send the <x:ref>Last-Modified</x:ref> value (using 
     <x:ref>If-Modified-Since</x:ref>) if the request is not for a subrange, 
     a single stored response is being validated, and that response 
     contains a Last-Modified value.</li>
   <li>&MAY; send the <x:ref>Last-Modified</x:ref> value (using 
     <x:ref>If-Unmodified-Since</x:ref> or <x:ref>If-Range</x:ref>) if 
     the request is for a subrange, a single stored response is being 
     validated, and that response contains only a Last-Modified value
     (not an entity tag).</li>
</ul>
<t>
   In most cases, both validators are generated in cache validation requests,
   even when entity tags are clearly superior, to allow old intermediaries
   that do not understand entity tag preconditions to respond appropriately.
</t>
</section>

<section title="Handling a Received Validation Request" anchor="validation.received">
<t>
   Each client in the request chain may have its own cache, so it is common
   for a cache at an intermediary to receive conditional requests from other
   (outbound) caches. Likewise, some user agents make use of conditional
   requests to limit data transfers to recently modified representations or to
   complete the transfer of a partially retrieved representation.
</t>
<t>
   If a cache receives a request that can be satisfied by reusing a stored
   <x:ref>200 (OK)</x:ref> or <x:ref>206 (Partial Content)</x:ref>
   response, as per <xref target="constructing.responses.from.caches"/>,
   the cache &SHOULD; evaluate any applicable conditional header
   field preconditions received in that request with respect to the
   corresponding validators contained within the stored response.
</t>
<t>
   A cache &MUST-NOT; evaluate conditional header fields that only
   apply to an origin server, occur in a request with semantics that
   cannot be satisfied with a cached response, or occur in a request with a target resource
   for which it has no stored responses; such preconditions are likely
   intended for some other (inbound) server.
</t>
<t>
   The proper evaluation of conditional requests by a cache depends on the
   received precondition header fields and their precedence. In summary, the <x:ref>If-Match</x:ref> and
   <x:ref>If-Unmodified-Since</x:ref> conditional header fields are not
   applicable to a cache, and <x:ref>If-None-Match</x:ref> takes precedence over
   <x:ref>If-Modified-Since</x:ref>. See <xref target="precedence"/> for
   a complete specification of precondition precedence.
</t>
<t>
   A request containing an <x:ref>If-None-Match</x:ref> header field (<xref
   target="field.if-none-match"/>) indicates that the client wants to
   validate one or more of its own stored responses in comparison to the
   stored response chosen by the cache (as per <xref
   target="constructing.responses.from.caches"/>).
</t>
<t>
   If an <x:ref>If-None-Match</x:ref> header field is not present, a request
   containing an <x:ref>If-Modified-Since</x:ref> header field
   (<xref target="field.if-modified-since"/>) indicates that the client wants to validate
   one or more of its own stored responses by modification date.
</t>
<t>
   If a request contains an <x:ref>If-Modified-Since</x:ref> header field and
   the <x:ref>Last-Modified</x:ref> header field is not present in a
   stored response, a cache &SHOULD; use the stored response's
   <x:ref>Date</x:ref> field value (or, if no Date field is present, the time
   that the stored response was received) to evaluate the conditional.
</t>
<t>
   A cache that implements partial responses to range requests, as defined in
   <xref target="field.range"/>, also needs to evaluate a received
   <x:ref>If-Range</x:ref> header field (<xref target="field.if-range"/>)
   with respect to the cache's chosen response.
</t>
<t>
   When a cache decides to forward a request to revalidate its own stored
   responses for a
   request that contains an <x:ref>If-None-Match</x:ref> list of entity tags,
   the cache &MAY; combine the received list with a list of entity tags
   from its own stored set of responses (fresh or stale) and send the union of
   the two lists as a replacement <x:ref>If-None-Match</x:ref> header
   field value in the forwarded request.
   If a stored response contains only partial content, the
   cache &MUST-NOT; include its entity tag in the union unless the request is
   for a range that would be fully satisfied by that partial stored response.
   If the response to the forwarded request is
   <x:ref>304 (Not Modified)</x:ref> and has an ETag field value with
   an entity tag that is not in the client's list, the cache &MUST;
   generate a <x:ref>200 (OK)</x:ref> response for the client by reusing its
   corresponding stored response, as updated by the 304 response metadata
   (<xref target="freshening.responses"/>).
</t>
</section>

<section title="Handling a Validation Response" anchor="validation.response">
<t>
   Cache handling of a response to a conditional request depends upon its
   status code:
</t>
<ul>
   <li>
     A <x:ref>304 (Not Modified)</x:ref> response status code indicates
     that the stored response can be updated and reused; see <xref
     target="freshening.responses"/>.
   </li>
   <li>
     A full response (i.e., one containing content) indicates that none
     of the stored responses nominated in the conditional request are
     suitable. Instead, the cache &MUST; use the full response to
     satisfy the request. The cache &MAY; store such a full response,
     subject to its constraints (see <xref target="response.cacheability"/>).
   </li>
   <li>
     However, if a cache receives a <x:ref>5xx (Server Error)</x:ref>
     response while attempting to validate a response, it can either
     forward this response to the requesting client or act as if the
     server failed to respond. In the latter case, the cache can send a
     previously stored response, subject to its constraints on doing so (see <xref
     target="serving.stale.responses" />), or retry the validation request.
   </li>
</ul>
</section>

<section title="Freshening Stored Responses upon Validation" anchor="freshening.responses">
<t>
   When a cache receives a <x:ref>304 (Not Modified)</x:ref> response, it needs
   to identify stored responses that are suitable for updating with the new information
   provided, and then do so.
</t>
<t>
   The initial set of stored responses to update are those that could have been chosen for
   that request &mdash; i.e., those that meet the requirements in <xref
   target="constructing.responses.from.caches"/>, except the last requirement
   to be fresh, able to be served stale, or just validated.
</t>
<t>
   Then, that initial set of stored responses is further filtered by the first match of:
</t>
<ul>
   <li>
     If the new response contains one or more <x:dfn>strong validators</x:dfn> (see
     <xref target="weak.and.strong.validators"/>), then each of those strong validators
     identifies a selected representation for update. All the stored
     responses in the initial set with one of those same strong validators are identified for update. If
     none of the initial set contains at least one of the same strong validators, then the
     cache &MUST-NOT; use the new response to update any stored responses.
   </li>
   <li>
     If the new response contains no strong validators but does contain
     one or more <x:dfn>weak validators</x:dfn>, and those
     validators correspond to one of the initial set's stored responses, then the most
     recent of those matching stored responses is identified for update.
   </li>
   <li>
     If the new response does not include any form of validator (such as
     where a client generates an <x:ref>If-Modified-Since</x:ref> request from
     a source other than the <x:ref>Last-Modified</x:ref> response header
     field), and there is only one stored response in the initial set, and that stored response
     also lacks a validator, then that stored response is identified for update.
   </li>
</ul>
<t>
   For each stored response identified, the cache &MUST; update
   its header fields with the header fields provided in the <x:ref>304 (Not
   Modified)</x:ref> response, as per <xref target="update"/>.
</t>
</section>

<section title="Freshening Responses with HEAD" anchor="head.effects">
<t>
   A response to the HEAD method is identical to what an equivalent request
   made with a GET would have been, without sending the content. This property
   of HEAD responses can be used to invalidate or update a cached GET
   response if the more efficient conditional GET request mechanism is not
   available (due to no validators being present in the stored response) or
   if transmission of the content is not desired even if it has
   changed.
</t>
<t>
   When a cache makes an inbound HEAD request for a target URI and
   receives a <x:ref>200 (OK)</x:ref> response, the cache &SHOULD; update or
   invalidate each of its stored GET responses that could have been chosen
   for that request (see <xref target="caching.negotiated.responses"/>).
</t>
<t>
   For each of the stored responses that could have been chosen, if the
   stored response and HEAD response have matching values for any received
   validator fields (<x:ref>ETag</x:ref> and <x:ref>Last-Modified</x:ref>)
   and, if the HEAD response has a <x:ref>Content-Length</x:ref> header field,
   the value of <x:ref>Content-Length</x:ref> matches that of the stored
   response, the cache &SHOULD; update the stored response as described below;
   otherwise, the cache &SHOULD; consider the stored response to be stale.
</t>
<t>
   If a cache updates a stored response with the metadata provided in a HEAD
   response, the cache &MUST; use the header fields provided in the HEAD
   response to update the stored response (see <xref
   target="update"/>).
</t>
</section>
</section>

<section title="Invalidating Stored Responses" anchor="invalidation">
<t>
   Because unsafe request methods (<xref target="safe.methods"/>) 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.
</t>
<t>
   A cache &MUST; invalidate the target URI
   (<xref target="target.resource"/>) when it receives a non-error status
   code in response to
   an unsafe request method (including methods whose safety is unknown).
</t>
<t>
   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
   <x:ref>Location</x:ref> and <x:ref>Content-Location</x:ref> 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 (<xref target="origin"/>) of the URI to be invalidated differs from that of the target URI
   (<xref target="target.resource"/>). This helps prevent denial-of-service attacks.
</t>
<t>
   <x:dfn>Invalidate</x:dfn> 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.
</t>
<t>
   A "non-error response" is one with a <x:ref>2xx (Successful)</x:ref>
   or <x:ref>3xx (Redirection)</x:ref> status code.
</t>
<t>
   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.
</t>
</section>
</section>



<section title="Field Definitions" anchor="header.field.definitions">
<t>
   This section defines the syntax and semantics of HTTP fields
   related to caching.
</t>

<section title="Age" anchor="field.age">
  <x:anchor-alias value="header.age"/>
   <iref primary="true" item="Fields" subitem="Age" x:for-anchor=""/><iref primary="true" item="Header Fields" subitem="Age" x:for-anchor=""/><iref primary="true" item="Fields" subitem="Age" x:for-anchor=""/><iref primary="true" item="Header Fields" subitem="Age" x:for-anchor=""/><iref item="Age header field" primary="true" x:for-anchor=""/>
   <x:anchor-alias value="Age"/>
   <x:anchor-alias value="age-value"/>
<t>
   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 <xref
   target="age.calculations" />.
</t>
<sourcecode type="abnf7230"><iref primary="true" item="Grammar" subitem="Age"/>
  <x:ref>Age</x:ref> = <x:ref>delta-seconds</x:ref>
</sourcecode>
<t>
   The Age field value is a non-negative integer, representing time in seconds
   (see <xref target="delta-seconds"/>).
</t>
<t>
   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.
</t>
<t>
   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.
</t>
<t>
   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.
</t>
</section>

<section title="Cache-Control" anchor="field.cache-control">
  <x:anchor-alias value="header.cache-control"/>
   <iref primary="true" item="Fields" subitem="Cache-Control" x:for-anchor=""/><iref primary="true" item="Header Fields" subitem="Cache-Control" x:for-anchor=""/><iref item="Cache-Control header field" primary="true" x:for-anchor=""/>
   <x:anchor-alias value="Cache-Control"/>
   <x:anchor-alias value="cache-directive"/>
<t>
   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.
</t>
<t>
   See <xref target="cache.control.extensions"/> for information about how
   Cache-Control directives defined elsewhere are handled.
</t>
<t>
   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.
</t>
<t>
   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.
</t>
<sourcecode type="abnf7230"><iref primary="true" item="Grammar" subitem="Cache-Control"/><iref primary="true" item="Grammar" subitem="cache-directive"/>
  <x:ref>Cache-Control</x:ref>   = #<x:ref>cache-directive</x:ref>

  <x:ref>cache-directive</x:ref> = <x:ref>token</x:ref> [ "=" ( <x:ref>token</x:ref> / <x:ref>quoted-string</x:ref> ) ]
</sourcecode>
<t>
   For the cache directives defined below, no argument is defined (nor allowed)
   unless stated otherwise.
</t>

<section title="Request Directives" anchor="cache-request-directive" x:assert-sorted-by="title">

<t>
   This section defines cache request directives. They are advisory; caches
   &MAY; implement them, but are not required to.
</t>


<section title="max-age" anchor="cache-request-directive.max-age">
   <iref item="max-age (cache directive)" primary="true"/>
<t>
   Argument syntax:
</t>
<ul empty="true">
  <li><x:ref>delta-seconds</x:ref> (see <xref target="delta-seconds"/>)</li>
</ul>
<t>
   The max-age request directive indicates that the client prefers a
   response whose age is less than or equal to the specified number of
   seconds. Unless the max-stale request directive is also present, the
   client does not wish to receive a stale response.
</t>
<t>
   This directive uses the token form of the argument syntax:
   e.g., 'max-age=5' not 'max-age="5"'. A sender &MUST-NOT; generate the
   quoted-string form.
</t>
</section>

<section title="max-stale" anchor="cache-request-directive.max-stale">
   <iref item="max-stale (cache directive)" primary="true"/>
<t>
   Argument syntax:
</t>
<ul empty="true">
  <li><x:ref>delta-seconds</x:ref> (see <xref target="delta-seconds"/>)</li>
</ul>
<t>
   The max-stale request directive indicates that the client will
   accept a response that has exceeded its freshness lifetime. If a value is
   present, then the client is willing to accept a response that has exceeded
   its freshness lifetime by no more than the specified number of seconds. If
   no value is assigned to max-stale, then the client will accept a
   stale response of any age.
</t>
<t>
   This directive uses the token form of the argument syntax:
   e.g., 'max-stale=10' not 'max-stale="10"'. A sender &MUST-NOT; generate
   the quoted-string form.
</t>
</section>

<section title="min-fresh" anchor="cache-request-directive.min-fresh">
   <iref item="min-fresh (cache directive)" primary="true"/>
<t>
   Argument syntax:
</t>
<ul empty="true">
  <li><x:ref>delta-seconds</x:ref> (see <xref target="delta-seconds"/>)</li>
</ul>
<t>
   The min-fresh request directive indicates that the client prefers a
   response whose freshness lifetime is no less than its current age plus the
   specified time in seconds. That is, the client wants a response that will
   still be fresh for at least the specified number of seconds.
</t>
<t>
   This directive uses the token form of the argument syntax:
   e.g., 'min-fresh=20' not 'min-fresh="20"'. A sender &MUST-NOT; generate
   the quoted-string form.
</t>
</section>

<section title="no-cache" anchor="cache-request-directive.no-cache">
   <iref item="no-cache (cache directive)" primary="true" />
<t>
   The no-cache request directive indicates that the client prefers
   a stored response not be used to satisfy the request without successful
   validation on the origin server.
</t>
</section>

<section title="no-store" anchor="cache-request-directive.no-store">
   <iref item="no-store (cache directive)" primary="true" />
<t>
   The no-store request directive indicates that a cache &MUST-NOT;
   store any part of either this request or any response to it. This
   directive applies to both private and shared caches. "MUST NOT
   store" in this context means that the cache &MUST-NOT; intentionally
   store the information in non-volatile storage and &MUST; make a
   best-effort attempt to remove the information from volatile storage as
   promptly as possible after forwarding it.
</t>
<t>
   This directive is not a reliable or sufficient mechanism for ensuring
   privacy. In particular, malicious or compromised caches might not
   recognize or obey this directive, and communications networks might be
   vulnerable to eavesdropping.
</t>
<t>
   Note that if a request containing this directive is satisfied from a
   cache, the no-store request directive does not apply to the already
   stored response.
</t>
</section>

<section title="no-transform" anchor="cache-request-directive.no-transform">
   <iref item="no-transform (cache directive)" primary="true" />
<t>
   The no-transform request directive indicates that the client is asking
   for intermediaries to avoid
   transforming the content, as defined in <xref
   target="message.transformations"/>.
</t>
</section>

<section title="only-if-cached" anchor="cache-request-directive.only-if-cached">
   <iref item="only-if-cached (cache directive)" primary="true" />
<t>
   The only-if-cached request directive indicates that the client only
   wishes to obtain a stored response. Caches that honor this request
   directive &SHOULD;, upon receiving it, respond with either a stored
   response consistent with the other constraints of the request or
   a <x:ref>504 (Gateway Timeout)</x:ref> status code.
</t>
</section>
</section>

<section title="Response Directives" anchor="cache-response-directive" x:assert-sorted-by="title">
   <x:anchor-alias value="cache-response-directive" />

<t>
   This section defines cache response directives. A cache &MUST; obey the
   Cache-Control directives defined in this section.
</t>

<section title="max-age" anchor="cache-response-directive.max-age">
   <iref item="max-age (cache directive)" primary="true" />
<t>
   Argument syntax:
</t>
<ul empty="true">
  <li><x:ref>delta-seconds</x:ref> (see <xref target="delta-seconds"/>)</li>
</ul>
<t>
   The max-age response directive indicates that the response is to be
   considered stale after its age is greater than the specified number of
   seconds.
</t>
<t>
   This directive uses the token form of the argument syntax:
   e.g., 'max-age=5' not 'max-age="5"'. A sender &MUST-NOT; generate the
   quoted-string form.
</t>
</section>

<section title="must-revalidate" anchor="cache-response-directive.must-revalidate">
   <iref item="must-revalidate (cache directive)" primary="true" />
<t>
   The must-revalidate response directive indicates that once the response
   has become stale, a cache &MUST-NOT; reuse that response to satisfy
   another request until it has been successfully validated by the origin, as
   defined by <xref target="validation.model"/>.
</t>
<t>
   The must-revalidate directive is necessary to support reliable operation
   for certain protocol features. In all circumstances, a cache &MUST-NOT; ignore
   the must-revalidate directive; in particular, if a cache is disconnected,
   the cache &MUST; generate an error response rather than reuse the stale response.
   The generated status code &SHOULD; be <x:ref>504 (Gateway Timeout)</x:ref>
   unless another error status code is more applicable.
</t>
<t>
   The must-revalidate directive ought to be used by servers if and only
   if failure to validate a request could cause
   incorrect operation, such as a silently unexecuted financial
   transaction.
</t>
<t>
   The must-revalidate directive also permits a shared cache to
   reuse a response to a request containing an <x:ref>Authorization</x:ref>
   header field (<xref target="field.authorization"/>),
   subject to the above requirement on revalidation
   (<xref target="caching.authenticated.responses"/>).
</t>
</section>

<section title="must-understand" anchor="cache-response-directive.must-understand">
   <iref item="must-understand (cache directive)" primary="true" />
<t>
   The must-understand response directive limits caching of the response to
   a cache that understands and conforms to the requirements for that
   response's status code.
</t>
<t>
   A response that contains the must-understand directive &SHOULD;
   also contain the no-store directive. When a cache that implements the
   must-understand directive receives a response that includes it,
   the cache &SHOULD; ignore the no-store directive if it
   understands and implements the status code's caching requirements.
</t>
</section>

<section title="no-cache" anchor="cache-response-directive.no-cache">
   <iref item="no-cache (cache directive)" primary="true" />
<t>
   Argument syntax:
</t>
<ul empty="true">
  <li>#<x:ref>field-name</x:ref></li>
</ul>
<t>
   The no-cache response directive, in its unqualified form (without an
   argument), indicates that the response &MUST-NOT; be used to satisfy any
   other request without forwarding it for validation and receiving a
   successful response; see <xref target="validation.model"/>.
</t>
<t>
   This allows an origin server to prevent a cache from using
   the response to satisfy a request without contacting it, even by caches that have
   been configured to send stale responses.
</t>
<t>
   The qualified form of the no-cache response directive, with an argument that
   lists one or more field names, indicates that a cache &MAY; use the
   response to satisfy a subsequent request, subject to any other restrictions
   on caching, if the listed header fields are excluded from the subsequent
   response or the subsequent response has been successfully revalidated with
   the origin server (updating or removing those fields).
   This allows an origin server to prevent the reuse of certain header
   fields in a response, while still allowing caching of the rest of the
   response.
</t>
<t>
   The field names given are not limited to the set of header
   fields defined by this specification. Field names are case-insensitive.
</t>
<t>
   This directive uses the quoted-string form of the argument syntax.
   A sender &SHOULD-NOT; generate the token form (even if quoting appears not
   to be needed for single-entry lists).
</t>
<aside><t>
   &Note; The
   qualified form of the directive is often handled by caches as if an
   unqualified no-cache directive was received; that is, the special handling
   for the qualified form is not widely implemented.
</t></aside>
</section>

<section title="no-store" anchor="cache-response-directive.no-store">
   <iref item="no-store (cache directive)" primary="true" />
<t>
   The no-store response directive indicates that a cache &MUST-NOT; store
   any part of either the immediate request or the response and &MUST-NOT; use
   the response to satisfy any other request.
</t>
<t>
   This directive applies to both private and shared caches. "MUST NOT
   store" in this context means that the cache &MUST-NOT; intentionally store
   the information in non-volatile storage and &MUST; make a best-effort
   attempt to remove the information from volatile storage as promptly as
   possible after forwarding it.
</t>
<t>
   This directive is not a reliable or sufficient mechanism for ensuring
   privacy. In particular, malicious or compromised caches might not
   recognize or obey this directive, and communications networks might be
   vulnerable to eavesdropping.
</t>
<t>
   Note that the must-understand cache directive overrides no-store in certain
   circumstances; see <xref target="cache-response-directive.must-understand"/>.
</t>
</section>

<section title="no-transform" anchor="cache-response-directive.no-transform">
   <iref item="no-transform (cache directive)" primary="true" />
<t>
   The no-transform response directive indicates that an intermediary
   (regardless of whether it implements a cache) &MUST-NOT; transform the
   content, as defined in <xref target="message.transformations"/>.
</t>
</section>

<section title="private" anchor="cache-response-directive.private">
   <iref item="private (cache directive)" primary="true" />
<t>
   Argument syntax:
</t>
<ul empty="true">
  <li>#<x:ref>field-name</x:ref></li>
</ul>
<t>
   The unqualified private response directive indicates that
   a shared cache &MUST-NOT; store the response (i.e., the response is
   intended for a single user).
   It also indicates that a private cache &MAY; store the response, subject
   to the constraints defined in <xref target="response.cacheability"/>, even if
   the response would not otherwise be heuristically cacheable by a private
   cache.
</t>
<t>
   If a qualified private response directive is present, with an argument that
   lists one or more field names, then only the listed header fields are limited to a
   single user: a shared cache &MUST-NOT; store the listed header fields if they
   are present in the original response but &MAY; store the remainder of the
   response message without those header fields, subject
   the constraints defined in <xref target="response.cacheability"/>.
</t>
<t>
   The field names given are not limited to the set of header
   fields defined by this specification. Field names are case-insensitive.
</t>
<t>
   This directive uses the quoted-string form of the argument syntax.
   A sender &SHOULD-NOT; generate the token form (even if quoting appears not
   to be needed for single-entry lists).
</t>
<aside><t>
   &Note; This usage of the word "private" only controls
   where the response can be stored; it cannot ensure the privacy of the
   message content. Also, the qualified form of the directive is
   often handled by caches as if an unqualified private directive
   was received; that is, the special handling for the qualified form is not
   widely implemented.
</t></aside>
</section>

<section title="proxy-revalidate" anchor="cache-response-directive.proxy-revalidate">
   <iref item="proxy-revalidate (cache directive)" primary="true" />
<t>
   The proxy-revalidate response directive indicates that once the response
   has become stale, a shared cache &MUST-NOT; reuse that response to satisfy
   another request until it has been successfully validated by the origin,
   as defined by <xref target="validation.model"/>. This is analogous to
   must-revalidate (<xref target="cache-response-directive.must-revalidate"/>),
   except that proxy-revalidate does not apply to private caches.
</t>
<t>
   Note that proxy-revalidate on its own does not imply that a response is
   cacheable. For example, it might be combined with the public directive
   (<xref target="cache-response-directive.public"/>), allowing the response
   to be cached while requiring only a shared cache to revalidate when stale.
</t>
</section>

<section title="public" anchor="cache-response-directive.public">
   <iref item="public (cache directive)" primary="true" />
<t>
   The public response directive indicates that a cache &MAY; store the
   response even if it would otherwise be prohibited, subject to the
   constraints defined in <xref target="response.cacheability"/>. In other words,
   public explicitly marks the response as cacheable. For example,
   public permits a shared cache to reuse a response to a request containing
   an Authorization header field (<xref target="caching.authenticated.responses"/>).
</t>
<t>
   Note that it is unnecessary to add the public directive to a response that
   is already cacheable according to <xref target="response.cacheability"/>.
</t>
<t>
   If a response with the public directive has no explicit freshness information,
   it is heuristically cacheable (<xref target="heuristic.freshness"/>).
</t>
</section>

<section title="s-maxage" anchor="cache-response-directive.s-maxage">
   <iref item="s-maxage (cache directive)" primary="true" />
<t>
   Argument syntax:
</t>
<ul empty="true">
  <li><x:ref>delta-seconds</x:ref> (see <xref target="delta-seconds"/>)</li>
</ul>
<t>
   The s-maxage response directive indicates that, for a shared cache, the
   maximum age specified by this directive overrides the maximum age
   specified by either the max-age directive or the <x:ref>Expires</x:ref>
   header field.
</t>
<t>
   The s-maxage directive incorporates the semantics of the
   proxy-revalidate response directive
   (<xref target="cache-response-directive.proxy-revalidate"/>)
   for a shared cache.
   A shared cache &MUST-NOT; reuse a stale response with s-maxage to satisfy
   another request until it has been successfully validated by the origin, as
   defined by <xref target="validation.model"/>.
   This directive also permits a shared cache to reuse a response to a
   request containing an Authorization header field, subject to the above
   requirements on maximum age and revalidation
   (<xref target="caching.authenticated.responses"/>).
</t>
<t>
   This directive uses the token form of the argument syntax:
   e.g., 's-maxage=10' not 's-maxage="10"'. A sender &MUST-NOT; generate
   the quoted-string form.
</t>
</section>
</section>

<section title="Extension Directives" anchor="cache.control.extensions">
<t>
   The Cache-Control header field can be extended through the use of one or
   more extension cache directives.
   A cache &MUST; ignore unrecognized cache directives.
</t>
<t>
   Informational extensions (those that do not require a change in cache
   behavior) can be added without changing the semantics of other directives.
</t>
<t>
   Behavioral extensions are designed to work by acting as modifiers to the
   existing base of cache directives.
   Both the new directive and the old directive are supplied, such that
   applications that do not understand the new directive will default to the
   behavior specified by the old directive, and those that understand the
   new directive will recognize it as modifying the requirements associated
   with the old directive. In this way, extensions to the existing
   cache directives can be made without breaking deployed caches.
</t>
<t>
   For example, consider a hypothetical new response directive called
   "community" that acts as a modifier to the private directive: in addition
   to private caches, only a cache that is shared by members of the named
   community is allowed to cache the response. An origin server wishing to
   allow the UCI community to use an otherwise private response in their
   shared cache(s) could do so by including
</t>
<sourcecode type="http-message">
Cache-Control: private, community="UCI"
</sourcecode>
<t>
   A cache that recognizes such a community cache directive could broaden its
   behavior in accordance with that extension.  A cache that does not
   recognize the community cache directive would ignore it and adhere to the
   private directive.
</t>
<t>
   New extension directives ought to consider defining:
</t>
<ul>
  <li>What it means for a directive to be specified multiple times,</li>
  <li>When the directive does not take an argument, what it means when an
  argument is present,</li>
  <li>When the directive requires an argument, what it means when it is
  missing, and</li>
  <li>Whether the directive is specific to requests, specific to responses, or able
  to be used in either.</li>
</ul>
</section>

<section title="Cache Directive Registry" anchor="cache.directive.registry">
<t>
   The "Hypertext Transfer Protocol (HTTP) Cache Directive Registry" defines the namespace for the
   cache directives. It has been created and is now maintained at
   <eref target="https://www.iana.org/assignments/http-cache-directives"/>.
</t>
<t>
   A registration &MUST; include the following fields:
</t>
<ul>
  <li>Cache Directive Name</li>
  <li>Pointer to specification text</li>
</ul>
<t>
   Values to be added to this namespace require IETF Review (see <xref
   target="RFC8126" x:fmt="," x:sec="4.8"/>).
</t>
</section>
</section>

<section title="Expires" anchor="field.expires">
  <x:anchor-alias value="header.expires"/>
   <iref primary="true" item="Fields" subitem="Expires" x:for-anchor=""/><iref primary="true" item="Header Fields" subitem="Expires" x:for-anchor=""/><iref primary="true" item="Fields" subitem="Expires" x:for-anchor=""/><iref primary="true" item="Header Fields" subitem="Expires" x:for-anchor=""/><iref item="Expires header field" primary="true" x:for-anchor=""/>
   <x:anchor-alias value="Expires"/>
<t>
   The "Expires" response header field gives the date/time after which the
   response is considered stale. See <xref target="expiration.model" /> for
   further discussion of the freshness model.
</t>
<t>
   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.
</t>
<t>
   The Expires field value is an HTTP-date timestamp, as defined in <xref target="http.date"/>.
   See also <xref target="expiration.model"/> for parsing requirements specific to caches.
</t>
<sourcecode type="abnf7230"><iref primary="true" item="Grammar" subitem="Expires"/>
  <x:ref>Expires</x:ref> = <x:ref>HTTP-date</x:ref>
</sourcecode>
<t>
   For example
</t>
<sourcecode type="http-message">
Expires: Thu, 01 Dec 1994 16:00:00 GMT
</sourcecode>
<t>
   A cache recipient &MUST; interpret invalid date formats, especially the
   value "0", as representing a time in the past (i.e., "already expired").
</t>
<t>
   If a response includes a <x:ref>Cache-Control</x:ref> header field with
   the max-age directive (<xref target="cache-response-directive.max-age"/>),
   a recipient &MUST; ignore the Expires header field.
   Likewise, if a response includes the s-maxage directive
   (<xref target="cache-response-directive.s-maxage" />), 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.
</t>
<t>
   An origin server without a clock (<xref target="http.date"/>)
   &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.
</t>
<t>
   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.
</t>
</section>

<section title="Pragma" anchor="field.pragma">
  <x:anchor-alias value="header.pragma"/>
   <iref primary="true" item="Fields" subitem="Pragma" x:for-anchor=""/><iref primary="true" item="Header Fields" subitem="Pragma" x:for-anchor=""/><iref primary="true" item="Fields" subitem="Pragma" x:for-anchor=""/><iref primary="true" item="Header Fields" subitem="Pragma" x:for-anchor=""/><iref item="Pragma header field" primary="true" x:for-anchor=""/>
   <x:anchor-alias value="extension-pragma"/>
   <x:anchor-alias value="Pragma"/>
   <x:anchor-alias value="pragma-directive"/>
   <rdf:Description>
     <status xmlns="urn:ietf:id:draft-ietf-httpbis-p2-semantics#">deprecated</status>
   </rdf:Description>
<t>
   The "Pragma" request header field was defined for HTTP/1.0 caches, so that clients
   could specify a "no-cache" request (as <x:ref>Cache-Control</x:ref> was
   not defined until HTTP/1.1).
</t>
<t>
   However, support for Cache-Control is now widespread. As a result, this
   specification deprecates Pragma.
</t>

<aside>
   <t>
      &Note; Because the meaning of "Pragma: no-cache" in responses was never
      specified, it does not provide a reliable replacement for
      "Cache-Control: no-cache" in them.
   </t>
</aside>
</section>

<section title="Warning" anchor="field.warning">
  <x:anchor-alias value="header.warning"/>
   <rdf:Description>
     <status xmlns="urn:ietf:id:draft-ietf-httpbis-p2-semantics#">obsoleted</status>
   </rdf:Description>
   <iref primary="true" item="Fields" subitem="Warning" x:for-anchor=""/><iref primary="true" item="Header Fields" subitem="Warning" x:for-anchor=""/><iref item="Warning header field" primary="true" x:for-anchor=""/>
   <x:anchor-alias value="Warning"/>
<t>
   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 <x:ref>Age</x:ref>.
</t>
</section>
</section>

<section title="Relationship to Applications and Other Caches" anchor="history.lists">
<t>
   Applications using HTTP often specify additional forms of caching. For
   example, Web browsers often have history mechanisms such as "Back" buttons
   that can be used to redisplay a representation retrieved earlier in a
   session.
</t>
<t>
   Likewise, some Web browsers implement caching of images and other assets
   within a page view; they may or may not honor HTTP caching semantics.
</t>
<t>
   The requirements in this specification do not necessarily apply to how
   applications use data after it is retrieved from an HTTP cache. For example, a
   history mechanism can display a previous representation even if it has
   expired, and an application can use cached data in other ways beyond its
   freshness lifetime.
</t>
<t>
   This specification does not prohibit the application from taking HTTP caching into
   account; for example, a history mechanism might tell the user that a view
   is stale, or it might honor cache directives (e.g., Cache-Control:
   no-store).
</t>
<t>
   However, when an application caches data and does not make this
   apparent to or easily controllable by the user, it is strongly encouraged to
   define its operation with respect to HTTP cache directives so as
   not to surprise authors who expect caching semantics
   to be honored. For example, while it might be reasonable to define an
   application cache "above" HTTP that allows a response containing
   Cache-Control: no-store to be reused for requests that are directly related
   to the request that fetched it (such as those created during the same page
   load), it would likely be surprising and confusing to users and authors if it
   were allowed to be reused for requests unrelated in any way to the one from
   which it was obtained.
</t>
</section>


<section title="Security Considerations" anchor="security.considerations">
<t>
   This section is meant to inform developers, information providers, and
   users of known security concerns specific to HTTP caching.
   More general security considerations are addressed in "HTTP/1.1"
   (<xref target="HTTP11" x:rel="#security.considerations"/>)
   and "HTTP Semantics"
   (<xref target="HTTP" x:rel="#security.considerations"/>).
</t>
<t>
   Caches expose an additional attack surface because the contents of
   the cache represent an attractive target for malicious exploitation.
   Since cache contents persist after an HTTP request is complete, an attack
   on the cache can reveal information long after a user believes that the
   information has been removed from the network. Therefore, cache contents
   need to be protected as sensitive information.
</t>
<t>
   In particular, because private caches are restricted to a single user,
   they can be used to reconstruct a user's activity. As a result, it is
   important for user agents to allow end users to control them, for example,
   by allowing stored responses to be removed for some or all origin servers.
</t>

<section title="Cache Poisoning" anchor="cache.poisoning">
<t>
   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.
</t>
<t>
   One common attack vector for cache poisoning is to exploit differences in
   message parsing on proxies and in user agents; see <xref
   target="message.body.length"/> for the relevant requirements regarding
   HTTP/1.1.
</t>
</section>

<section title="Timing Attacks" anchor="security.timing">
<t>
   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.
</t>
<t>
   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.
</t>
<t>
   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".
</t>
</section>

<section title="Caching of Sensitive Information" anchor="caching.of.sensitive.information">
<t>
   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.
</t>
<t>
   Note that the Set-Cookie response header field <xref target="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.
</t>
</section>
</section>


<section title="IANA Considerations" anchor="iana.considerations">
<t>
   The change controller for the following registrations is:
   "IETF (iesg@ietf.org) - Internet Engineering Task Force".
</t>

<section title="Field Name Registration" anchor="field.name.registration">
<t>
   IANA has updated the "Hypertext Transfer Protocol (HTTP) Field
   Name Registry" at <eref target="https://www.iana.org/assignments/http-fields"/>,
   as described in <xref target="HTTP" x:rel="#field.name.registration"/>,
   with the field names listed in the table below:
</t>
<?BEGININC build/draft-ietf-httpbis-cache-latest.iana-headers ?>
<!--AUTOGENERATED FROM extract-header-defs.xslt, do not edit manually-->
<table align="left" anchor="iana.header.registration.table">
   <thead>
      <tr>
         <th>Field Name</th>
         <th>Status</th>
         <th>Section</th>
         <th>Comments</th>
      </tr>
   </thead>
   <tbody>
      <tr>
         <td>Age</td>
         <td>permanent</td>
         <td>
            <xref target="field.age" format="counter"/>
         </td>
         <td/>
      </tr>
      <tr>
         <td>Cache-Control</td>
         <td>permanent</td>
         <td>
            <xref target="field.cache-control" format="counter"/>
         </td>
         <td/>
      </tr>
      <tr>
         <td>Expires</td>
         <td>permanent</td>
         <td>
            <xref target="field.expires" format="counter"/>
         </td>
         <td/>
      </tr>
      <tr>
         <td>Pragma</td>
         <td>deprecated</td>
         <td>
            <xref target="field.pragma" format="counter"/>
         </td>
         <td/>
      </tr>
      <tr>
         <td>Warning</td>
         <td>obsoleted</td>
         <td>
            <xref target="field.warning" format="counter"/>
         </td>
         <td/>
      </tr>
   </tbody>
</table>
<!--(END)-->
<?ENDINC build/draft-ietf-httpbis-cache-latest.iana-headers ?>
</section>

<section title="Cache Directive Registration" anchor="cache.directive.registration">
<t>
   IANA has updated the
   "Hypertext Transfer Protocol (HTTP) Cache Directive Registry"
   at <eref target="https://www.iana.org/assignments/http-cache-directives"/>
   with the registration procedure per <xref target="cache.directive.registry"/>
   and the cache directive names summarized in the table below.
</t>
<?BEGININC build/draft-ietf-httpbis-cache-latest.cache-directives ?>
<!--AUTOGENERATED FROM extract-cache-directives-defs.xslt, do not edit manually-->
<table align="left" anchor="iana.cache.directive.registration.table">
   <thead>
      <tr>
         <th>Cache Directive</th>
         <th>Section</th>
      </tr>
   </thead>
   <tbody>
      <tr>
         <td>max-age</td>
         <td>
            <xref target="cache-request-directive.max-age" format="counter"/>, <xref target="cache-response-directive.max-age" format="counter"/>
         </td>
      </tr>
      <tr>
         <td>max-stale</td>
         <td>
            <xref target="cache-request-directive.max-stale" format="counter"/>
         </td>
      </tr>
      <tr>
         <td>min-fresh</td>
         <td>
            <xref target="cache-request-directive.min-fresh" format="counter"/>
         </td>
      </tr>
      <tr>
         <td>must-revalidate</td>
         <td>
            <xref target="cache-response-directive.must-revalidate" format="counter"/>
         </td>
      </tr>
      <tr>
         <td>must-understand</td>
         <td>
            <xref target="cache-response-directive.must-understand" format="counter"/>
         </td>
      </tr>
      <tr>
         <td>no-cache</td>
         <td>
            <xref target="cache-request-directive.no-cache" format="counter"/>, <xref target="cache-response-directive.no-cache" format="counter"/>
         </td>
      </tr>
      <tr>
         <td>no-store</td>
         <td>
            <xref target="cache-request-directive.no-store" format="counter"/>, <xref target="cache-response-directive.no-store" format="counter"/>
         </td>
      </tr>
      <tr>
         <td>no-transform</td>
         <td>
            <xref target="cache-request-directive.no-transform" format="counter"/>, <xref target="cache-response-directive.no-transform" format="counter"/>
         </td>
      </tr>
      <tr>
         <td>only-if-cached</td>
         <td>
            <xref target="cache-request-directive.only-if-cached" format="counter"/>
         </td>
      </tr>
      <tr>
         <td>private</td>
         <td>
            <xref target="cache-response-directive.private" format="counter"/>
         </td>
      </tr>
      <tr>
         <td>proxy-revalidate</td>
         <td>
            <xref target="cache-response-directive.proxy-revalidate" format="counter"/>
         </td>
      </tr>
      <tr>
         <td>public</td>
         <td>
            <xref target="cache-response-directive.public" format="counter"/>
         </td>
      </tr>
      <tr>
         <td>s-maxage</td>
         <td>
            <xref target="cache-response-directive.s-maxage" format="counter"/>
         </td>
      </tr>
   </tbody>
</table>
<!--(END)-->
<?ENDINC build/draft-ietf-httpbis-cache-latest.cache-directives ?>
</section>

<section title="Warn Code Registry" anchor="warn.code.registration">
<t>
   IANA has added the following note to the "Hypertext Transfer Protocol (HTTP) Warn Codes"
   registry at <eref target="https://www.iana.org/assignments/http-warn-codes"/>
   stating that "Warning" has been obsoleted:
</t>
<blockquote><t>
   The Warning header field (and the warn codes that it uses) has been obsoleted
   for HTTP per [RFC9111].
</t></blockquote>
</section>
</section>
</middle>

<back>
<displayreference target="HTTP11" to="HTTP/1.1"/>

<references title="Normative References">

  <reference anchor="HTTP">
    <x:source basename="draft-ietf-httpbis-semantics-latest" href="draft-ietf-httpbis-semantics-latest.xml">
      <x:defines>Content-Length</x:defines>
      <x:has anchor="GET"/>
      <x:has anchor="abnf.extension"/>
      <x:has anchor="combining.byte.ranges"/>
      <x:has anchor="conditional.requests"/>
      <x:has anchor="field.authorization"/>
      <x:has anchor="field.connection"/>
      <x:has anchor="field.date"/>
      <x:has anchor="field.etag"/>
      <x:has anchor="field.if-modified-since"/>
      <x:has anchor="field.if-none-match"/>
      <x:has anchor="field.if-range"/>
      <x:has anchor="field.last-modified"/>
      <x:has anchor="field.lines"/>
      <x:has anchor="field.proxy-authenticate"/>
      <x:has anchor="field.proxy-authentication-info"/>
      <x:has anchor="field.proxy-authorization"/>
      <x:has anchor="field.range"/>
      <x:has anchor="field.vary"/>
      <x:has anchor="http.date"/>
      <x:has anchor="intermediaries"/>
      <x:has anchor="fields.names"/>
      <x:has anchor="message.framing"/>
      <x:has anchor="message.transformations"/>
      <x:has anchor="origin"/>
      <x:has anchor="overview.of.status.codes"/>
      <x:has anchor="precedence"/>
      <x:has anchor="quoted.strings"/>
      <x:has anchor="safe.methods"/>
      <x:has anchor="status.codes"/>
      <x:has anchor="target.resource"/>
      <x:has anchor="tokens"/>
      <x:has anchor="weak.and.strong.validators"/>
      <x:has anchor="whitespace"/>
    </x:source>
  </reference>

  <reference anchor="RFC2119">
    <front>
      <title>Key words for use in RFCs to Indicate Requirement Levels</title>
      <author initials="S." surname="Bradner" fullname="Scott Bradner"/>
      <date month="March" year="1997"/>
    </front>
    <seriesInfo name="BCP" value="14"/>
    <seriesInfo name="RFC" value="2119"/>
  </reference>

  <reference anchor="RFC8174">
    <front>
      <title>Ambiguity of Uppercase vs Lowercase in RFC 2119 Key Words</title>
      <author initials='B.' surname='Leiba' fullname='Barry Leiba'/>
      <date year="2017" month="May"/>
    </front>
    <seriesInfo name='BCP' value='14'/>
    <seriesInfo name='RFC' value='8174'/>
  </reference>

  <reference anchor="RFC5234">
    <front>
      <title abbrev="ABNF for Syntax Specifications">Augmented BNF for Syntax Specifications: ABNF</title>
      <author initials="D." surname="Crocker" fullname="Dave Crocker" role="editor"/>
      <author initials="P." surname="Overell" fullname="Paul Overell"/>
      <date month="January" year="2008"/>
    </front>
    <seriesInfo name="STD" value="68"/>
    <seriesInfo name="RFC" value="5234"/>
  </reference>

  <reference anchor="RFC7405">
    <front>
      <title>Case-Sensitive String Support in ABNF</title>
      <author initials="P." surname="Kyzivat" fullname="Dave Kyzivat"/>
      <date month="December" year="2014"/>
    </front>
    <seriesInfo name="RFC" value="7405"/>
  </reference>

</references>

<references title="Informative References">

  <reference anchor="HTTP11">
    <x:source basename="draft-ietf-httpbis-messaging-latest" href="draft-ietf-httpbis-messaging-latest.xml">
      <x:has anchor="message.body.length"/>
    </x:source>
  </reference>

  <reference anchor="RFC2616">
    <front>
      <title>Hypertext Transfer Protocol -- HTTP/1.1</title>
      <author fullname="R. Fielding" initials="R." surname="Fielding"/>
      <author fullname="J. Gettys" initials="J." surname="Gettys"/>
      <author fullname="J. Mogul" initials="J." surname="Mogul"/>
      <author fullname="H. Frystyk" initials="H." surname="Frystyk"/>
      <author fullname="L. Masinter" initials="L." surname="Masinter"/>
      <author fullname="P. Leach" initials="P." surname="Leach"/>
      <author fullname="T. Berners-Lee" initials="T." surname="Berners-Lee"/>
      <date month="June" year="1999" />
    </front>
    <seriesInfo name="RFC" value="2616" />
  </reference>

  <reference anchor="RFC7234">
    <front>
      <title>Hypertext Transfer Protocol (HTTP/1.1): Caching</title>
      <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor"/>
      <author initials="M." surname="Nottingham" fullname="Mark Nottingham" role="editor"/>
      <author initials="J." surname="Reschke" fullname="Julian F. Reschke" role="editor"/>
      <date month="June" year="2014"/>
    </front>
    <seriesInfo name="RFC" value="7234"/>
  </reference>

  <reference anchor="RFC5861">
    <front>
      <title abbrev="HTTP stale controls">HTTP Cache-Control Extensions for Stale Content</title>
      <author initials="M." surname="Nottingham" fullname="Mark Nottingham"/>
      <date month="May" year="2010"/>
    </front>
    <seriesInfo name='RFC' value='5861' />
  </reference>

  <reference anchor="COOKIE">
    <front>
      <title>HTTP State Management Mechanism</title>
      <author initials="A." surname="Barth" fullname="Adam Barth"/>
      <date year="2011" month="April" />
    </front>
    <seriesInfo name="RFC" value="6265"/>
  </reference>

  <reference anchor="RFC8126">
    <front>
      <title>Guidelines for Writing an IANA Considerations Section in RFCs</title>
      <author initials="M." surname="Cotton" fullname="M. Cotton"/>
      <author initials="B." surname="Leiba" fullname="B. Leiba"/>
      <author initials="T." surname="Narten" fullname="T. Narten"/>
      <date year="2017" month="June" />
    </front>
    <seriesInfo name="BCP" value="26"/>
    <seriesInfo name="RFC" value="8126"/>
  </reference>

</references>

<?BEGININC build/draft-ietf-httpbis-cache-latest.abnf-appendix ?>
<section xmlns:x="http://purl.org/net/xml2rfc/ext" title="Collected ABNF" anchor="collected.abnf"><t>In the collected ABNF below, list rules are expanded per <xref target="HTTP" x:rel="#abnf.extension"/>.</t><sourcecode type="abnf" name="rfc9111.parsed-abnf">
<x:ref>Age</x:ref> = delta-seconds

<x:ref>Cache-Control</x:ref> = [ cache-directive *( OWS "," OWS cache-directive ) ]

<x:ref>Expires</x:ref> = HTTP-date

<x:ref>HTTP-date</x:ref> = &lt;HTTP-date, see <xref target="HTTP" x:fmt="," x:sec="5.6.7"/>&gt;

<x:ref>OWS</x:ref> = &lt;OWS, see <xref target="HTTP" x:fmt="," x:sec="5.6.3"/>&gt;

<x:ref>cache-directive</x:ref> = token [ "=" ( token / quoted-string ) ]

<x:ref>delta-seconds</x:ref> = 1*DIGIT

<x:ref>field-name</x:ref> = &lt;field-name, see <xref target="HTTP" x:fmt="," x:sec="5.1"/>&gt;

<x:ref>quoted-string</x:ref> = &lt;quoted-string, see <xref target="HTTP" x:fmt="," x:sec="5.6.4"/>&gt;

<x:ref>token</x:ref> = &lt;token, see <xref target="HTTP" x:fmt="," x:sec="5.6.2"/>&gt;
</sourcecode>
</section>
<?ENDINC build/draft-ietf-httpbis-cache-latest.abnf-appendix ?>

<section title="Changes from RFC 7234" anchor="changes.from.rfc.7234">
<t>
   Handling of duplicate and conflicting cache directives has been clarified.
   (<xref target="calculating.freshness.lifetime"/>)
</t>
<t>
   Cache invalidation of the URIs in the Location and Content-Location
   header fields is no longer required but is still allowed.
   (<xref target="invalidation"/>)
</t>
<t>
   Cache invalidation of the URIs in the Location and Content-Location header fields is disallowed
   when the origin is different; previously, it was the host.
   (<xref target="invalidation"/>)
</t>
<t>
   Handling invalid and multiple Age header field values has been clarified.
   (<xref target="field.age"/>)
</t>
<t>
   Some cache directives defined by this specification now have stronger
   prohibitions against generating the quoted form of their values, since
   this has been found to create interoperability problems. Consumers of
   extension cache directives are no longer required to accept both token and
   quoted-string forms, but they still need to parse them properly for
   unknown extensions.
   (<xref target="field.cache-control"/>)
</t>
<t>
   The public and private cache directives were clarified, so that they
   do not make responses reusable under any condition.
   (<xref target="cache-response-directive"/>)
</t>
<t>
   The must-understand cache directive was introduced; caches are no
   longer required to understand the semantics of new response status codes
   unless it is present.
   (<xref target="cache-response-directive.must-understand"/>)
</t>
<t>
   The Warning response header was obsoleted. Much of the information
   supported by Warning could be gleaned by examining the response, and the
   remaining information &mdash; although potentially useful &mdash; was entirely
   advisory. In practice, Warning was not added by caches or intermediaries.
   (<xref target="field.warning"/>)
</t>
</section>

<section title="Change Log" anchor="change.log" removeInRFC="true">
<t>
  See <eref target="https://www.ietf.org/archive/id/draft-ietf-httpbis-cache-19.html#appendix-C"/>
  for changes up to version 19 of this document.
</t>

<section title="Since draft-ietf-httpbis-cache-19" anchor="changes.since.19">
<t>
  This (unpublished) draft contains changes that were made after draft 19
  was approved by the IESG. Most changes are editorial only.<!-- Furthermore:-->
</t>
<ul x:when-empty="None.">
</ul>
</section>
</section>

<section title="Acknowledgements" anchor="acks" numbered="false">
<t>
  See <xref target="HTTP" x:rel="#acks"/>, which applies to this document as well.
</t>
</section>
</back>
</rfc>