draft-ietf-httpbis-priority-06.txt   draft-ietf-httpbis-priority-07.txt 
HTTP K. Oku HTTP K. Oku
Internet-Draft Fastly Internet-Draft Fastly
Intended status: Standards Track L. Pardue Intended status: Standards Track L. Pardue
Expires: 4 April 2022 Cloudflare Expires: 28 April 2022 Cloudflare
1 October 2021 25 October 2021
Extensible Prioritization Scheme for HTTP Extensible Prioritization Scheme for HTTP
draft-ietf-httpbis-priority-06 draft-ietf-httpbis-priority-07
Abstract Abstract
This document describes a scheme for prioritizing HTTP responses. This document describes a scheme that allows an HTTP client to
This scheme expresses the priority of each HTTP response using communicate its preferences for how the upstream server prioritizes
absolute values, rather than as a relative relationship between a responses to its requests, and also allows a server to hint to a
group of HTTP responses. downstream intermediary how its responses should be prioritized when
they are forwarded. This document defines the Priority header field
This document defines the Priority header field for communicating the for communicating the initial priority in an HTTP version-independent
initial priority in an HTTP version-independent manner, as well as manner, as well as HTTP/2 and HTTP/3 frames for reprioritizing
HTTP/2 and HTTP/3 frames for reprioritizing the responses. These responses. These share a common format structure that is designed to
share a common format structure that is designed to provide future provide future extensibility.
extensibility.
Note to Readers Note to Readers
_RFC EDITOR: please remove this section before publication_ _RFC EDITOR: please remove this section before publication_
Discussion of this draft takes place on the HTTP working group Discussion of this draft takes place on the HTTP working group
mailing list (ietf-http-wg@w3.org), which is archived at mailing list (ietf-http-wg@w3.org), which is archived at
https://lists.w3.org/Archives/Public/ietf-http-wg/ https://lists.w3.org/Archives/Public/ietf-http-wg/
(https://lists.w3.org/Archives/Public/ietf-http-wg/). (https://lists.w3.org/Archives/Public/ietf-http-wg/).
skipping to change at page 2, line 10 skipping to change at page 2, line 10
Internet-Drafts are working documents of the Internet Engineering Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF). Note that other groups may also distribute Task Force (IETF). Note that other groups may also distribute
working documents as Internet-Drafts. The list of current Internet- working documents as Internet-Drafts. The list of current Internet-
Drafts is at https://datatracker.ietf.org/drafts/current/. Drafts is at https://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any and may be updated, replaced, or obsoleted by other documents at any
time. It is inappropriate to use Internet-Drafts as reference time. It is inappropriate to use Internet-Drafts as reference
material or to cite them other than as "work in progress." material or to cite them other than as "work in progress."
This Internet-Draft will expire on 4 April 2022. This Internet-Draft will expire on 28 April 2022.
Copyright Notice Copyright Notice
Copyright (c) 2021 IETF Trust and the persons identified as the Copyright (c) 2021 IETF Trust and the persons identified as the
document authors. All rights reserved. document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents (https://trustee.ietf.org/ Provisions Relating to IETF Documents (https://trustee.ietf.org/
license-info) in effect on the date of publication of this document. license-info) in effect on the date of publication of this document.
Please review these documents carefully, as they describe your rights Please review these documents carefully, as they describe your rights
and restrictions with respect to this document. Code Components and restrictions with respect to this document. Code Components
extracted from this document must include Simplified BSD License text extracted from this document must include Simplified BSD License text
as described in Section 4.e of the Trust Legal Provisions and are as described in Section 4.e of the Trust Legal Provisions and are
provided without warranty as described in the Simplified BSD License. provided without warranty as described in the Simplified BSD License.
Table of Contents Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1. Notational Conventions . . . . . . . . . . . . . . . . . 4 1.1. Notational Conventions . . . . . . . . . . . . . . . . . 4
2. Motivation for Replacing RFC 7540 Priorities . . . . . . . . 4 2. Motivation for Replacing RFC 7540 Priorities . . . . . . . . 5
2.1. Disabling RFC 7540 Priorities . . . . . . . . . . . . . . 6 2.1. Disabling RFC 7540 Priorities . . . . . . . . . . . . . . 6
2.1.1. Advice when Using Extensible Priorities as the 2.1.1. Advice when Using Extensible Priorities as the
Alternative . . . . . . . . . . . . . . . . . . . . . 7 Alternative . . . . . . . . . . . . . . . . . . . . . 7
3. Applicability of the Extensible Priority Scheme . . . . . . . 7 3. Applicability of the Extensible Priority Scheme . . . . . . . 7
4. Priority Parameters . . . . . . . . . . . . . . . . . . . . . 7 4. Priority Parameters . . . . . . . . . . . . . . . . . . . . . 8
4.1. Urgency . . . . . . . . . . . . . . . . . . . . . . . . . 8 4.1. Urgency . . . . . . . . . . . . . . . . . . . . . . . . . 8
4.2. Incremental . . . . . . . . . . . . . . . . . . . . . . . 8 4.2. Incremental . . . . . . . . . . . . . . . . . . . . . . . 9
4.3. Defining New Parameters . . . . . . . . . . . . . . . . . 9 4.3. Defining New Parameters . . . . . . . . . . . . . . . . . 10
4.3.1. Registration . . . . . . . . . . . . . . . . . . . . 9 4.3.1. Registration . . . . . . . . . . . . . . . . . . . . 10
5. The Priority HTTP Header Field . . . . . . . . . . . . . . . 10 5. The Priority HTTP Header Field . . . . . . . . . . . . . . . 11
6. Reprioritization . . . . . . . . . . . . . . . . . . . . . . 11 6. Reprioritization . . . . . . . . . . . . . . . . . . . . . . 12
7. The PRIORITY_UPDATE Frame . . . . . . . . . . . . . . . . . . 11 7. The PRIORITY_UPDATE Frame . . . . . . . . . . . . . . . . . . 12
7.1. HTTP/2 PRIORITY_UPDATE Frame . . . . . . . . . . . . . . 12 7.1. HTTP/2 PRIORITY_UPDATE Frame . . . . . . . . . . . . . . 13
7.2. HTTP/3 PRIORITY_UPDATE Frame . . . . . . . . . . . . . . 13 7.2. HTTP/3 PRIORITY_UPDATE Frame . . . . . . . . . . . . . . 14
8. Merging Client- and Server-Driven Parameters . . . . . . . . 14 8. Merging Client- and Server-Driven Parameters . . . . . . . . 15
9. Client Scheduling . . . . . . . . . . . . . . . . . . . . . . 15 9. Client Scheduling . . . . . . . . . . . . . . . . . . . . . . 16
10. Server Scheduling . . . . . . . . . . . . . . . . . . . . . . 15 10. Server Scheduling . . . . . . . . . . . . . . . . . . . . . . 17
10.1. Intermediaries with Multiple Backend Connections . . . . 17 10.1. Intermediaries with Multiple Backend Connections . . . . 18
11. Scheduling and the CONNECT Method . . . . . . . . . . . . . . 17 11. Scheduling and the CONNECT Method . . . . . . . . . . . . . . 19
12. Retransmission Scheduling . . . . . . . . . . . . . . . . . . 18 12. Retransmission Scheduling . . . . . . . . . . . . . . . . . . 19
13. Fairness . . . . . . . . . . . . . . . . . . . . . . . . . . 18 13. Fairness . . . . . . . . . . . . . . . . . . . . . . . . . . 20
13.1. Coalescing Intermediaries . . . . . . . . . . . . . . . 18 13.1. Coalescing Intermediaries . . . . . . . . . . . . . . . 20
13.2. HTTP/1.x Back Ends . . . . . . . . . . . . . . . . . . . 19 13.2. HTTP/1.x Back Ends . . . . . . . . . . . . . . . . . . . 21
13.3. Intentional Introduction of Unfairness . . . . . . . . . 19 13.3. Intentional Introduction of Unfairness . . . . . . . . . 21
14. Why use an End-to-End Header Field? . . . . . . . . . . . . . 20 14. Why use an End-to-End Header Field? . . . . . . . . . . . . . 21
15. Security Considerations . . . . . . . . . . . . . . . . . . . 20 15. Security Considerations . . . . . . . . . . . . . . . . . . . 22
16. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 21 16. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 22
17. References . . . . . . . . . . . . . . . . . . . . . . . . . 22 17. References . . . . . . . . . . . . . . . . . . . . . . . . . 23
17.1. Normative References . . . . . . . . . . . . . . . . . . 22 17.1. Normative References . . . . . . . . . . . . . . . . . . 23
17.2. Informative References . . . . . . . . . . . . . . . . . 23 17.2. Informative References . . . . . . . . . . . . . . . . . 24
Appendix A. Acknowledgements . . . . . . . . . . . . . . . . . . 24 Appendix A. Acknowledgements . . . . . . . . . . . . . . . . . . 25
Appendix B. Change Log . . . . . . . . . . . . . . . . . . . . . 24 Appendix B. Change Log . . . . . . . . . . . . . . . . . . . . . 25
B.1. Since draft-ietf-httpbis-priority-05 . . . . . . . . . . 24 B.1. Since draft-ietf-httpbis-priority-06 . . . . . . . . . . 25
B.2. Since draft-ietf-httpbis-priority-04 . . . . . . . . . . 25 B.2. Since draft-ietf-httpbis-priority-05 . . . . . . . . . . 26
B.3. Since draft-ietf-httpbis-priority-03 . . . . . . . . . . 25 B.3. Since draft-ietf-httpbis-priority-04 . . . . . . . . . . 26
B.4. Since draft-ietf-httpbis-priority-02 . . . . . . . . . . 25 B.4. Since draft-ietf-httpbis-priority-03 . . . . . . . . . . 26
B.5. Since draft-ietf-httpbis-priority-01 . . . . . . . . . . 25 B.5. Since draft-ietf-httpbis-priority-02 . . . . . . . . . . 26
B.6. Since draft-ietf-httpbis-priority-00 . . . . . . . . . . 25 B.6. Since draft-ietf-httpbis-priority-01 . . . . . . . . . . 26
B.7. Since draft-kazuho-httpbis-priority-04 . . . . . . . . . 26 B.7. Since draft-ietf-httpbis-priority-00 . . . . . . . . . . 27
B.8. Since draft-kazuho-httpbis-priority-03 . . . . . . . . . 26 B.8. Since draft-kazuho-httpbis-priority-04 . . . . . . . . . 27
B.9. Since draft-kazuho-httpbis-priority-02 . . . . . . . . . 26 B.9. Since draft-kazuho-httpbis-priority-03 . . . . . . . . . 27
B.10. Since draft-kazuho-httpbis-priority-01 . . . . . . . . . 26 B.10. Since draft-kazuho-httpbis-priority-02 . . . . . . . . . 27
B.11. Since draft-kazuho-httpbis-priority-00 . . . . . . . . . 26 B.11. Since draft-kazuho-httpbis-priority-01 . . . . . . . . . 27
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 27 B.12. Since draft-kazuho-httpbis-priority-00 . . . . . . . . . 28
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 28
1. Introduction 1. Introduction
It is common for an HTTP [HTTP] resource representation to have It is common for representations of an HTTP [HTTP] resource to have
relationships to one or more other resources. Clients will often relationships to one or more other resources. Clients will often
discover these relationships while processing a retrieved discover these relationships while processing a retrieved
representation, leading to further retrieval requests. Meanwhile, representation, which may lead to further retrieval requests.
the nature of the relationship determines whether the client is Meanwhile, the nature of the relationship determines whether the
blocked from continuing to process locally available resources. For client is blocked from continuing to process locally available
example, visual rendering of an HTML document could be blocked by the resources. An example of this is visual rendering of an HTML
retrieval of a CSS file that the document refers to. In contrast, document, which could be blocked by the retrieval of a CSS file that
inline images do not block rendering and get drawn incrementally as the document refers to. In contrast, inline images do not block
the chunks of the images arrive. rendering and get drawn incrementally as the chunks of the images
arrive.
To provide meaningful presentation of a document at the earliest HTTP/2 [HTTP2] and HTTP/3 [HTTP3] support multiplexing of requests
and responses in a single connection. An important feature of any
implementation of a protocol that provides multiplexing is the
ability to prioritize the sending of information. For example, to
provide meaningful presentation of an HTML document at the earliest
moment, it is important for an HTTP server to prioritize the HTTP moment, it is important for an HTTP server to prioritize the HTTP
responses, or the chunks of those HTTP responses, that it sends. responses, or the chunks of those HTTP responses, that it sends to a
client.
A server that operates in ignorance of how clients issue requests and
consume responses can cause suboptimal client application
performance. Priority signals allow clients to communicate their
view of request priority. Servers have their own needs that are
independent from client needs, so they often combine priority signals
with other available information in order to inform scheduling of
response data.
RFC 7540 [RFC7540] stream priority allowed a client to send a series RFC 7540 [RFC7540] stream priority allowed a client to send a series
of priority signals that communicate to the server a "priority tree"; of priority signals that communicate to the server a "priority tree";
the structure of this tree represents the client's preferred relative the structure of this tree represents the client's preferred relative
ordering and weighted distribution of the bandwidth among HTTP ordering and weighted distribution of the bandwidth among HTTP
responses. Servers could use these priority signals as input into responses. Servers could use these priority signals as input into
prioritization decision making. prioritization decision making.
The design and implementation of RFC 7540 stream priority was The design and implementation of RFC 7540 stream priority was
observed to have shortcomings, explained in Section 2. HTTP/2 observed to have shortcomings, explained in Section 2. HTTP/2
[HTTP2] has consequently deprecated the use of these stream priority [HTTP2] has consequently deprecated the use of these stream priority
signals. signals.
This document describes an extensible scheme for prioritizing HTTP This document describes an extensible scheme for prioritizing HTTP
responses that uses absolute values. Section 4 defines priority responses that uses absolute values. Section 4 defines priority
parameters, which are a standardized and extensible format of parameters, which are a standardized and extensible format of
priority information. Section 5 defines the Priority HTTP header priority information. Section 5 defines the Priority HTTP header
field that can be used by both client and server to exchange field, a protocol-version-independent and end-to-end priority signal.
parameters in order to specify the precedence of HTTP responses in a Clients can use this header to signal priority to servers in order to
protocol-version-independent and end-to-end manner. Section 7.1 and specify the precedence of HTTP responses. Similarly, servers behind
Section 7.2 define version-specific frames that carry parameters for an intermediary can use it to signal priority to the intermediary.
reprioritization. This prioritization scheme and its signals can act Section 7.1 and Section 7.2 define version-specific frames that carry
parameters, which clients can use for reprioritization.
Header field and frame priority signals are input to a server's
response prioritization process. They are only a suggestion and do
not guarantee any particular processing or transmission order for one
response relative to any other response. Section 10 and Section 12
provide consideration and guidance about how servers might act upon
signals.
The prioritization scheme and priority signals defined herein can act
as a substitute for RFC 7540 stream priority. as a substitute for RFC 7540 stream priority.
1.1. Notational Conventions 1.1. Notational Conventions
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in [RFC2119]. document are to be interpreted as described in [RFC2119].
The terms sf-integer and sf-boolean are imported from The terms Dictionary, sf-boolean, sf-dictionary, and sf-integer are
[STRUCTURED-FIELDS]. imported from [STRUCTURED-FIELDS].
Example HTTP requests and responses use the HTTP/2-style formatting Example HTTP requests and responses use the HTTP/2-style formatting
from [HTTP2]. from [HTTP2].
This document uses the variable-length integer encoding from [QUIC]. This document uses the variable-length integer encoding from [QUIC].
The term control stream is used to describe the HTTP/2 stream with The term control stream is used to describe the HTTP/2 stream with
identifier 0x0, and HTTP/3 control stream; see Section 6.2.1 of identifier 0x0, and HTTP/3 control stream; see Section 6.2.1 of
[HTTP3]. [HTTP3].
The term HTTP/2 priority signal is used to describe the priority The term HTTP/2 priority signal is used to describe the priority
information sent from clients to servers in HTTP/2 frames; see information sent from clients to servers in HTTP/2 frames; see
Section 5.3.2 of [HTTP2]. Section 5.3.2 of [HTTP2].
2. Motivation for Replacing RFC 7540 Priorities 2. Motivation for Replacing RFC 7540 Priorities
An important feature of any implementation of a protocol that
provides multiplexing is the ability to prioritize the sending of
information. Prioritization is a difficult problem, so it will
always be suboptimal, particularly if one endpoint operates in
ignorance of the needs of its peer. Priority signalling allows
endpoints to communicate their own view of priority, which can be
combined with information the peer has to inform scheduling.
RFC 7540 stream priority (see Section 5.3 of [RFC7540]) is a complex RFC 7540 stream priority (see Section 5.3 of [RFC7540]) is a complex
system where clients signal stream dependencies and weights to system where clients signal stream dependencies and weights to
describe an unbalanced tree. It suffered from limited deployment and describe an unbalanced tree. It suffered from limited deployment and
interoperability and was deprecated in a revision of HTTP/2 [HTTP2]. interoperability and was deprecated in a revision of HTTP/2 [HTTP2].
However, in order to maintain wire compatibility, HTTP/2 priority However, in order to maintain wire compatibility, HTTP/2 priority
signals are still mandatory to handle (see Section 5.3.2 of [HTTP2]). signals are still mandatory to handle (see Section 5.3.2 of [HTTP2]).
Clients can build RFC 7540 trees with rich flexibility but experience Clients can build RFC 7540 trees with rich flexibility but experience
has shown this is rarely exercised. Instead they tend to choose a has shown this is rarely exercised. Instead they tend to choose a
single model optimized for a single use case and experiment within single model optimized for a single use case and experiment within
the model constraints, or do nothing at all. Furthermore, many the model constraints, or do nothing at all. Furthermore, many
clients build their prioritization tree in a unique way, which makes clients build their prioritization tree in a unique way, which makes
it difficult for servers to understand their intent and act or it difficult for servers to understand their intent and act or
intervene accordingly. intervene accordingly.
Many RFC 7540 server implementations do not act on HTTP/2 priority Many RFC 7540 server implementations do not act on HTTP/2 priority
signals. Some instead favor custom server-driven schemes based on signals. Some instead favor custom server-driven schemes based on
heuristics or other hints, such as resource content type or request heuristics or other hints, such as resource content type or request
generation order. For example, a server, with knowledge of the generation order. For example, a server, with knowledge of an HTML
document structure, might want to prioritize the delivery of images document structure, might want to prioritize the delivery of images
that are critical to user experience above other images, but below that are critical to user experience above other images, but below
the CSS files. Since client trees vary, it is impossible for the the CSS files. Since client trees vary, it is impossible for the
server to determine how such images should be prioritized against server to determine how such images should be prioritized against
other responses. other responses.
RFC 7540 allows intermediaries to coalesce multiple client trees into RFC 7540 allows intermediaries to coalesce multiple client trees into
a single tree that is used for a single upstream HTTP/2 connection. a single tree that is used for a single upstream HTTP/2 connection.
However, most intermediaries do not support this. Additionally, RFC However, most intermediaries do not support this. Additionally, RFC
7540 does not define a method that can be used by a server to express 7540 does not define a method that can be used by a server to express
skipping to change at page 6, line 30 skipping to change at page 7, line 5
independent research have shown that simpler schemes can reach at independent research have shown that simpler schemes can reach at
least equivalent performance characteristics compared to the more least equivalent performance characteristics compared to the more
complex RFC 7540 setups seen in practice, at least for the web use complex RFC 7540 setups seen in practice, at least for the web use
case. case.
2.1. Disabling RFC 7540 Priorities 2.1. Disabling RFC 7540 Priorities
The problems and insights set out above provided the motivation for The problems and insights set out above provided the motivation for
deprecating RFC 7540 stream priority (see Section 5.3 of [RFC7540]). deprecating RFC 7540 stream priority (see Section 5.3 of [RFC7540]).
The SETTINGS_NO_RFC7540_PRIORITIES setting is defined by this The SETTINGS_NO_RFC7540_PRIORITIES HTTP/2 setting is defined by this
document in order to allow endpoints to explicitly opt out of using document in order to allow endpoints to explicitly opt out of using
HTTP/2 priority signals (see Section 5.3.2 of [HTTP2]). Endpoints HTTP/2 priority signals (see Section 5.3.2 of [HTTP2]). Endpoints
are encouraged to use alternative priority signals (for example, are encouraged to use alternative priority signals (for example,
Section 5 or Section 7.1) but there is no requirement to use a Section 5 or Section 7.1) but there is no requirement to use a
specific signal type. specific signal type.
The value of SETTINGS_NO_RFC7540_PRIORITIES MUST be 0 or 1. Any The value of SETTINGS_NO_RFC7540_PRIORITIES MUST be 0 or 1. Any
value other than 0 or 1 MUST be treated as a connection error (see value other than 0 or 1 MUST be treated as a connection error (see
Section 5.4.1 of [HTTP2]) of type PROTOCOL_ERROR. Section 5.4.1 of [HTTP2]) of type PROTOCOL_ERROR. The initial value
is 0.
Endpoints MUST send this SETTINGS parameter as part of the first Endpoints MUST send this SETTINGS parameter as part of the first
SETTINGS frame. A sender MUST NOT change the SETTINGS frame. A sender MUST NOT change the
SETTINGS_NO_RFC7540_PRIORITIES parameter value after the first SETTINGS_NO_RFC7540_PRIORITIES parameter value after the first
SETTINGS frame. Detection of a change by a receiver MUST be treated SETTINGS frame. Detection of a change by a receiver MUST be treated
as a connection error of type PROTOCOL_ERROR. as a connection error of type PROTOCOL_ERROR.
The SETTINGS frame precedes any HTTP/2 priority signal sent from a The SETTINGS frame precedes any HTTP/2 priority signal sent from a
client, so a server can determine if it needs to allocate any client, so a server can determine if it needs to allocate any
resource to signal handling before they arrive. A server that resource to signal handling before they arrive. A server that
skipping to change at page 7, line 38 skipping to change at page 8, line 14
4. Priority Parameters 4. Priority Parameters
The priority information is a sequence of key-value pairs, providing The priority information is a sequence of key-value pairs, providing
room for future extensions. Each key-value pair represents a room for future extensions. Each key-value pair represents a
priority parameter. priority parameter.
The Priority HTTP header field (Section 5) is an end-to-end way to The Priority HTTP header field (Section 5) is an end-to-end way to
transmit this set of parameters when a request or a response is transmit this set of parameters when a request or a response is
issued. In order to reprioritize a request, HTTP-version-specific issued. In order to reprioritize a request, HTTP-version-specific
frames (Section 7.1 and Section 7.2) are used by clients to transmit PRIORITY_UPDATE frames (Section 7.1 and Section 7.2) are used by
the same information on a single hop. If intermediaries want to clients to transmit the same information on a single hop. If
specify prioritization on a multiplexed HTTP connection, they SHOULD intermediaries want to specify prioritization on a multiplexed HTTP
use a PRIORITY_UPDATE frame and SHOULD NOT change the Priority header connection, they SHOULD use a PRIORITY_UPDATE frame and SHOULD NOT
field. change the Priority header field.
In both cases, the set of priority parameters is encoded as a In both cases, the set of priority parameters is encoded as a
Structured Fields Dictionary (see Section 3.2 of Structured Fields Dictionary (see Section 3.2 of
[STRUCTURED-FIELDS]). [STRUCTURED-FIELDS]).
This document defines the urgency(u) and incremental(i) parameters. This document defines the urgency(u) and incremental(i) parameters.
When receiving an HTTP request that does not carry these priority When receiving an HTTP request that does not carry these priority
parameters, a server SHOULD act as if their default values were parameters, a server SHOULD act as if their default values were
specified. Note that handling of omitted parameters is different specified. Note that handling of omitted parameters is different
when processing an HTTP response; see Section 8. when processing an HTTP response; see Section 8.
Unknown parameters, parameters with out-of-range values or values of Receivers parse the Dictionary as defined in Section 4.2 of
[STRUCTURED-FIELDS]. Where the Dictionary is successfully parsed,
this document places the additional requirement that unknown priority
parameters, parameters with out-of-range values, or values of
unexpected types MUST be ignored. unexpected types MUST be ignored.
4.1. Urgency 4.1. Urgency
The urgency parameter (u) takes an integer between 0 and 7, in The urgency parameter (u) takes an integer between 0 and 7, in
descending order of priority. This range provides sufficient descending order of priority.
granularity for prioritizing responses for ordinary web browsing, at
minimal complexity.
The value is encoded as an sf-integer. The default value is 3. The value is encoded as an sf-integer. The default value is 3.
This parameter indicates the sender's recommendation, based on the Endpoints use this parameter to communicate their view of the
expectation that the server would transmit HTTP responses in the precedence of HTTP responses. The chosen value of urgency can be
order of their urgency values if possible. The smaller the value, based on the expectation that servers might use this information to
the higher the precedence. transmit HTTP responses in the order of their urgency. The smaller
the value, the higher the precedence.
The following example shows a request for a CSS file with the urgency The following example shows a request for a CSS file with the urgency
set to 0: set to 0:
:method = GET :method = GET
:scheme = https :scheme = https
:authority = example.net :authority = example.net
:path = /style.css :path = /style.css
priority = u=0 priority = u=0
skipping to change at page 8, line 43 skipping to change at page 9, line 23
to the main resource. This convention allows servers to refine the to the main resource. This convention allows servers to refine the
urgency using knowledge specific to the web-site (see Section 8). urgency using knowledge specific to the web-site (see Section 8).
The lowest urgency level (7) is reserved for background tasks such as The lowest urgency level (7) is reserved for background tasks such as
delivery of software updates. This urgency level SHOULD NOT be used delivery of software updates. This urgency level SHOULD NOT be used
for fetching responses that have impact on user interaction. for fetching responses that have impact on user interaction.
4.2. Incremental 4.2. Incremental
The incremental parameter (i) takes an sf-boolean as the value that The incremental parameter (i) takes an sf-boolean as the value that
indicates if an HTTP response can be processed incrementally, i.e. indicates if an HTTP response can be processed incrementally, i.e.,
provide some meaningful output as chunks of the response arrive. provide some meaningful output as chunks of the response arrive.
The default value of the incremental parameter is false (0). The default value of the incremental parameter is false (0).
A server might distribute the bandwidth of a connection between If a client makes concurrent requests with the incremental parameter
incremental responses that share the same urgency, hoping that set to false, there is no benefit serving responses with the same
providing those responses in parallel would be more helpful to the urgency concurrently because the client is not going to process those
client than delivering the responses one by one. responses incrementally. Serving non-incremental responses with the
same urgency one by one, in the order in which those requests were
generated is considered to be the best strategy.
If a client makes concurrent requests with the incremental parameter If a client makes concurrent requests with the incremental parameter
set to false, there is no benefit serving responses in parallel set to true, serving requests with the same urgency concurrently
because the client is not going to process those responses might be beneficial. Doing this distributes the connection
incrementally. Serving non-incremental responses one by one, in the bandwidth, meaning that responses take longer to complete.
order in which those requests were generated is considered to be the Incremental delivery is most useful where multiple partial responses
best strategy. might provide some value to clients ahead of a complete response
being available.
The following example shows a request for a JPEG file with the The following example shows a request for a JPEG file with the
urgency parameter set to 5 and the incremental parameter set to true. urgency parameter set to 5 and the incremental parameter set to true.
:method = GET :method = GET
:scheme = https :scheme = https
:authority = example.net :authority = example.net
:path = /image.jpg :path = /image.jpg
priority = u=5, i priority = u=5, i
4.3. Defining New Parameters 4.3. Defining New Parameters
When attempting to define new parameters, care must be taken so that When attempting to define new parameters, care must be taken so that
they do not adversely interfere with prioritization performed by they do not adversely interfere with prioritization performed by
existing endpoints or intermediaries that do not understand the newly existing endpoints or intermediaries that do not understand the newly
defined parameter. Since unknown parameters are ignored, new defined parameter. Since unknown parameters are ignored, new
parameters should not change the interpretation of or modify the parameters should not change the interpretation of, or modify, the
predefined parameters in a way that is not backwards compatible or urgency (see Section 4.1) or incremental (see Section 4.2) parameters
fallback safe. in a way that is not backwards compatible or fallback safe.
For example, if there is a need to provide more granularity than For example, if there is a need to provide more granularity than
eight urgency levels, it would be possible to subdivide the range eight urgency levels, it would be possible to subdivide the range
using an additional parameter. Implementations that do not recognize using an additional parameter. Implementations that do not recognize
the parameter can safely continue to use the less granular eight the parameter can safely continue to use the less granular eight
levels. levels.
Alternatively, the urgency can be augmented. For example, a Alternatively, the urgency can be augmented. For example, a
graphical user agent could send a visible parameter to indicate if graphical user agent could send a visible parameter to indicate if
the resource being requested is within the viewport. the resource being requested is within the viewport.
Generic parameters are preferred over vendor-specific, application- Generic parameters are preferred over vendor-specific, application-
specific or deployment-specific values. If a generic value cannot be specific or deployment-specific values. If a generic value cannot be
agreed upon in the community, the parameter's name should be agreed upon in the community, the parameter's name should be
correspondingly specific (e.g., with a prefix that identifies the correspondingly specific (e.g., with a prefix that identifies the
vendor, application or deployment). vendor, application or deployment).
4.3.1. Registration 4.3.1. Registration
New Priority parameters can be defined by registering them in the New Priority parameters can be defined by registering them in the
HTTP Priority Parameters Registry. HTTP Priority Parameters Registry. The registry governs the keys
(short textual strings) used in Structured Fields Dictionary (see
Registration requests are reviewed and approved by a Designated Section 3.2 of [STRUCTURED-FIELDS]). Since each HTTP request can
Expert, as per Section 4.5 of [RFC8126]. A specification document is have associated priority signals, there is value in having short key
appreciated, but not required. lengths, especially single-character strings. In order to encourage
extension while avoiding unintended conflict among attractive key
values, the HTTP Priority Parameters Registry operates two
registration policies depending on key length.
The Expert(s) should consider the following factors when evaluating * Registration requests for parameters with a key length of one use
requests: the Specification Required policy, as per Section 4.6 of
[RFC8126].
* Community feedback * Registration requests for parameters with a key length greater
than one use the Expert Review policy, as per Section 4.5 of
[RFC8126]. A specification document is appreciated, but not
required.
* If the parameters are sufficiently well-defined and adhere to the When reviewing registration requests, the designated expert(s) can
guidance provided in Section 4.3. consider the additional guidance provided in Section 4.3 but cannot
use it as a basis for rejection.
Registration requests should use the following template: Registration requests should use the following template:
* Name: [a name for the Priority Parameter that matches key] Name: [a name for the Priority Parameter that matches key]
* Description: [a description of the parameter semantics and value] Description: [a description of the parameter semantics and value]
* Reference: [to a specification defining this parameter] Reference: [to a specification defining this parameter]
See the registry at https://iana.org/assignments/http-priority See the registry at https://iana.org/assignments/http-priority
(https://iana.org/assignments/http-priority) for details on where to (https://iana.org/assignments/http-priority) for details on where to
send registration requests. send registration requests.
5. The Priority HTTP Header Field 5. The Priority HTTP Header Field
The Priority HTTP header field can appear in requests and responses. The Priority HTTP header field carries priority parameters Section 4.
A client uses it to specify the priority of the response. A server It can appear in requests and responses. It is an end-to-end signal
uses it to inform the client that the priority was overwritten. An of the request priority from the client or the response priority from
intermediary can use the Priority information from client requests the server. Section 8 describes how intermediaries can combine the
and server responses to correct or amend the precedence to suit it priority information from client requests and server responses to
(see Section 8). correct or amend the precedence. Clients cannot interpret the
appearance or omission of a Priority response header as
acknowledgement that any prioritization has occurred. Guidance for
how endpoints can act on Priority header values is given in
Section 10 and Section 9.
The Priority header field is an end-to-end signal of the request Priority is a Dictionary (Section 3.2 of [STRUCTURED-FIELDS]):
priority from the client or the response priority from the server.
Priority = sf-dictionary
As is the ordinary case for HTTP caching [CACHING], a response with a As is the ordinary case for HTTP caching [CACHING], a response with a
Priority header field might be cached and re-used for subsequent Priority header field might be cached and re-used for subsequent
requests. When an origin server generates the Priority response requests. When an origin server generates the Priority response
header field based on properties of an HTTP request it receives, the header field based on properties of an HTTP request it receives, the
server is expected to control the cacheability or the applicability server is expected to control the cacheability or the applicability
of the cached response, by using header fields that control the of the cached response, by using header fields that control the
caching behavior (e.g., Cache-Control, Vary). caching behavior (e.g., Cache-Control, Vary).
An endpoint that fails to parse the Priority header field SHOULD use
default parameter values.
6. Reprioritization 6. Reprioritization
After a client sends a request, it may be beneficial to change the After a client sends a request, it may be beneficial to change the
priority of the response. As an example, a web browser might issue a priority of the response. As an example, a web browser might issue a
prefetch request for a JavaScript file with the urgency parameter of prefetch request for a JavaScript file with the urgency parameter of
the Priority request header field set to u=7 (background). Then, the Priority request header field set to u=7 (background). Then,
when the user navigates to a page which references the new JavaScript when the user navigates to a page which references the new JavaScript
file, while the prefetch is in progress, the browser would send a file, while the prefetch is in progress, the browser would send a
reprioritization signal with the priority field value set to u=0. reprioritization signal with the priority field value set to u=0.
The PRIORITY_UPDATE frame (Section 7) can be used for such The PRIORITY_UPDATE frame (Section 7) can be used for such
skipping to change at page 12, line 50 skipping to change at page 13, line 50
fields: fields:
Reserved: A reserved 1-bit field. The semantics of this bit are Reserved: A reserved 1-bit field. The semantics of this bit are
undefined, and the bit MUST remain unset (0x0) when sending and undefined, and the bit MUST remain unset (0x0) when sending and
MUST be ignored when receiving. MUST be ignored when receiving.
Prioritized Stream ID: A 31-bit stream identifier for the stream Prioritized Stream ID: A 31-bit stream identifier for the stream
that is the target of the priority update. that is the target of the priority update.
Priority Field Value: The priority update value in ASCII text, Priority Field Value: The priority update value in ASCII text,
encoded using Structured Fields. encoded using Structured Fields. This is the same representation
as the Priority header field value.
When the PRIORITY_UPDATE frame applies to a request stream, clients When the PRIORITY_UPDATE frame applies to a request stream, clients
SHOULD provide a Prioritized Stream ID that refers to a stream in the SHOULD provide a Prioritized Stream ID that refers to a stream in the
"open", "half-closed (local)", or "idle" state. Servers can discard "open", "half-closed (local)", or "idle" state. Servers can discard
frames where the Prioritized Stream ID refers to a stream in the frames where the Prioritized Stream ID refers to a stream in the
"half-closed (local)" or "closed" state. The number of streams which "half-closed (local)" or "closed" state. The number of streams which
have been prioritized but remain in the "idle" state plus the number have been prioritized but remain in the "idle" state plus the number
of active streams (those in the "open" or either "half-closed" state; of active streams (those in the "open" or either "half-closed" state;
see Section 5.1.2 of [HTTP2]) MUST NOT exceed the value of the see Section 5.1.2 of [HTTP2]) MUST NOT exceed the value of the
SETTINGS_MAX_CONCURRENT_STREAMS parameter. Servers that receive such SETTINGS_MAX_CONCURRENT_STREAMS parameter. Servers that receive such
skipping to change at page 14, line 20 skipping to change at page 15, line 20
} }
Figure 2: HTTP/3 PRIORITY_UPDATE Frame Figure 2: HTTP/3 PRIORITY_UPDATE Frame
The PRIORITY_UPDATE frame payload has the following fields: The PRIORITY_UPDATE frame payload has the following fields:
Prioritized Element ID: The stream ID or push ID that is the target Prioritized Element ID: The stream ID or push ID that is the target
of the priority update. of the priority update.
Priority Field Value: The priority update value in ASCII text, Priority Field Value: The priority update value in ASCII text,
encoded using Structured Fields. encoded using Structured Fields. This is the same representation
as the Priority header field value.
The request-stream variant of PRIORITY_UPDATE (type=0xF0700) MUST The request-stream variant of PRIORITY_UPDATE (type=0xF0700) MUST
reference a request stream. If a server receives a PRIORITY_UPDATE reference a request stream. If a server receives a PRIORITY_UPDATE
(type=0xF0700) for a Stream ID that is not a request stream, this (type=0xF0700) for a Stream ID that is not a request stream, this
MUST be treated as a connection error of type H3_ID_ERROR. The MUST be treated as a connection error of type H3_ID_ERROR. The
Stream ID MUST be within the client-initiated bidirectional stream Stream ID MUST be within the client-initiated bidirectional stream
limit. If a server receives a PRIORITY_UPDATE (type=0xF0700) with a limit. If a server receives a PRIORITY_UPDATE (type=0xF0700) with a
Stream ID that is beyond the stream limits, this SHOULD be treated as Stream ID that is beyond the stream limits, this SHOULD be treated as
a connection error of type H3_ID_ERROR. a connection error of type H3_ID_ERROR. Generating an error is not
mandatory because HTTP/3 implementations might have practical
barriers to determining the active stream concurrency limit that is
applied by the QUIC layer.
The push-stream variant PRIORITY_UPDATE (type=0xF0701) MUST reference The push-stream variant PRIORITY_UPDATE (type=0xF0701) MUST reference
a promised push stream. If a server receives a PRIORITY_UPDATE a promised push stream. If a server receives a PRIORITY_UPDATE
(type=0xF0701) with a Push ID that is greater than the maximum Push (type=0xF0701) with a Push ID that is greater than the maximum Push
ID or which has not yet been promised, this MUST be treated as a ID or which has not yet been promised, this MUST be treated as a
connection error of type H3_ID_ERROR. connection error of type H3_ID_ERROR.
PRIORITY_UPDATE frames of either type are only sent by clients. If a PRIORITY_UPDATE frames of either type are only sent by clients. If a
client receives a PRIORITY_UPDATE frame, this MUST be treated as a client receives a PRIORITY_UPDATE frame, this MUST be treated as a
connection error of type H3_FRAME_UNEXPECTED. connection error of type H3_FRAME_UNEXPECTED.
skipping to change at page 16, line 20 skipping to change at page 17, line 28
perfect scheduler and this document only provides some basic perfect scheduler and this document only provides some basic
recommendations for implementations. recommendations for implementations.
Clients cannot depend on particular treatment based on priority Clients cannot depend on particular treatment based on priority
signals. Servers can use other information to prioritize responses. signals. Servers can use other information to prioritize responses.
It is RECOMMENDED that, when possible, servers respect the urgency It is RECOMMENDED that, when possible, servers respect the urgency
parameter (Section 4.1), sending higher urgency responses before parameter (Section 4.1), sending higher urgency responses before
lower urgency responses. lower urgency responses.
It is RECOMMENDED that, when possible, servers respect the
incremental parameter (Section 4.2). Non-incremental responses of
the same urgency SHOULD be served one-by-one based on the Stream ID,
which corresponds to the order in which clients make requests. Doing
so ensures that clients can use request ordering to influence
response order. Incremental responses of the same urgency SHOULD be
served in round-robin manner.
The incremental parameter indicates how a client processes response The incremental parameter indicates how a client processes response
bytes as they arrive. Non-incremental resources are only used when bytes as they arrive. It is RECOMMENDED that, when possible, servers
all of the response payload has been received. Incremental resources respect the incremental parameter (Section 4.2). Non-incremental
are used as parts, or chunks, of the response payload are received. resources can only be used when all of the response payload has been
Therefore, the timing of response data reception at the client, such received. Therefore, non-incremental responses of the same urgency
as the time to early bytes or the time to receive the entire payload, SHOULD be served in their entirety, one-by-one, based on the stream
plays an important role in perceived performance. Timings depend on ID, which corresponds to the order in which clients make requests.
resource size but this scheme provides no explicit guidance about how Doing so ensures that clients can use request ordering to influence
a server should use size as an input to prioritization. Instead, the response order.
following examples demonstrate how a server that strictly abides the
scheduling guidance based on urgency and request generation order Incremental responses of the same urgency SHOULD be served by sharing
could find that early requests prevent serving of later requests. bandwidth amongst them. Incremental resources are used as parts, or
chunks, of the response payload are received. A client might benefit
more from receiving a portion of all these resources rather than the
entirety of a single resource. How large a portion of the resource
is needed to be useful in improving performance varies. Some
resource types place critical elements early, others can use
information progressively. This scheme provides no explicit mandate
about how a server should use size, type or any other input to decide
how to prioritize.
There can be scenarios where a server will need to schedule multiple
incremental and non-incremental responses at the same urgency level.
Strictly abiding the scheduling guidance based on urgency and request
generation order might lead to sub-optimal results at the client, as
early non-incremental responses might prevent serving of incremental
responses issued later. The following are examples of such
challenges.
1. At the same urgency level, a non-incremental request for a large 1. At the same urgency level, a non-incremental request for a large
resource followed by an incremental request for a small resource. resource followed by an incremental request for a small resource.
2. At the same urgency level, an incremental request of 2. At the same urgency level, an incremental request of
indeterminate length followed by a non-incremental large indeterminate length followed by a non-incremental large
resource. resource.
It is RECOMMENDED that servers avoid such starvation where possible. It is RECOMMENDED that servers avoid such starvation where possible.
The method to do so is an implementation decision. For example, a The method to do so is an implementation decision. For example, a
skipping to change at page 17, line 21 skipping to change at page 18, line 36
on the best way to apply these. A server that is too simple could on the best way to apply these. A server that is too simple could
easily push at too high a priority and block client requests, or push easily push at too high a priority and block client requests, or push
at too low a priority and delay the response, negating intended goals at too low a priority and delay the response, negating intended goals
of server push. of server push.
Priority signals are a factor for server push scheduling. The Priority signals are a factor for server push scheduling. The
concept of parameter value defaults applies slightly differently concept of parameter value defaults applies slightly differently
because there is no explicit client-signalled initial priority. A because there is no explicit client-signalled initial priority. A
server can apply priority signals provided in an origin response; see server can apply priority signals provided in an origin response; see
the merging guidance given in Section 8. In the absence of origin the merging guidance given in Section 8. In the absence of origin
signals, applying default parameter values could be suboptimal. How signals, applying default parameter values could be suboptimal. By
ever a server decides to schedule a pushed response, it can signal whatever means a server decides to schedule a pushed response, it can
the intended priority to the client by including the Priority field signal the intended priority to the client by including the Priority
in a PUSH_PROMISE or HEADERS frame. field in a PUSH_PROMISE or HEADERS frame.
10.1. Intermediaries with Multiple Backend Connections 10.1. Intermediaries with Multiple Backend Connections
An intermediary serving an HTTP connection might split requests over An intermediary serving an HTTP connection might split requests over
multiple backend connections. When it applies prioritization rules multiple backend connections. When it applies prioritization rules
strictly, low priority requests cannot make progress while requests strictly, low priority requests cannot make progress while requests
with higher priorities are inflight. This blocking can propagate to with higher priorities are inflight. This blocking can propagate to
backend connections, which the peer might interpret as a connection backend connections, which the peer might interpret as a connection
stall. Endpoints often implement protections against stalls, such as stall. Endpoints often implement protections against stalls, such as
abruptly closing connections after a certain time period. To reduce abruptly closing connections after a certain time period. To reduce
skipping to change at page 18, line 30 skipping to change at page 19, line 43
when scheduling both new data and retransmission data might better when scheduling both new data and retransmission data might better
match the expectations of the application. However, there are no match the expectations of the application. However, there are no
requirements on how a transport chooses to schedule based on this requirements on how a transport chooses to schedule based on this
information because the decision depends on several factors and information because the decision depends on several factors and
trade-offs. It could prioritize new data for a higher urgency stream trade-offs. It could prioritize new data for a higher urgency stream
over retransmission data for a lower priority stream, or it could over retransmission data for a lower priority stream, or it could
prioritize retransmission data over new data irrespective of prioritize retransmission data over new data irrespective of
urgencies. urgencies.
Section 6.2.4 of [QUIC-RECOVERY], also highlights consideration of Section 6.2.4 of [QUIC-RECOVERY], also highlights consideration of
application priorities when sending probe packets after PTO timer application priorities when sending probe packets after Probe Timeout
expiration. An QUIC implementation supporting application-indicated timer expiration. A QUIC implementation supporting application-
priorities might use the relative priority of streams when choosing indicated priorities might use the relative priority of streams when
probe data. choosing probe data.
13. Fairness 13. Fairness
As a general guideline, a server SHOULD NOT use priority information As a general guideline, a server SHOULD NOT use priority information
for making schedule decisions across multiple connections, unless it for making scheduling decisions across multiple connections, unless
knows that those connections originate from the same client. Due to it knows that those connections originate from the same client. Due
this, priority information conveyed over a non-coalesced HTTP to this, priority information conveyed over a non-coalesced HTTP
connection (e.g., HTTP/1.1) might go unused. connection (e.g., HTTP/1.1) might go unused.
The remainder of this section discusses scenarios where unfairness is The remainder of this section discusses scenarios where unfairness is
problematic and presents possible mitigations, or where unfairness is problematic and presents possible mitigations, or where unfairness is
desirable. desirable.
13.1. Coalescing Intermediaries 13.1. Coalescing Intermediaries
When an intermediary coalesces HTTP requests coming from multiple When an intermediary coalesces HTTP requests coming from multiple
clients into one HTTP/2 or HTTP/3 connection going to the backend clients into one HTTP/2 or HTTP/3 connection going to the backend
skipping to change at page 19, line 17 skipping to change at page 20, line 36
an example, a resource-constrained server might defer the an example, a resource-constrained server might defer the
transmission of software update files that would have the background transmission of software update files that would have the background
urgency being associated. However, in the worst case, the asymmetry urgency being associated. However, in the worst case, the asymmetry
between the precedence declared by multiple clients might cause between the precedence declared by multiple clients might cause
responses going to one user agent to be delayed totally after those responses going to one user agent to be delayed totally after those
going to another. going to another.
In order to mitigate this fairness problem, a server could use In order to mitigate this fairness problem, a server could use
knowledge about the intermediary as another signal in its knowledge about the intermediary as another signal in its
prioritization decisions. For instance, if a server knows the prioritization decisions. For instance, if a server knows the
intermediary is coalescing requests, then it could serve the intermediary is coalescing requests, then it could avoid serving the
responses in round-robin manner. This can work if the constrained responses in their entirety and instead distribute bandwidth (for
example, in a round-robin manner). This can work if the constrained
resource is network capacity between the intermediary and the user resource is network capacity between the intermediary and the user
agent, as the intermediary buffers responses and forwards the chunks agent, as the intermediary buffers responses and forwards the chunks
based on the prioritization scheme it implements. based on the prioritization scheme it implements.
A server can determine if a request came from an intermediary through A server can determine if a request came from an intermediary through
configuration, or by consulting if that request contains one of the configuration, or by consulting if that request contains one of the
following header fields: following header fields:
* Forwarded [FORWARDED], X-Forwarded-For * Forwarded [FORWARDED], X-Forwarded-For
* Via (see Section 7.6.3 of [HTTP]) * Via (see Section 7.6.3 of [HTTP])
13.2. HTTP/1.x Back Ends 13.2. HTTP/1.x Back Ends
It is common for CDN infrastructure to support different HTTP It is common for CDN infrastructure to support different HTTP
versions on the front end and back end. For instance, the client- versions on the front end and back end. For instance, the client-
facing edge might support HTTP/2 and HTTP/3 while communication to facing edge might support HTTP/2 and HTTP/3 while communication to
back end servers is done using HTTP/1.1. Unlike with connection back end servers is done using HTTP/1.1. Unlike with connection
coalescing, the CDN will "de-mux" requests into discrete connections coalescing, the CDN will "de-mux" requests into discrete connections
to the back end. As HTTP/1.1 and older do not provide a way to to the back end. HTTP/1.1 and older do not support response
concurrently transmit multiple responses, there is no immediate multiplexing in a single connection, so there is not a fairness
fairness issue in protocol. However, back end servers MAY still use problem. However, back end servers MAY still use client headers for
client headers for request scheduling. Back end servers SHOULD only request scheduling. Back end servers SHOULD only schedule based on
schedule based on client priority information where that information client priority information where that information can be scoped to
can be scoped to individual end clients. Authentication and other individual end clients. Authentication and other session information
session information might provide this linkability. might provide this linkability.
13.3. Intentional Introduction of Unfairness 13.3. Intentional Introduction of Unfairness
It is sometimes beneficial to deprioritize the transmission of one It is sometimes beneficial to deprioritize the transmission of one
connection over others, knowing that doing so introduces a certain connection over others, knowing that doing so introduces a certain
amount of unfairness between the connections and therefore between amount of unfairness between the connections and therefore between
the requests served on those connections. the requests served on those connections.
For example, a server might use a scavenging congestion controller on For example, a server might use a scavenging congestion controller on
connections that only convey background priority responses such as connections that only convey background priority responses such as
skipping to change at page 20, line 37 skipping to change at page 22, line 11
value of the cached header field when serving the cached response, value of the cached header field when serving the cached response,
only because the header field is defined as end-to-end rather than only because the header field is defined as end-to-end rather than
hop-by-hop. hop-by-hop.
It should also be noted that the use of a header field carrying a It should also be noted that the use of a header field carrying a
textual value makes the prioritization scheme extensible; see the textual value makes the prioritization scheme extensible; see the
discussion below. discussion below.
15. Security Considerations 15. Security Considerations
[CVE-2019-9513] aka "Resource Loop", is a DoS attack based on [RFC7540] stream prioritization relies on dependencies.
manipulation of the RFC 7540 priority tree. Extensible priorities Considerations are presented to implementations, describing how
does not use stream dependencies, which mitigates this vulnerability. limiting state or work commitments can avoid some types of problems.
In addition, [CVE-2019-9513] aka "Resource Loop", is an example of a
Section 5.3.4 of [RFC7540] describes a scenario where closure of DoS attack that abuses stream dependencies. Extensible priorities
streams in the priority tree could cause suboptimal prioritization. does not use dependencies, which avoids these issues.
To avoid this, [RFC7540] states that "an endpoint SHOULD retain
stream prioritization state for a period after streams become
closed". Retaining state for streams no longer counted towards
stream concurrency consumes server resources. Furthermore, [RFC7540]
identifies that reprioritization of a closed stream could affect
dependents; it recommends updating the priority tree if sufficient
state is stored, which will also consume server resources. To limit
this commitment, it is stated that "The amount of prioritization
state that is retained MAY be limited" and "If a limit is applied,
endpoints SHOULD maintain state for at least as many streams as
allowed by their setting for SETTINGS_MAX_CONCURRENT_STREAMS.".
Extensible priorities does not use stream dependencies, which
minimizes most of the resource concerns related to this scenario.
Section 5.3.4 of [RFC7540] also presents considerations about the Section 7 describes considerations for server buffering of
state required to store priority information about streams in an PRIORITY_UPDATE frames.
"idle" state. This state can be limited by adopting the guidance
about concurrency limits described above. Extensible priorities is
subject to a similar consideration because PRIORITY_UPDATE frames may
arrive before the request that they reference. A server is required
to store the information in order to apply the most up-to-date signal
to the request. However, HTTP/3 implementations might have practical
barriers to determining reasonable stream concurrency limits
depending on the information that is available to them from the QUIC
transport layer.
16. IANA Considerations Section 10 presents examples where servers that prioritize responses
in a certain way might be starved of the ability to transmit payload.
This specification registers the following entry in the Permanent The security considerations from [STRUCTURED-FIELDS] apply to
Message Header Field Names registry established by [RFC3864]: processing of priority parameters defined in Section 4.
Header field name: Priority 16. IANA Considerations
Applicable protocol: http This specification registers the following entry in the the Hypertext
Transfer Protocol (HTTP) Field Name Registry established by [HTTP]:
Status: standard Field name: Priority
Author/change controller: IETF Status: permanent
Specification document(s): This document Specification document(s): This document
Related information: n/a
This specification registers the following entry in the HTTP/2 This specification registers the following entry in the HTTP/2
Settings registry established by [HTTP2]: Settings registry established by [RFC7540]:
Name: SETTINGS_NO_RFC7540_PRIORITIES Name: SETTINGS_NO_RFC7540_PRIORITIES
Code: 0x9 Code: 0x9
Initial value: 0 Initial value: 0
Specification: This document Specification: This document
This specification registers the following entry in the HTTP/2 Frame This specification registers the following entry in the HTTP/2 Frame
Type registry established by [HTTP2]: Type registry established by [RFC7540]:
Frame Type: PRIORITY_UPDATE Frame Type: PRIORITY_UPDATE
Code: 0x10 Code: 0x10
Specification: This document Specification: This document
This specification registers the following entries in the HTTP/3 This specification registers the following entries in the HTTP/3
Frame Type registry established by [HTTP3]: Frame Type registry established by [HTTP3]:
Frame Type: PRIORITY_UPDATE Frame Type: PRIORITY_UPDATE
skipping to change at page 24, line 5 skipping to change at page 25, line 5
Priorities", Work in Progress, Internet-Draft, draft- Priorities", Work in Progress, Internet-Draft, draft-
lassey-priority-setting-00, 25 July 2019, lassey-priority-setting-00, 25 July 2019,
<https://datatracker.ietf.org/doc/html/draft-lassey- <https://datatracker.ietf.org/doc/html/draft-lassey-
priority-setting-00>. priority-setting-00>.
[QUIC-RECOVERY] [QUIC-RECOVERY]
Iyengar, J., Ed. and I. Swett, Ed., "QUIC Loss Detection Iyengar, J., Ed. and I. Swett, Ed., "QUIC Loss Detection
and Congestion Control", RFC 9002, DOI 10.17487/RFC9002, and Congestion Control", RFC 9002, DOI 10.17487/RFC9002,
May 2021, <https://www.rfc-editor.org/rfc/rfc9002>. May 2021, <https://www.rfc-editor.org/rfc/rfc9002>.
[RFC3864] Klyne, G., Nottingham, M., and J. Mogul, "Registration
Procedures for Message Header Fields", BCP 90, RFC 3864,
DOI 10.17487/RFC3864, September 2004,
<https://www.rfc-editor.org/rfc/rfc3864>.
[RFC7540] Belshe, M., Peon, R., and M. Thomson, Ed., "Hypertext [RFC7540] Belshe, M., Peon, R., and M. Thomson, Ed., "Hypertext
Transfer Protocol Version 2 (HTTP/2)", RFC 7540, Transfer Protocol Version 2 (HTTP/2)", RFC 7540,
DOI 10.17487/RFC7540, May 2015, DOI 10.17487/RFC7540, May 2015,
<https://www.rfc-editor.org/rfc/rfc7540>. <https://www.rfc-editor.org/rfc/rfc7540>.
[RFC8081] Lilley, C., "The "font" Top-Level Media Type", RFC 8081, [RFC8081] Lilley, C., "The "font" Top-Level Media Type", RFC 8081,
DOI 10.17487/RFC8081, February 2017, DOI 10.17487/RFC8081, February 2017,
<https://www.rfc-editor.org/rfc/rfc8081>. <https://www.rfc-editor.org/rfc/rfc8081>.
Appendix A. Acknowledgements Appendix A. Acknowledgements
Roy Fielding presented the idea of using a header field for Roy Fielding presented the idea of using a header field for
representing priorities in http://tools.ietf.org/agenda/83/slides/ representing priorities in http://tools.ietf.org/agenda/83/slides/
slides-83-httpbis-5.pdf (http://tools.ietf.org/agenda/83/slides/ slides-83-httpbis-5.pdf (http://tools.ietf.org/agenda/83/slides/
slides-83-httpbis-5.pdf). In https://github.com/pmeenan/http3- slides-83-httpbis-5.pdf). In https://github.com/pmeenan/http3-
prioritization-proposal (https://github.com/pmeenan/http3- prioritization-proposal (https://github.com/pmeenan/http3-
prioritization-proposal), Patrick Meenan advocates for representing prioritization-proposal), Patrick Meenan advocated for representing
the priorities using a tuple of urgency and concurrency. The ability the priorities using a tuple of urgency and concurrency. The ability
to disable HTTP/2 prioritization is inspired by to disable HTTP/2 prioritization is inspired by
[I-D.lassey-priority-setting], authored by Brad Lassey and Lucas [I-D.lassey-priority-setting], authored by Brad Lassey and Lucas
Pardue, with modifications based on feedback that was not Pardue, with modifications based on feedback that was not
incorporated into an update to that document. incorporated into an update to that document.
The motivation for defining an alternative to HTTP/2 priorities is The motivation for defining an alternative to HTTP/2 priorities is
drawn from discussion within the broad HTTP community. Special drawn from discussion within the broad HTTP community. Special
thanks to Roberto Peon, Martin Thomson and Netflix for text that was thanks to Roberto Peon, Martin Thomson and Netflix for text that was
incorporated explicitly in this document. incorporated explicitly in this document.
In addition to the people above, this document owes a lot to the In addition to the people above, this document owes a lot to the
extensive discussion in the HTTP priority design team, consisting of extensive discussion in the HTTP priority design team, consisting of
Alan Frindell, Andrew Galloni, Craig Taylor, Ian Swett, Kazuho Oku, Alan Frindell, Andrew Galloni, Craig Taylor, Ian Swett, Kazuho Oku,
Lucas Pardue, Matthew Cox, Mike Bishop, Roberto Peon, Robin Marx, Roy Lucas Pardue, Matthew Cox, Mike Bishop, Roberto Peon, Robin Marx, Roy
Fielding. Fielding.
Yang Chi contributed the section on retransmission scheduling.
Appendix B. Change Log Appendix B. Change Log
B.1. Since draft-ietf-httpbis-priority-05 B.1. Since draft-ietf-httpbis-priority-06
* Focus on editorial changes
* Clarify rules about Sf-Dictionary handling in headers
* Split policy for parameter IANA registry into two sections based
on key length
B.2. Since draft-ietf-httpbis-priority-05
* Renamed SETTINGS_DEPRECATE_RFC7540_PRIORITIES to * Renamed SETTINGS_DEPRECATE_RFC7540_PRIORITIES to
SETTINGS_NO_RFC7540_PRIORITIES SETTINGS_NO_RFC7540_PRIORITIES
* Clarify that senders of the HTTP/2 setting can use any alternative * Clarify that senders of the HTTP/2 setting can use any alternative
(#1679, #1705) (#1679, #1705)
B.2. Since draft-ietf-httpbis-priority-04 B.3. Since draft-ietf-httpbis-priority-04
* Renamed SETTINGS_DEPRECATE_HTTP2_PRIORITIES to * Renamed SETTINGS_DEPRECATE_HTTP2_PRIORITIES to
SETTINGS_DEPRECATE_RFC7540_PRIORITIES (#1601) SETTINGS_DEPRECATE_RFC7540_PRIORITIES (#1601)
* Reoriented text towards RFC7540bis (#1561, #1601) * Reoriented text towards RFC7540bis (#1561, #1601)
* Clarify intermediary behavior (#1562) * Clarify intermediary behavior (#1562)
B.3. Since draft-ietf-httpbis-priority-03 B.4. Since draft-ietf-httpbis-priority-03
* Add statement about what this scheme applies to. Clarify * Add statement about what this scheme applies to. Clarify
extensions can use it but must define how themselves (#1550, extensions can use it but must define how themselves (#1550,
#1559) #1559)
* Describe scheduling considerations for the CONNECT method (#1495, * Describe scheduling considerations for the CONNECT method (#1495,
#1544) #1544)
* Describe scheduling considerations for retransmitted data (#1429, * Describe scheduling considerations for retransmitted data (#1429,
#1504) #1504)
* Suggest intermediaries might avoid strict prioritization (#1562) * Suggest intermediaries might avoid strict prioritization (#1562)
B.4. Since draft-ietf-httpbis-priority-02 B.5. Since draft-ietf-httpbis-priority-02
* Describe considerations for server push prioritization (#1056, * Describe considerations for server push prioritization (#1056,
#1345) #1345)
* Define HTTP/2 PRIORITY_UPDATE ID limits in HTTP/2 terms (#1261, * Define HTTP/2 PRIORITY_UPDATE ID limits in HTTP/2 terms (#1261,
#1344) #1344)
* Add a Parameters registry (#1371) * Add a Parameters registry (#1371)
B.5. Since draft-ietf-httpbis-priority-01 B.6. Since draft-ietf-httpbis-priority-01
* PRIORITY_UPDATE frame changes (#1096, #1079, #1167, #1262, #1267, * PRIORITY_UPDATE frame changes (#1096, #1079, #1167, #1262, #1267,
#1271) #1271)
* Add section to describe server scheduling considerations (#1215, * Add section to describe server scheduling considerations (#1215,
#1232, #1266) #1232, #1266)
* Remove specific instructions related to intermediary fairness * Remove specific instructions related to intermediary fairness
(#1022, #1264) (#1022, #1264)
B.6. Since draft-ietf-httpbis-priority-00 B.7. Since draft-ietf-httpbis-priority-00
* Move text around (#1217, #1218) * Move text around (#1217, #1218)
* Editorial change to the default urgency. The value is 3, which * Editorial change to the default urgency. The value is 3, which
was always the intent of previous changes. was always the intent of previous changes.
B.7. Since draft-kazuho-httpbis-priority-04 B.8. Since draft-kazuho-httpbis-priority-04
* Minimize semantics of Urgency levels (#1023, #1026) * Minimize semantics of Urgency levels (#1023, #1026)
* Reduce guidance about how intermediary implements merging priority * Reduce guidance about how intermediary implements merging priority
signals (#1026) signals (#1026)
* Remove mention of CDN-Loop (#1062) * Remove mention of CDN-Loop (#1062)
* Editorial changes * Editorial changes
* Make changes due to WG adoption * Make changes due to WG adoption
* Removed outdated Consideration (#118) * Removed outdated Consideration (#118)
B.8. Since draft-kazuho-httpbis-priority-03 B.9. Since draft-kazuho-httpbis-priority-03
* Changed numbering from [-1,6] to [0,7] (#78) * Changed numbering from [-1,6] to [0,7] (#78)
* Replaced priority scheme negotiation with HTTP/2 priority * Replaced priority scheme negotiation with HTTP/2 priority
deprecation (#100) deprecation (#100)
* Shorten parameter names (#108) * Shorten parameter names (#108)
* Expand on considerations (#105, #107, #109, #110, #111, #113) * Expand on considerations (#105, #107, #109, #110, #111, #113)
B.9. Since draft-kazuho-httpbis-priority-02 B.10. Since draft-kazuho-httpbis-priority-02
* Consolidation of the problem statement (#61, #73) * Consolidation of the problem statement (#61, #73)
* Define SETTINGS_PRIORITIES for negotiation (#58, #69) * Define SETTINGS_PRIORITIES for negotiation (#58, #69)
* Define PRIORITY_UPDATE frame for HTTP/2 and HTTP/3 (#51) * Define PRIORITY_UPDATE frame for HTTP/2 and HTTP/3 (#51)
* Explain fairness issue and mitigations (#56) * Explain fairness issue and mitigations (#56)
B.10. Since draft-kazuho-httpbis-priority-01 B.11. Since draft-kazuho-httpbis-priority-01
* Explain how reprioritization might be supported. * Explain how reprioritization might be supported.
B.11. Since draft-kazuho-httpbis-priority-00 B.12. Since draft-kazuho-httpbis-priority-00
* Expand urgency levels from 3 to 8. * Expand urgency levels from 3 to 8.
Authors' Addresses Authors' Addresses
Kazuho Oku Kazuho Oku
Fastly Fastly
Email: kazuhooku@gmail.com Email: kazuhooku@gmail.com
 End of changes. 71 change blocks. 
233 lines changed or deleted 263 lines changed or added

This html diff was produced by rfcdiff 1.48. The latest version is available from http://tools.ietf.org/tools/rfcdiff/