Extensible Prioritization Scheme for HTTP

Status of This Memo

This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶

Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶

Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶

This Internet-Draft will expire on 12 June 2022.¶

Copyright Notice

Copyright (c) 2021 IETF Trust and the persons identified as the document authors. All rights reserved.¶

This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶

Table of Contents

1. Introduction

It is common for representations of an HTTP [HTTP] resource to have relationships to one or more other resources. Clients will often discover these relationships while processing a retrieved representation, which may lead to further retrieval requests. Meanwhile, the nature of the relationship determines whether the client is blocked from continuing to process locally available resources. An example of this is visual rendering of an HTML document, which could be blocked by the retrieval of a CSS file that the document refers to. In contrast, inline images do not block rendering and get drawn incrementally as the chunks of the images arrive.¶

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 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 of priority signals that communicate to the server a "priority tree"; the structure of this tree represents the client's preferred relative ordering and weighted distribution of the bandwidth among HTTP responses. Servers could use these priority signals as input into prioritization decision making.¶

The design and implementation of RFC 7540 stream priority was observed to have shortcomings, explained in Section 2. HTTP/2 [HTTP2] has consequently deprecated the use of these stream priority signals. The prioritization scheme and priority signals defined herein can act as a substitute for RFC 7540 stream priority.¶

This document describes an extensible scheme for prioritizing HTTP responses that uses absolute values. Section 4 defines priority parameters, which are a standardized and extensible format of priority information. Section 5 defines the Priority HTTP header field, a protocol-version-independent and end-to-end priority signal. Clients can send this header field to signal their view of how responses should be prioritized. Similarly, servers behind an intermediary can use it to signal priority to the intermediary. After sending a request, a client can change the priority of the response (see Section 6) using HTTP-version-specific frames defined in Section 7.1 and Section 7.2.¶

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.¶

2. Motivation for Replacing RFC 7540 Priorities

RFC 7540 stream priority (see Section 5.3 of [RFC7540]) is a complex system where clients signal stream dependencies and weights to describe an unbalanced tree. It suffered from limited deployment and interoperability and was deprecated in a revision of HTTP/2 [HTTP2]. HTTP/2 retains these protocol elements in order to maintain wire compatibility (see Section 5.3.2 of [HTTP2]), which means that they might still be used even in the presence of alternative signaling, such as the scheme this document describes.¶

Many RFC 7540 server implementations do not act on HTTP/2 priority signals.¶

Prioritization can use information that servers have about resources or the order in which requests are generated. For example, a server, with knowledge of an HTML document structure, might want to prioritize the delivery of images that are critical to user experience above other images. With RFC 7540 it is difficult for servers to interpret signals from clients for prioritization as the same conditions could result in very different signaling from different clients. This document describes signaling that is simpler and more constrained, requiring less interpretation and allowing less variation.¶

RFC 7540 does not define a method that can be used by a server to provide a priority signal for intermediaries.¶

RFC 7540 priority is expressed relative to other requests on the same connection. Many requests are generated without knowledge of how other requests might share a connection, which makes this difficult to use reliably, especially in protocols that do not have strong ordering guarantees, like HTTP/3 [HTTP3].¶

Multiple experiments from independent research ([MARX], [MEENAN]) have shown that simpler schemes can reach at least equivalent performance characteristics compared to the more complex RFC 7540 setups seen in practice, at least for the web use case.¶

3. Applicability of the Extensible Priority Scheme

The priority scheme defined by this document is primarily focused on the prioritization of HTTP response messages (see Section 3.4 of [HTTP]). It defines new priority parameters (Section 4) and their conveyors (Section 5 and Section 7) intended to communicate the priority of responses to a server that is responsible for prioritizing them. Section 10 provides considerations for servers about acting on those signals in combination with other inputs and factors.¶

The CONNECT method (see Section 9.3.6 of [HTTP]) can be used to establish tunnels. Signaling applies similarly to tunnels; additional considerations for server prioritization are given in Section 11.¶

Section 9 describes how clients can optionally apply elements of this scheme locally to the request messages that they generate.¶

Some forms of HTTP extensions might change HTTP/2 or HTTP/3 stream behavior or define new data carriage mechanisms. Such extensions can define themselves how this priority scheme is to be applied.¶

4. Priority Parameters

The priority information is a sequence of key-value pairs, providing room for future extensions. Each key-value pair represents a priority parameter.¶

The Priority HTTP header field (Section 5) is an end-to-end way to transmit this set of priority parameters when a request or a response is issued. In order to reprioritize a request (Section 6), HTTP-version-specific PRIORITY_UPDATE frames (Section 7.1 and Section 7.2) are used by clients to transmit the same information on a single hop.¶

Intermediaries can consume and produce priority signals in a PRIORITY_UPDATE frame or Priority header field. Sending a PRIORITY_UPDATE frame preserves the signal from the client, but provides a signal that overrides that for the next hop; see Section 14. Replacing or adding a Priority header field overrides any signal from a client and can affect prioritization for all subsequent recipients.¶

For both the Priority header field and the PRIORITY_UPDATE frame, the set of priority parameters is encoded as a Structured Fields Dictionary (see Section 3.2 of [STRUCTURED-FIELDS]).¶

This document defines the urgency(u) and incremental(i) priority parameters. When receiving an HTTP request that does not carry these priority parameters, a server SHOULD act as if their default values were specified.¶

An intermediary can combine signals from requests and responses that it forwards. Note that omission of priority parameters in responses is handled differently from omission in requests; see Section 8.¶

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, priority parameters with out-of-range values, or values of unexpected types MUST be ignored.¶

:method = GET
:scheme = https
:authority = example.net
:path = /style.css
priority = u=0

:method = GET
:scheme = https
:authority = example.net
:path = /image.jpg
priority = u=5, i

5. The Priority HTTP Header Field

The Priority HTTP header field carries priority parameters (see Section 4). It can appear in requests and responses. It is an end-to-end signal of the request priority from the client or the response priority from the server. Section 8 describes how intermediaries can combine the priority information sent from clients and servers. Clients cannot interpret the appearance or omission of a Priority response header field 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.¶

Priority is a Dictionary (Section 3.2 of [STRUCTURED-FIELDS]):¶

Priority   = sf-dictionary

An HTTP request with a Priority header field might be cached and re-used for subsequent requests; see [CACHING]. When an origin server generates the Priority response header field based on properties of an HTTP request it receives, the server is expected to control the cacheability or the applicability of the cached response, by using header fields that control the caching behavior (e.g., Cache-Control, Vary).¶

6. Reprioritization

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 prefetch request for a JavaScript file with the urgency parameter of the Priority request header field set to u=7 (background). Then, when the user navigates to a page which references the new JavaScript file, while the prefetch is in progress, the browser would send a reprioritization signal with the priority field value set to u=0. The PRIORITY_UPDATE frame (Section 7) can be used for such reprioritization.¶

7. The PRIORITY_UPDATE Frame

This document specifies a new PRIORITY_UPDATE frame for HTTP/2 [HTTP2] and HTTP/3 [HTTP3]. It carries priority parameters and references the target of the prioritization based on a version-specific identifier. In HTTP/2, this identifier is the Stream ID; in HTTP/3, the identifier is either the Stream ID or Push ID. Unlike the Priority header field, the PRIORITY_UPDATE frame is a hop-by-hop signal.¶

PRIORITY_UPDATE frames are sent by clients on the control stream, allowing them to be sent independent from the stream that carries the response. This means they can be used to reprioritize a response or a push stream; or signal the initial priority of a response instead of the Priority header field.¶

A PRIORITY_UPDATE frame communicates a complete set of all priority parameters in the Priority Field Value field. Omitting a priority parameter is a signal to use its default value. Failure to parse the Priority Field Value MAY be treated as a connection error. In HTTP/2 the error is of type PROTOCOL_ERROR; in HTTP/3 the error is of type H3_GENERAL_PROTOCOL_ERROR.¶

A client MAY send a PRIORITY_UPDATE frame before the stream that it references is open (except for HTTP/2 push streams; see Section 7.1). Furthermore, HTTP/3 offers no guaranteed ordering across streams, which could cause the frame to be received earlier than intended. Either case leads to a race condition where a server receives a PRIORITY_UPDATE frame that references a request stream that is yet to be opened. To solve this condition, for the purposes of scheduling, the most recently received PRIORITY_UPDATE frame can be considered as the most up-to-date information that overrides any other signal. Servers SHOULD buffer the most recently received PRIORITY_UPDATE frame and apply it once the referenced stream is opened. Holding PRIORITY_UPDATE frames for each stream requires server resources, which can can be bounded by local implementation policy. Although there is no limit to the number of PRIORITY_UPDATES that can be sent, storing only the most recently received frame limits resource commitment.¶

HTTP/2 PRIORITY_UPDATE Frame {
  Length (24),
  Type (i) = 10,

  Unused Flags (8).

  Reserved (1),
  Stream Identifier (31),

  Reserved (1),
  Prioritized Stream ID (31),
  Priority Field Value (..),
}

HTTP/3 PRIORITY_UPDATE Frame {
  Type (i) = 0xF0700..0xF0701,
  Length (i),
  Prioritized Element ID (i),
  Priority Field Value (..),
}

8. Merging Client- and Server-Driven Priority Parameters

It is not always the case that the client has the best understanding of how the HTTP responses deserve to be prioritized. The server might have additional information that can be combined with the client's indicated priority in order to improve the prioritization of the response. For example, use of an HTML document might depend heavily on one of the inline images; existence of such dependencies is typically best known to the server. Or, a server that receives requests for a font [RFC8081] and images with the same urgency might give higher precedence to the font, so that a visual client can render textual information at an early moment.¶

An origin can use the Priority response header field to indicate its view on how an HTTP response should be prioritized. An intermediary that forwards an HTTP response can use the priority parameters found in the Priority response header field, in combination with the client Priority request header field, as input to its prioritization process. No guidance is provided for merging priorities, this is left as an implementation decision.¶

Absence of a priority parameter in an HTTP response indicates the server's disinterest in changing the client-provided value. This is different from the the request header field, in which omission of a priority parameter implies the use of their default values (see Section 4).¶

As a non-normative example, when the client sends an HTTP request with the urgency parameter set to 5 and the incremental parameter set to true¶

:method = GET
:scheme = https
:authority = example.net
:path = /menu.png
priority = u=5, i

and the origin responds with¶

:status = 200
content-type = image/png
priority = u=1

the intermediary might alter its understanding of the urgency from 5 to 1, because it prefers the server-provided value over the client's. The incremental value continues to be true, the value specified by the client, as the server did not specify the incremental(i) parameter.¶

9. Client Scheduling

A client MAY use priority values to make local processing or scheduling choices about the requests it initiates.¶

10. Server Scheduling

It is generally beneficial for an HTTP server to send all responses as early as possible. However, when serving multiple requests on a single connection, there could be competition between the requests for resources such as connection bandwidth. This section describes considerations regarding how servers can schedule the order in which the competing responses will be sent, when such competition exists.¶

Server scheduling is a prioritization process based on many inputs, with priority signals being only one form of input. Factors such as implementation choices or deployment environment also play a role. Any given connection is likely to have many dynamic permutations. For these reasons, there is no unilateral perfect scheduler. This document provides some basic, non-exhaustive, recommendations for how servers might act on priority parameters. It does not describe in detail how servers might combine priority signals with other factors. Endpoints cannot depend on particular treatment based on priority signals. Expressing priority is only a suggestion.¶

It is RECOMMENDED that, when possible, servers respect the urgency parameter (Section 4.1), sending higher urgency responses before lower urgency responses.¶

The incremental parameter indicates how a client processes response bytes as they arrive. It is RECOMMENDED that, when possible, servers respect the incremental parameter (Section 4.2).¶

Non-incremental responses of the same urgency SHOULD be served by prioritizing bandwidth allocation in ascending order of 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 by sharing 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 resource followed by an incremental request for a small resource.¶
2. At the same urgency level, an incremental request of indeterminate length followed by a non-incremental large resource.¶

It is RECOMMENDED that servers avoid such starvation where possible. The method to do so is an implementation decision. For example, a server might pre-emptively send responses of a particular incremental type based on other information such as content size.¶

Optimal scheduling of server push is difficult, especially when pushed resources contend with active concurrent requests. Servers can consider many factors when scheduling, such as the type or size of resource being pushed, the priority of the request that triggered the push, the count of active concurrent responses, the priority of other active concurrent responses, etc. There is no general guidance 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 at too low a priority and delay the response, negating intended goals of server push.¶

Priority signals are a factor for server push scheduling. The concept of parameter value defaults applies slightly differently because there is no explicit client-signalled initial priority. A server can apply priority signals provided in an origin response; see the merging guidance given in Section 8. In the absence of origin signals, applying default parameter values could be suboptimal. By whatever means a server decides to schedule a pushed response, it can signal the intended priority to the client by including the Priority field in a PUSH_PROMISE or HEADERS frame.¶

11. Scheduling and the CONNECT Method

When a request stream carries the CONNECT method, the scheduling guidance in this document applies to the frames on the stream. A client that issues multiple CONNECT requests can set the incremental parameter to true. Servers that implement the recommendations for handling of the incremental parameter in Section 10 are likely to schedule these fairly, avoiding one CONNECT stream from blocking others.¶

12. Retransmission Scheduling

Transport protocols such as TCP and QUIC provide reliability by detecting packet losses and retransmitting lost information. In addition to the considerations in Section 10, scheduling of retransmission data could compete with new data. The remainder of this section discusses considerations when using QUIC.¶

Section 13.3 of [QUIC] states "Endpoints SHOULD prioritize retransmission of data over sending new data, unless priorities specified by the application indicate otherwise". When an HTTP/3 application uses the priority scheme defined in this document and the QUIC transport implementation supports application indicated stream priority, a transport that considers the relative priority of streams when scheduling both new data and retransmission data might better match the expectations of the application. However, there are no requirements on how a transport chooses to schedule based on this information because the decision depends on several factors and trade-offs. It could prioritize new data for a higher urgency stream over retransmission data for a lower priority stream, or it could prioritize retransmission data over new data irrespective of urgencies.¶

Section 6.2.4 of [QUIC-RECOVERY] also highlights consideration of application priorities when sending probe packets after Probe Timeout timer expiration. A QUIC implementation supporting application-indicated priorities might use the relative priority of streams when choosing probe data.¶

13. Fairness

Typically, HTTP implementations depend on the underlying transport to maintain fairness between connections competing for bandwidth. When HTTP requests are forwarded through intermediaries, progress made by each connection originating from end clients can become different over time, depending on how intermediaries coalesce or split requests into backend connections. This unfairness can expand if priority signals are used. Section 13.1 and Section 13.2 discuss mitigations against this expansion of unfairness.¶

Conversely, Section 13.3 discusses how servers might intentionally allocate unequal bandwidth to some connections depending on the priority signals.¶

14. Why use an End-to-End Header Field?

In contrast to the prioritization scheme of HTTP/2 that uses a hop-by-hop frame, the Priority header field is defined as end-to-end.¶

The way that a client processes a response is a property associated with the client generating that request. Not that of an intermediary. Therefore, it is an end-to-end property. How these end-to-end properties carried by the Priority header field affect the prioritization between the responses that share a connection is a hop-by-hop issue.¶

Having the Priority header field defined as end-to-end is important for caching intermediaries. Such intermediaries can cache the value of the Priority header field along with the response, and utilize the value of the cached header field when serving the cached response, only because the header field is defined as end-to-end rather than hop-by-hop.¶

15. Security Considerations

Section 7 describes considerations for server buffering of PRIORITY_UPDATE frames.¶

Section 10 presents examples where servers that prioritize responses in a certain way might be starved of the ability to transmit payload.¶

The security considerations from [STRUCTURED-FIELDS] apply to processing of priority parameters defined in Section 4.¶

16. IANA Considerations

This specification registers the following entry in the the Hypertext Transfer Protocol (HTTP) Field Name Registry established by [HTTP]:¶

Field name:

Priority¶

Status:

permanent¶

Specification document(s):

This document¶

This specification registers the following entry in the HTTP/2 Settings registry established by [RFC7540]:¶

Name:

SETTINGS_NO_RFC7540_PRIORITIES¶

Code:

0x9¶

Initial value:

0¶

Specification:

This document¶

This specification registers the following entry in the HTTP/2 Frame Type registry established by [RFC7540]:¶

Frame Type:

PRIORITY_UPDATE¶

Code:

0x10¶

Specification:

This document¶

This specification registers the following entries in the HTTP/3 Frame Type registry established by [HTTP3]:¶

Frame Type:

PRIORITY_UPDATE¶

Code:

0xF0700 and 0xF0701¶

Specification:

This document¶

Upon publication, please create the HTTP Priority Parameters registry at https://iana.org/assignments/http-priority and populate it with the entries in Table 1; see Section 4.3.1 for its associated procedures.¶

Appendix A. Acknowledgements

Roy Fielding presented the idea of using a header field for representing priorities in http://tools.ietf.org/agenda/83/slides/slides-83-httpbis-5.pdf. In https://github.com/pmeenan/http3-prioritization-proposal, Patrick Meenan advocated for representing the priorities using a tuple of urgency and concurrency. The ability to disable HTTP/2 prioritization is inspired by [I-D.lassey-priority-setting], authored by Brad Lassey and Lucas Pardue, with modifications based on feedback that was not incorporated into an update to that document.¶

The motivation for defining an alternative to HTTP/2 priorities is drawn from discussion within the broad HTTP community. Special thanks to Roberto Peon, Martin Thomson and Netflix for text that was incorporated explicitly in this document.¶

In addition to the people above, this document owes a lot to the extensive discussion in the HTTP priority design team, consisting of Alan Frindell, Andrew Galloni, Craig Taylor, Ian Swett, Kazuho Oku, Lucas Pardue, Matthew Cox, Mike Bishop, Roberto Peon, Robin Marx, Roy Fielding.¶

Yang Chi contributed the section on retransmission scheduling.¶

Appendix B. Change Log

RFC EDITOR: please remove this section before publication¶

Authors' Addresses