draft-ietf-httpbis-http2-02.txt   draft-ietf-httpbis-http2-03.txt 
HTTPbis Working Group M. Belshe HTTPbis Working Group M. Belshe
Internet-Draft Twist Internet-Draft Twist
Intended status: Standards Track R. Peon Intended status: Standards Track R. Peon
Expires: October 5, 2013 Google, Inc Expires: November 30, 2013 Google, Inc
M. Thomson, Ed. M. Thomson, Ed.
Microsoft Microsoft
A. Melnikov, Ed. A. Melnikov, Ed.
Isode Ltd Isode Ltd
April 3, 2013 May 29, 2013
Hypertext Transfer Protocol version 2.0 Hypertext Transfer Protocol version 2.0
draft-ietf-httpbis-http2-02 draft-ietf-httpbis-http2-03
Abstract Abstract
This specification describes an optimised expression of the syntax of This specification describes an optimized expression of the syntax of
the Hypertext Transfer Protocol (HTTP). The HTTP/2.0 encapsulation the Hypertext Transfer Protocol (HTTP). The HTTP/2.0 encapsulation
enables more efficient transfer of representations by providing enables more efficient use of network resources and reduced
compressed header fields, simultaneous requests, and also introduces perception of latency by allowing header field compression and
unsolicited push of representations from server to client. multiple concurrent messages on the same connection. It also
introduces unsolicited push of representations from servers to
clients.
This document is an alternative to, but does not obsolete the HTTP This document is an alternative to, but does not obsolete the
message format. HTTP semantics remain unchanged. HTTP/1.1 message format or protocol. HTTP's existing semantics
remain unchanged.
Editorial Note (To be removed by RFC Editor) Editorial Note (To be removed by RFC Editor)
Discussion of this draft takes place on the HTTPBIS working group Discussion of this draft takes place on the HTTPBIS working group
mailing list (ietf-http-wg@w3.org), which is archived at mailing list (ietf-http-wg@w3.org), which is archived at
<http://lists.w3.org/Archives/Public/ietf-http-wg/>. <http://lists.w3.org/Archives/Public/ietf-http-wg/>.
Working Group information and related documents can be found at Working Group information and related documents can be found at
<http://tools.ietf.org/wg/httpbis/> (Wiki) and <http://tools.ietf.org/wg/httpbis/> (Wiki) and
<https://github.com/http2/http2-spec> (source code and issues <https://github.com/http2/http2-spec> (source code and issues
skipping to change at page 2, line 10 skipping to change at page 2, line 12
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 http://datatracker.ietf.org/drafts/current/. Drafts is at http://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 October 5, 2013. This Internet-Draft will expire on November 30, 2013.
Copyright Notice Copyright Notice
Copyright (c) 2013 IETF Trust and the persons identified as the Copyright (c) 2013 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 Provisions Relating to IETF Documents
(http://trustee.ietf.org/license-info) in effect on the date of (http://trustee.ietf.org/license-info) in effect on the date of
publication of this document. Please review these documents publication of this document. Please review these documents
skipping to change at page 2, line 32 skipping to change at page 2, line 34
to this document. Code Components extracted from this document must to this document. Code Components extracted from this document must
include Simplified BSD License text as described in Section 4.e of include Simplified BSD License text as described in Section 4.e of
the Trust Legal Provisions and are provided without warranty as the Trust Legal Provisions and are provided without warranty as
described in the Simplified BSD License. described in the Simplified BSD License.
Table of Contents Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 5 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.1. Document Organization . . . . . . . . . . . . . . . . . . 5 1.1. Document Organization . . . . . . . . . . . . . . . . . . 5
1.2. Conventions and Terminology . . . . . . . . . . . . . . . 6 1.2. Conventions and Terminology . . . . . . . . . . . . . . . 6
2. Starting HTTP/2.0 . . . . . . . . . . . . . . . . . . . . . . 7 2. Starting HTTP/2.0 . . . . . . . . . . . . . . . . . . . . . . 6
2.1. HTTP/2.0 Version Identification . . . . . . . . . . . . . 7 2.1. HTTP/2.0 Version Identification . . . . . . . . . . . . . 7
2.2. Starting HTTP/2.0 for "http:" URIs . . . . . . . . . . . . 8 2.2. Starting HTTP/2.0 for "http:" URIs . . . . . . . . . . . . 8
2.3. Starting HTTP/2.0 for "https:" URIs . . . . . . . . . . . 8 2.3. Starting HTTP/2.0 for "https:" URIs . . . . . . . . . . . 8
2.4. Starting HTTP/2.0 with Prior Knowledge . . . . . . . . . . 9 2.4. Starting HTTP/2.0 with Prior Knowledge . . . . . . . . . . 9
3. HTTP/2.0 Framing Layer . . . . . . . . . . . . . . . . . . . . 9 3. HTTP/2.0 Framing Layer . . . . . . . . . . . . . . . . . . . . 9
3.1. Session . . . . . . . . . . . . . . . . . . . . . . . . . 9 3.1. Connection . . . . . . . . . . . . . . . . . . . . . . . . 9
3.2. Session Header . . . . . . . . . . . . . . . . . . . . . . 9 3.2. Connection Header . . . . . . . . . . . . . . . . . . . . 9
3.3. Framing . . . . . . . . . . . . . . . . . . . . . . . . . 10 3.3. Framing . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.3.1. Frame Header . . . . . . . . . . . . . . . . . . . . . 10 3.3.1. Frame Header . . . . . . . . . . . . . . . . . . . . . 10
3.3.2. Frame Processing . . . . . . . . . . . . . . . . . . . 11 3.3.2. Frame Size . . . . . . . . . . . . . . . . . . . . . . 12
3.4. Streams . . . . . . . . . . . . . . . . . . . . . . . . . 11 3.4. Streams . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.4.1. Stream Creation . . . . . . . . . . . . . . . . . . . 12 3.4.1. Stream Creation . . . . . . . . . . . . . . . . . . . 13
3.4.2. Stream priority . . . . . . . . . . . . . . . . . . . 12 3.4.2. Stream priority . . . . . . . . . . . . . . . . . . . 13
3.4.3. Stream headers . . . . . . . . . . . . . . . . . . . . 13 3.4.3. Stream half-close . . . . . . . . . . . . . . . . . . 14
3.4.4. Stream data exchange . . . . . . . . . . . . . . . . . 13 3.4.4. Stream close . . . . . . . . . . . . . . . . . . . . . 14
3.4.5. Stream half-close . . . . . . . . . . . . . . . . . . 13 3.5. Error Handling . . . . . . . . . . . . . . . . . . . . . . 15
3.4.6. Stream close . . . . . . . . . . . . . . . . . . . . . 13 3.5.1. Connection Error Handling . . . . . . . . . . . . . . 15
3.5. Error Handling . . . . . . . . . . . . . . . . . . . . . . 14 3.5.2. Stream Error Handling . . . . . . . . . . . . . . . . 16
3.5.1. Session Error Handling . . . . . . . . . . . . . . . . 14 3.5.3. Error Codes . . . . . . . . . . . . . . . . . . . . . 16
3.5.2. Stream Error Handling . . . . . . . . . . . . . . . . 15 3.6. Stream Flow Control . . . . . . . . . . . . . . . . . . . 17
3.5.3. Error Codes . . . . . . . . . . . . . . . . . . . . . 15 3.6.1. Flow Control Principles . . . . . . . . . . . . . . . 17
3.6. Stream Flow Control . . . . . . . . . . . . . . . . . . . 16 3.6.2. Appropriate Use of Flow Control . . . . . . . . . . . 18
3.6.1. Flow Control Principles . . . . . . . . . . . . . . . 16 3.7. Header Blocks . . . . . . . . . . . . . . . . . . . . . . 19
3.6.2. Appropriate Use of Flow Control . . . . . . . . . . . 17 3.8. Frame Types . . . . . . . . . . . . . . . . . . . . . . . 19
3.7. Frame Types . . . . . . . . . . . . . . . . . . . . . . . 18 3.8.1. DATA Frames . . . . . . . . . . . . . . . . . . . . . 20
3.7.1. DATA Frames . . . . . . . . . . . . . . . . . . . . . 18 3.8.2. HEADERS+PRIORITY . . . . . . . . . . . . . . . . . . . 20
3.7.2. HEADERS+PRIORITY . . . . . . . . . . . . . . . . . . . 18 3.8.3. RST_STREAM . . . . . . . . . . . . . . . . . . . . . . 21
3.7.3. RST_STREAM . . . . . . . . . . . . . . . . . . . . . . 18 3.8.4. SETTINGS . . . . . . . . . . . . . . . . . . . . . . . 21
3.7.4. SETTINGS . . . . . . . . . . . . . . . . . . . . . . . 19 3.8.5. PUSH_PROMISE . . . . . . . . . . . . . . . . . . . . . 25
3.7.5. PUSH_PROMISE . . . . . . . . . . . . . . . . . . . . . 22 3.8.6. PING . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.7.6. PING . . . . . . . . . . . . . . . . . . . . . . . . . 23 3.8.7. GOAWAY . . . . . . . . . . . . . . . . . . . . . . . . 26
3.7.7. GOAWAY . . . . . . . . . . . . . . . . . . . . . . . . 23 3.8.8. HEADERS . . . . . . . . . . . . . . . . . . . . . . . 28
3.7.8. HEADERS . . . . . . . . . . . . . . . . . . . . . . . 24 3.8.9. WINDOW_UPDATE . . . . . . . . . . . . . . . . . . . . 29
3.7.9. WINDOW_UPDATE . . . . . . . . . . . . . . . . . . . . 25 4. HTTP Message Exchanges . . . . . . . . . . . . . . . . . . . . 32
3.7.10. Header Block . . . . . . . . . . . . . . . . . . . . . 28 4.1. Connection Management . . . . . . . . . . . . . . . . . . 32
4. HTTP Message Exchanges . . . . . . . . . . . . . . . . . . . . 28 4.2. HTTP Request/Response . . . . . . . . . . . . . . . . . . 33
4.1. Connection Management . . . . . . . . . . . . . . . . . . 28 4.2.1. HTTP Header Fields and HTTP/2.0 Headers . . . . . . . 33
4.1.1. Use of GOAWAY . . . . . . . . . . . . . . . . . . . . 29 4.2.2. Request . . . . . . . . . . . . . . . . . . . . . . . 33
4.2. HTTP Request/Response . . . . . . . . . . . . . . . . . . 29 4.2.3. Response . . . . . . . . . . . . . . . . . . . . . . . 34
4.2.1. HTTP Header Fields and HTTP/2.0 Headers . . . . . . . 29 4.3. Server Push Transactions . . . . . . . . . . . . . . . . . 35
4.2.2. Request . . . . . . . . . . . . . . . . . . . . . . . 29 4.3.1. Server implementation . . . . . . . . . . . . . . . . 36
4.2.3. Response . . . . . . . . . . . . . . . . . . . . . . . 31 4.3.2. Client implementation . . . . . . . . . . . . . . . . 37
4.3. Server Push Transactions . . . . . . . . . . . . . . . . . 32 5. Design Rationale and Notes . . . . . . . . . . . . . . . . . . 38
4.3.1. Server implementation . . . . . . . . . . . . . . . . 33 5.1. Separation of Framing Layer and Application Layer . . . . 38
4.3.2. Client implementation . . . . . . . . . . . . . . . . 34 5.2. Error handling - Framing Layer . . . . . . . . . . . . . . 39
5. Design Rationale and Notes . . . . . . . . . . . . . . . . . . 35 5.3. One Connection per Domain . . . . . . . . . . . . . . . . 39
5.1. Separation of Framing Layer and Application Layer . . . . 35 5.4. Fixed vs Variable Length Fields . . . . . . . . . . . . . 39
5.2. Error handling - Framing Layer . . . . . . . . . . . . . . 35 5.5. Server Push . . . . . . . . . . . . . . . . . . . . . . . 40
5.3. One Connection Per Domain . . . . . . . . . . . . . . . . 36 6. Security Considerations . . . . . . . . . . . . . . . . . . . 40
5.4. Fixed vs Variable Length Fields . . . . . . . . . . . . . 36 6.1. Server Authority and Same-Origin . . . . . . . . . . . . . 40
5.5. Server Push . . . . . . . . . . . . . . . . . . . . . . . 36 6.2. Cross-Protocol Attacks . . . . . . . . . . . . . . . . . . 40
6. Security Considerations . . . . . . . . . . . . . . . . . . . 37 6.3. Cacheability of Pushed Resources . . . . . . . . . . . . . 41
6.1. Use of Same-origin constraints . . . . . . . . . . . . . . 37 7. Privacy Considerations . . . . . . . . . . . . . . . . . . . . 41
6.2. Cross-Protocol Attacks . . . . . . . . . . . . . . . . . . 37 7.1. Long Lived Connections . . . . . . . . . . . . . . . . . . 41
6.3. Cacheability of Pushed Resources . . . . . . . . . . . . . 37 7.2. SETTINGS frame . . . . . . . . . . . . . . . . . . . . . . 41
7. Privacy Considerations . . . . . . . . . . . . . . . . . . . . 37 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 42
7.1. Long Lived Connections . . . . . . . . . . . . . . . . . . 38 8.1. Frame Type Registry . . . . . . . . . . . . . . . . . . . 42
7.2. SETTINGS frame . . . . . . . . . . . . . . . . . . . . . . 38 8.2. Error Code Registry . . . . . . . . . . . . . . . . . . . 43
8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 38 8.3. Settings Registry . . . . . . . . . . . . . . . . . . . . 43
8.1. Frame Type Registry . . . . . . . . . . . . . . . . . . . 38 9. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 44
8.2. Error Code Registry . . . . . . . . . . . . . . . . . . . 39 10. References . . . . . . . . . . . . . . . . . . . . . . . . . . 44
8.3. Settings Registry . . . . . . . . . . . . . . . . . . . . 39 10.1. Normative References . . . . . . . . . . . . . . . . . . . 44
9. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 40 10.2. Informative References . . . . . . . . . . . . . . . . . . 45
10. References . . . . . . . . . . . . . . . . . . . . . . . . . . 41
10.1. Normative References . . . . . . . . . . . . . . . . . . . 41
10.2. Informative References . . . . . . . . . . . . . . . . . . 42
Appendix A. Change Log (to be removed by RFC Editor before Appendix A. Change Log (to be removed by RFC Editor before
publication) . . . . . . . . . . . . . . . . . . . . 42 publication) . . . . . . . . . . . . . . . . . . . . 46
A.1. Since draft-ietf-httpbis-http2-01 . . . . . . . . . . . . 42 A.1. Since draft-ietf-httpbis-http2-02 . . . . . . . . . . . . 46
A.2. Since draft-ietf-httpbis-http2-00 . . . . . . . . . . . . 43 A.2. Since draft-ietf-httpbis-http2-01 . . . . . . . . . . . . 46
A.3. Since draft-mbelshe-httpbis-spdy-00 . . . . . . . . . . . 43 A.3. Since draft-ietf-httpbis-http2-00 . . . . . . . . . . . . 47
A.4. Since draft-mbelshe-httpbis-spdy-00 . . . . . . . . . . . 47
1. Introduction 1. Introduction
The Hypertext Transfer Protocol (HTTP) is a wildly successful The Hypertext Transfer Protocol (HTTP) is a wildly successful
protocol. The HTTP/1.1 message encapsulation ([HTTP-p1], Section 3) protocol. However, the HTTP/1.1 message encapsulation ([HTTP-p1],
is optimized for implementation simplicity and accessibility, not Section 3) is optimized for implementation simplicity and
application performance. As such it has several characteristics that accessibility, not application performance. As such it has several
have a negative overall effect on application performance. characteristics that have a negative overall effect on application
performance.
The HTTP/1.1 encapsulation ensures that only one request can be In particular, HTTP/1.0 only allows one request to be delivered at a
delivered at a time on a given connection. HTTP/1.1 pipelining, time on a given connection. HTTP/1.1 pipelining only partially
which is not widely deployed, only partially addresses these addressed request concurrency, and is not widely deployed.
concerns. Clients that need to make multiple requests therefore use Therefore, clients that need to make many requests (as is common on
commonly multiple connections to a server or servers in order to the Web) typically use multiple connections to a server in order to
reduce the overall latency of those requests. [[anchor1: Need to tune reduce perceived latency.
the anti-pipelining comments here.]]
Furthermore, HTTP/1.1 header fields are represented in an inefficient Furthermore, HTTP/1.1 header fields are often repetitive and verbose,
fashion, which, in addition to generating more or larger network which, in addition to generating more or larger network packets, can
packets, can cause the small initial TCP window to fill more quickly cause the small initial TCP congestion window to quickly fill. This
than is ideal. This results in excessive latency where multiple can result in excessive latency when multiple requests are made on a
requests are made on a new TCP connection. single new TCP connection.
This document defines an optimized mapping of the HTTP semantics to a This document addresses these issues by defining an optimized mapping
TCP connection. This optimization reduces the latency costs of HTTP of HTTP's semantics to an underlying connection. Specifically, it
by allowing parallel requests on the same connection and by using an allows interleaving of request and response messages on the same
efficient coding for HTTP header fields. Prioritization of requests connection and uses an efficient coding for HTTP header fields. It
lets more important requests complete faster, further improving also allows prioritization of requests, letting more important
application performance. requests complete more quickly, further improving perceived
performance.
HTTP/2.0 applications have an improved impact on network congestion The resulting protocol is designed to have be more friendly to the
due to the use of fewer TCP connections to achieve the same effect. network, because fewer TCP connections can be used, in comparison to
Fewer TCP connections compete more fairly with other flows. Long- HTTP/1.x. This means less competition with other flows, and longer-
lived connections are also more able to take better advantage of the lived connections, which in turn leads to better utilization of
available network capacity, rather than operating in the slow start available network capacity.
phase of TCP.
The HTTP/2.0 encapsulation also enables more efficient processing of Finally, this encapsulation also enables more scalable processing of
messages by providing efficient message framing. Processing of messages through use of binary message framing.
header fields in HTTP/2.0 messages is more efficient (for entities
that process many messages).
1.1. Document Organization 1.1. Document Organization
The HTTP/2.0 Specification is split into three parts: starting The HTTP/2.0 Specification is split into three parts: starting
HTTP/2.0 (Section 2), which covers how a HTTP/2.0 is started; a HTTP/2.0 (Section 2), which covers how a HTTP/2.0 connection is
framing layer (Section 3), which multiplexes a TCP connection into initiated; a framing layer (Section 3), which multiplexes a single
independent, length-prefixed frames; and an HTTP layer (Section 4), TCP connection into independent frames of various types; and an HTTP
which specifies the mechanism for overlaying HTTP request/response layer (Section 4), which specifies the mechanism for expressing HTTP
pairs on top of the framing layer. While some of the framing layer interactions using the framing layer. While some of the framing
concepts are isolated from the HTTP layer, building a generic framing layer concepts are isolated from HTTP, building a generic framing
layer has not been a goal. The framing layer is tailored to the layer has not been a goal. The framing layer is tailored to the
needs of the HTTP protocol and server push. needs of the HTTP protocol and server push.
1.2. Conventions and Terminology 1.2. Conventions and Terminology
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 RFC 2119 [RFC2119]. document are to be interpreted as described in RFC 2119 [RFC2119].
All numeric values are in network byte order. Values are unsigned All numeric values are in network byte order. Values are unsigned
unless otherwise indicated. Literal values are provided in decimal unless otherwise indicated. Literal values are provided in decimal
or hexadecimal as appropriate. Hexadecimal literals are prefixed or hexadecimal as appropriate. Hexadecimal literals are prefixed
with "0x" to distinguish them from decimal literals. with "0x" to distinguish them from decimal literals.
The following terms are used: The following terms are used:
client: The endpoint initiating the HTTP/2.0 session. client: The endpoint initiating the HTTP connection.
connection: A transport-level connection between two endpoints. connection: A transport-level connection between two endpoints.
endpoint: Either the client or server of a connection. endpoint: Either the client or server of the connection.
frame: The smallest unit of communication, each containing a frame frame: The smallest unit of communication within an HTTP/2.0
header. connection, consisting of a header and a variable-length sequence
of bytes structured according to the frame type.
message: A complete sequence of frames. peer: An endpoint. When discussing a particular endpoint, "peer"
refers to the endpoint that is remote to the primary subject of
discussion.
receiver: An endpoint that is receiving frames. receiver: An endpoint that is receiving frames.
sender: An endpoint that is transmitting frames. sender: An endpoint that is transmitting frames.
server: The endpoint which did not initiate the HTTP/2.0 session. server: The endpoint which did not initiate the HTTP connection.
session: A synonym for a connection.
session error: An error on the HTTP/2.0 session. connection error: An error on the HTTP/2.0 connection.
stream: A bi-directional flow of bytes across a virtual channel stream: A bi-directional flow of frames across a virtual channel
within a HTTP/2.0 session. within the HTTP/2.0 connection.
stream error: An error on an individual HTTP/2.0 stream. stream error: An error on the individual HTTP/2.0 stream.
2. Starting HTTP/2.0 2. Starting HTTP/2.0
Just as HTTP/1.1 does, HTTP/2.0 uses the "http:" and "https:" URI HTTP/2.0 uses the same "http:" and "https:" URI schemes used by
schemes. An HTTP/2.0-capable client is therefore required to HTTP/1.1. As a result, implementations processing requests for
discover whether a server (or intermediary) supports HTTP/2.0. target resource URIs like "http://example.org/foo" or
"https://example.com/bar" are required to first discover whether the
upstream server (the immediate peer to which the client wishes to
establish a connection) supports HTTP/2.0.
Different discovery mechanisms are defined for "http:" and "https:" The means by which support for HTTP/2.0 is determined is different
URIs. Discovery for "http:" URIs is described in Section 2.2; for "http" and "https" URIs. Discovery for "https:" URIs is
discovery for "https:" URIs is described in Section 2.3. described in Section 2.3. Discovery for "http" URIs is described
here.
2.1. HTTP/2.0 Version Identification 2.1. HTTP/2.0 Version Identification
HTTP/2.0 is identified using the string "HTTP/2.0". This The protocol defined in this document is identified using the string
identification is used in the HTTP/1.1 Upgrade header field, in the "HTTP/2.0". This identification is used in the HTTP/1.1 Upgrade
TLS-NPN [TLSNPN] [[anchor4: TBD]] field and other places where header field, in the TLS application layer protocol negotiation
protocol identification is required. extension [TLSALPN] field and other places where protocol
identification is required.
Negotiating "HTTP/2.0" implies the use of the transport, security, Negotiating "HTTP/2.0" implies the use of the transport, security,
framing and message semantics described in this document. framing and message semantics described in this document.
[[anchor5: Editor's Note: please remove the following text prior to [[anchor3: Editor's Note: please remove the following text prior to
the publication of a final version of this document.]] the publication of a final version of this document.]]
Only implementations of the final, published RFC can identify Only implementations of the final, published RFC can identify
themselves as "HTTP/2.0". Until such an RFC exists, implementations themselves as "HTTP/2.0". Until such an RFC exists, implementations
MUST NOT identify themselves using "HTTP/2.0". MUST NOT identify themselves using "HTTP/2.0".
Examples and text throughout the rest of this document use "HTTP/2.0" Examples and text throughout the rest of this document use "HTTP/2.0"
as a matter of editorial convenience only. Implementations of draft as a matter of editorial convenience only. Implementations of draft
versions MUST NOT identify using this string. versions MUST NOT identify using this string.
Implementations of draft versions of the protocol MUST add the string Implementations of draft versions of the protocol MUST add the string
"-draft-" and the corresponding draft number to the identifier before "-draft-" and the corresponding draft number to the identifier before
the separator ('/'). For example, draft-ietf-httpbis-http2-03 is the separator ('/'). For example, draft-ietf-httpbis-http2-03 is
identified using the string "HTTP-draft-03/2.0". identified using the string "HTTP-draft-03/2.0".
Non-compatible experiments that are based on these draft versions Non-compatible experiments that are based on these draft versions
MUST instead replace the string "draft" with a different identifier. MUST instead replace the string "draft" with a different identifier.
For example, an experimental implementation of packet mood-based For example, an experimental implementation of packet mood-based
encoding based on draft-ietf-httpbis-http2-07 might identify itself encoding based on draft-ietf-httpbis-http2-07 might identify itself
as "HTTP-emo-07/2.0". Note that any label MUST conform with the as "HTTP-emo-07/2.0". Note that any label MUST conform to the
"token" syntax defined in Section 3.2.6 of [HTTP-p1]. Experimenters "token" syntax defined in Section 3.2.6 of [HTTP-p1]. Experimenters
are encouraged to coordinate their experiments on the are encouraged to coordinate their experiments on the
ietf-http-wg@w3.org mailing list. ietf-http-wg@w3.org mailing list.
2.2. Starting HTTP/2.0 for "http:" URIs 2.2. Starting HTTP/2.0 for "http:" URIs
A client that makes a request to an "http:" URI without prior A client that makes a request to an "http:" URI without prior
knowledge about support for HTTP/2.0 uses the HTTP Upgrade mechanism knowledge about support for HTTP/2.0 uses the HTTP Upgrade mechanism
(Section 6.7 of [HTTP-p1]). The client makes an HTTP/1.1 request (Section 6.7 of [HTTP-p1]). The client makes an HTTP/1.1 request
that includes an Upgrade header field identifying HTTP/2.0. that includes an Upgrade header field identifying HTTP/2.0.
skipping to change at page 8, line 37 skipping to change at page 8, line 37
A server that supports HTTP/2.0 can accept the upgrade with a 101 A server that supports HTTP/2.0 can accept the upgrade with a 101
(Switching Protocols) status code. After the empty line that (Switching Protocols) status code. After the empty line that
terminates the 101 response, the server can begin sending HTTP/2.0 terminates the 101 response, the server can begin sending HTTP/2.0
frames. These frames MUST include a response to the request that frames. These frames MUST include a response to the request that
initiated the Upgrade. initiated the Upgrade.
HTTP/1.1 101 Switching Protocols HTTP/1.1 101 Switching Protocols
Connection: Upgrade Connection: Upgrade
Upgrade: HTTP/2.0 Upgrade: HTTP/2.0
[ HTTP/2.0 session ... [ HTTP/2.0 connection ...
Once the server returns the 101 response, both the client and the The first HTTP/2.0 frame sent by the server is a SETTINGS frame
server send a session header (Section 3.2). (Section 3.8.4). Upon receiving the 101 response, the client sends a
connection header (Section 3.2), which includes a SETTINGS frame.
2.3. Starting HTTP/2.0 for "https:" URIs 2.3. Starting HTTP/2.0 for "https:" URIs
A client that makes a request to an "https:" URI without prior A client that makes a request to an "https:" URI without prior
knowledge about support for HTTP/2.0 uses TLS [RFC5246] with TLS-NPN knowledge about support for HTTP/2.0 uses TLS [RFC5246] with the
[TLSNPN] extension. [[anchor6: TBD, maybe ALPN]] application layer protocol negotiation extension [TLSALPN].
Once TLS negotiation is complete, both the client and the server send Once TLS negotiation is complete, both the client and the server send
a session header (Section 3.2). a connection header (Section 3.2).
2.4. Starting HTTP/2.0 with Prior Knowledge 2.4. Starting HTTP/2.0 with Prior Knowledge
A client can learn that a particular server supports HTTP/2.0 by A client can learn that a particular server supports HTTP/2.0 by
other means. A client MAY immediately send HTTP/2.0 frames to a other means. A client MAY immediately send HTTP/2.0 frames to a
server that is known to support HTTP/2.0. This only affects the server that is known to support HTTP/2.0. This only affects the
resolution of "http:" URIs, servers supporting HTTP/2.0 are required resolution of "http:" URIs, servers supporting HTTP/2.0 are required
to support protocol negotiation in TLS [TLSNPN]. to support protocol negotiation in TLS [TLSALPN] for "https:" URIs.
Prior support for HTTP/2.0 is not a strong signal that a given server Prior support for HTTP/2.0 is not a strong signal that a given server
will support HTTP/2.0 for future sessions. It is possible for server will support HTTP/2.0 for future connections. It is possible for
configurations to change or for configurations to differ between server configurations to change or for configurations to differ
instances in clustered server. Different "transparent" between instances in clustered server. Interception proxies (a.k.a.
intermediaries - intermediaries that are not explicitly selected by "transparent" proxies) are another source of variability.
either client or server - are another source of variability.
3. HTTP/2.0 Framing Layer 3. HTTP/2.0 Framing Layer
3.1. Session 3.1. Connection
The HTTP/2.0 session runs atop TCP ([RFC0793]). The client is the The HTTP/2.0 connection is an Application Level protocol running on
TCP connection initiator. top of a TCP connection ([RFC0793]). The client is the TCP
connection initiator.
HTTP/2.0 connections are persistent connections. For best HTTP/2.0 connections are persistent. That is, for best performance,
performance, it is expected that clients will not close open it is expected a clients will not close connections until it is
connections until the user navigates away from all web pages determined that no further communication with a server is necessary
referencing a connection, or until the server closes the connection. (for example, when a user navigates away from a particular web page),
Servers are encouraged to leave connections open for as long as or until the server closes the connection.
possible, but can terminate idle connections if necessary. When
either endpoint closes the transport-level connection, it MUST first
send a GOAWAY (Section 3.7.7) frame so that the endpoints can
reliably determine if requests finished before the close.
3.2. Session Header Servers are encouraged to maintain open connections for as long as
possible, but are permitted to terminate idle connections if
necessary. When either endpoint chooses to close the transport-level
TCP connection, the terminating endpoint MUST first send a GOAWAY
(Section 3.8.7) frame so that both endpoints can reliably determine
whether previously sent frames have been processed and gracefully
complete or terminate any necessary remaining tasks.
After opening a TCP connection and performing either an HTTP/1.1 3.2. Connection Header
Upgrade or TLS handshake, the client sends the client session header.
The server replies with a server session header.
The session header provides a final confirmation that both peers Upon establishment of a TCP connection and determination that
agree to use the HTTP/2.0 protocol. The SETTINGS frame ensures that HTTP/2.0 will be used by both peers to communicate, each endpoint
client or server configuration is known as quickly as possible. MUST send a connection header as a final confirmation and to
establish the default parameters for the HTTP/2.0 connection.
The client session header is the 25 byte sequence The client connection header is a sequence of 24 octets (in hex
0x464f4f202a20485454502f322e300d0a0d0a4241520d0a0d0a (the string "FOO notation)
* HTTP/2.0\r\n\r\nBAR\r\n\r\n") followed by a SETTINGS frame
(Section 3.7.4). The client sends the client session header
immediately after receiving an HTTP/1.1 Upgrade, or after receiving a
TLS Finished message from the server.
The client session header is selected so that a large proportion 464f4f202a20485454502f322e300d0a0d0a42410d0a0d0a
of HTTP/1.1 or HTTP/1.0 servers and intermediaries do not attempt (the string "FOO * HTTP/2.0\r\n\r\nBA\r\n\r\n") followed by a
to process further frames. This doesn't address the concerns SETTINGS frame (Section 3.8.4). The client sends the client
raised in [TALKING]. connection header immediately upon receipt of a 101 Switching
Protocols response (indicating a successful upgrade), or after
receiving a TLS Finished message from the server. If starting an
HTTP/2.0 connection with prior knowledge of server support for the
protocol, the client connection header is sent upon connection
establishment.
The server session header is a SETTINGS frame (Section 3.7.4). The The client connection header is selected so that a large
server sends the server session header immediately after receiving proportion of HTTP/1.1 or HTTP/1.0 servers and intermediaries do
and validating the client session header. not attempt to process further frames. Note that this does not
address the concerns raised in [TALKING].
The client sends requests immediately after sending the session The server connection header consists of just a SETTINGS frame
header, without waiting to receive a server session header. This (Section 3.8.4) that MUST be the first frame the server sends in the
ensures that confirming session headers does not add latency. HTTP/2.0 connection.
Both client and server MUST close the connection if it does not begin To avoid unnecessary latency, clients are permitted to send
with a valid session header. A GOAWAY frame (Section 3.7.7) MAY be additional frames to the server immediately after sending the client
omitted if it is clear that the peer is not using HTTP/2.0. connection header, without waiting to receive the server connection
header. It is important to note, however, that the server connection
header SETTINGS frame might include parameters that necessarily alter
how a client is expected to communicate with the server. Upon
receiving the SETTINGS frame, the client is expected to honor any
parameters established.
Clients and servers MUST terminate the TCP connection if either peer
does not begin with a valid connection header. A GOAWAY frame
(Section 3.8.7) MAY be omitted if it is clear that the peer is not
using HTTP/2.0.
3.3. Framing 3.3. Framing
Once the connection is established, clients and servers exchange Once the HTTP/2.0 connection is established, clients and servers can
HTTP/2.0 frames. Frames are the basic unit of communication. begin exchanging frames.
3.3.1. Frame Header 3.3.1. Frame Header
HTTP/2.0 frames share a common header format. Frames have an 8 byte HTTP/2.0 frames share a common base format consisting of an 8-byte
header with between 0 and 65535 bytes of data. header followed by 0 to 65535 bytes of data.
0 1 2 3 0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Length (16) | Type (8) | Flags (8) | | Length (16) | Type (8) | Flags (8) |
+-+-------------+---------------+-------------------------------+ +-+-------------+---------------+-------------------------------+
|R| Stream Identifier (31) | |R| Stream Identifier (31) |
+-+-------------------------------------------------------------+ +-+-------------------------------------------------------------+
| Frame Data (0...) ... | Frame Data (0...) ...
+---------------------------------------------------------------+ +---------------------------------------------------------------+
Frame Header Frame Header
The fields of the frame header are defined as: The fields of the frame header are defined as:
Length: The 16-bit length of the frame payload in bytes. The length Length: The length of the frame data expressed as an unsigned 16-bit
of the frame header is not included in this sum. integer. The 8 bytes of the frame header are not included in this
value.
Type: The 8-bit type of the frame. The frame type determines how Type: The 8-bit type of the frame. The frame type determines how
the remainder of the frame header and payload are interpreted. the remainder of the frame header and data are interpreted.
Implementations MUST ignore frames that use types that they do not Implementations MUST ignore unsupported and unrecognized frame
support. types.
Flags: An 8-bit field reserved for flags. Bits that have undefined Flags: An 8-bit field reserved for frame-type specific boolean
semantics are reserved. The following flags are defined for all flags.
frame types:
FINAL (0x1): Bit 1 (the least significant bit) indicates that The least significant bit (0x1) - the FINAL bit - is defined for
this is the last frame in a stream. This places the stream all frame types as an indication that this frame is the last the
into a half-closed state (Section 3.4.5). No further frames endpoint will send for the identified stream. Setting this flag
follow in the direction of the carrying frame. causes the stream to enter the half-closed state (Section 3.4.3).
Implementations MUST process the FINAL bit for all frames whose
stream identifier field is not 0x0. The FINAL bit MUST NOT be set
on frames that use a stream identifier of 0.
Frame types can define semantics for frame-specific flags. The remaining flags can be assigned semantics specific to the
indicated frame type. Flags that have no defined semantics for a
particular frame type MUST be ignored, and MUST be left unset (0)
when sending.
R: A reserved 1-bit field. The semantics of this bit are not R: A reserved 1-bit field. The semantics of this bit are undefined
defined. and the bit MUST remain unset (0) when sending and MUST be ignored
when receiving.
Stream Identifier: A 31-bit stream identifier (see Section 3.4.1). Stream Identifier: A 31-bit stream identifier (see Section 3.4.1).
A value 0 is reserved for frames that are directed at the session A value 0 is reserved for frames that are associated with the
as a whole instead of a single stream. connection as a whole as opposed to an individual stream.
Frame Data: Frames contain between 0 and 65535 bytes of data.
Reserved bits in the frame header MUST be set to zero when sending The structure and content of the remaining frame data is dependent
and MUST be ignored when receiving frames, unless the semantics of entirely on the frame type.
the bit are known.
3.3.2. Frame Processing 3.3.2. Frame Size
A frame of the maximum size might be too large for implementations Implementations with limited resources might not be capable of
with limited resources to process. Implementations MAY choose to processing large frame sizes. Such implementations MAY choose to
support frames smaller than the maximum possible size. However, place additional limits on the maximum frame size. However, all
implementations MUST be able to receive frames containing at least implementations MUST be capable of receiving and processing frames
8192 octets of payload. containing at least 8192 octets of data. [[anchor6: Ed. Question:
Does this minimum include the 8-byte header or just the frame data?]]
An implementation MUST immediately close a stream if it is unable to An implementation MUST terminate a stream immediately if it is unable
process a frame related to that stream due to it exceeding a size to process a frame due it's size. This is done by sending an
limit. The implementation MUST send a RST_STREAM frame RST_STREAM frame (Section 3.8.3) containing the FRAME_TOO_LARGE error
(Section 3.7.3) containing FRAME_TOO_LARGE error code if the frame code.
size limit is exceeded.
[[anchor9: <https://github.com/http2/http2-spec/issues/28>: Need a [[anchor7: <https://github.com/http2/http2-spec/issues/28>: Need a
way to signal the maximum frame size; no way to RST_STREAM on non- way to signal the maximum frame size; no way to RST_STREAM on non-
stream-related frames.]] stream-related frames.]]
3.4. Streams 3.4. Streams
Streams are independent sequences of bi-directional data divided into A "stream" is an independent, bi-directional sequence of frames
frames with several properties: exchanged between the client and server within an HTTP/2.0
connection. Streams have several important characteristics:
o Streams can be created by either the client or server. o Streams can be established and used unilaterally or shared by
either the client or server.
o Streams optionally carry a set of name-value header pairs. o Streams can be rejected or cancelled by either endpoint.
o Streams can concurrently send data interleaved with other streams. o Multiple types of frames can be sent by either endpoint within a
single stream.
o Streams can be established and used unilaterally. o The order in which frames are sent within a stream is significant.
Recipients are required to process frames in the order they are
received.
o Streams can be cancelled. o Streams optionally carry a set of name-value header pairs that are
expressed within the headers block of HEADERS+PRIORITY, HEADERS,
or PUSH_PROMISE frames.
3.4.1. Stream Creation o A single HTTP/2.0 connection can contain multiple concurrently
active streams, with either endpoint interleaving frames from
multiple streams.
Use of streams does not require negotiation. A stream is not 3.4.1. Stream Creation
created, streams are used by sending a frame on the stream.
Streams are identified by a 31-bit numeric identifier. Streams There is no coordination or shared action between the client and
initiated by a client use odd numbered stream identifiers. Streams server required to create a stream. Rather, new streams are
initiated by the server use odd numbered stream identifiers. A established by sending a frame whose stream identifier field
stream identifier of zero MUST NOT be used to create a new stream. references a previously unused stream identifier.
The stream identifier of a new stream MUST be greater than all other All streams are identified by an unsigned 31-bit integer. Streams
streams from that endpoint, unless the stream identifier was initiated by a client use odd numbered stream identifiers; those
previously reserved (such as the promised stream identifier in a initiated by the server use even numbered stream identifiers. A
PUSH_PROMISE (Section 3.7.5) frame). An endpoint that receives an stream identifier of zero MUST NOT be used to establish a new stream.
unexpected stream identifier MUST treat this as a session error
(Section 3.5.1) of type PROTOCOL_ERROR.
A long-lived session can result in available stream identifiers being The identifier of a newly established stream MUST be numerically
exhausted. An endpoint that is unable to create a new stream greater than all previously established streams from that endpoint
identifier can establish a new session for any new streams. within the HTTP/2.0 connection, unless the identifier has been
reserved using a PUSH_PROMISE (Section 3.8.5) frame. An endpoint
that receives an unexpected stream identifier MUST respond with a
connection error (Section 3.5.1) of type PROTOCOL_ERROR.
An endpoint cannot prevent the creation of a new stream, but it can A peer can limit the total number of concurrently active streams
request the early termination of an unwanted stream. Upon receipt of using the SETTINGS_MAX_CONCURRENT_STREAMS parameters within a
a frame, the recipient can terminate the corresponding stream by SETTINGS frame. The maximum concurrent streams setting is specific
sending a stream error (Section 3.5.2) of type REFUSED_STREAM. This to each endpoint and applies only to the peer. That is, clients
cannot prevent the initiating endpoint from sending frames for that specify the maximum number of concurrent streams the server can
stream prior to receiving this request. initiate, and servers specify the maximum number of concurrent
streams the client can initiate. Peer endpoints MUST NOT exceed this
limit. All concurrently active streams initiated by an endpoint,
including streams that are half-open (Section 3.4.3) in any
direction, count toward that endpoint's limit.
3.4.2. Stream priority Stream identifiers cannot be reused within a connection. Long-lived
connections can cause an endpoint to exhaust the available range of
stream identifiers. A client that is unable to establish a new
stream identifier can establish a new connection for new streams.
The creator of a stream assigns a priority for that stream. Priority Either endpoint can request the early termination of an unwanted
is represented as a 31 bit integer. 0 represents the highest priority stream by sending an RST_STREAM frame (Section 3.5.2) with an error
and 2^31-1 represents the lowest priority. code of either REFUSED_STREAM (if no frames have been processed) or
CANCEL (if at least one frame has been processed). Such termination
might not take effect immediately as the peer might have sent
additional frames on the stream prior to receiving the termination
request.
The sender and recipient SHOULD use best-effort to process streams in 3.4.2. Stream priority
the order of highest priority to lowest priority. [[anchor11: ED:
toothless, useless "SHOULD": reword]]
3.4.3. Stream headers The endpoint establishing a new stream can assign a priority for the
stream. Priority is represented as an unsigned 31-bit integer. 0
represents the highest priority and 2^31-1 represents the lowest
priority.
Streams carry optional sets of header fields which carry metadata The purpose of this value is to allow the initiating endpoint to
about the stream. After the stream has been created, and as long as request that frames for the stream be processed with higher priority
the sender is not closed (Section 3.4.6) or half-closed relative to any other concurrently active streams. That is, if an
(Section 3.4.5), each side may send HEADERS frame(s) containing the endpoint receives interleaved frames for multiple streams, the
header data. Header data can be sent in multiple HEADERS frames, and endpoint ought to make a best-effort attempt at processing frames for
HEADERS frames may be interleaved with data frames. higher priority streams before processing those for lower priority
streams.
3.4.4. Stream data exchange Explicitly setting the priority for a stream does not guarantee any
particular processing order for the stream relative to any other
stream. Nor is there is any mechanism provided by which the
initiator of a stream can force or require a receiving endpoint to
process frames from one stream before processing frames from another.
Once a stream is created, it can be used to send arbitrary amounts of 3.4.3. Stream half-close
data. Generally this means that a series of data frames will be sent
on the stream until a frame containing the FINAL flag (Section 3.3.1)
is set. Once the FINAL flag has been set on any frame, the stream is
considered to be half-closed.
3.4.5. Stream half-close When an endpoint sends a frame for a stream with the FINAL flag set,
the stream is considered to be half-closed for that endpoint.
Subsequent frames MUST NOT be sent by that endpoint for the half
closed stream for the remaining duration of the HTTP/2.0 connection.
When both endpoints have sent frames with the FINAL flag set, the
stream is considered to be fully closed.
When one side of the stream sends a frame with the FINAL flag set, If an endpoint receives additional frames for a stream that was
the stream is half-closed from that endpoint. The sender of the previously half-closed by the sending peer, the recipient MUST
FINAL flag MUST NOT send further frames on that stream. When both respond with a stream error (Section 3.5.2) of type STREAM_CLOSED.
sides have half-closed, the stream is closed.
An endpoint MUST treat the receipt of a data frame on a half-closed An endpoint that has not yet half-closed a stream by sending the
stream as a stream error (Section 3.5.2) of type STREAM_CLOSED. FINAL flag can continue sending frames on the stream.
Streams that have never received packets can be considered to be It is not necessary for an endpoint to half-close a stream for which
half-closed in the direction that is silent. This allows either peer it has not sent any frames. This allows endpoints to use fully
to create a unidirectional stream, which does not require an explicit unidirectional streams that do not require explicit action or
close from the peer that does not transmit frames. acknowledgement from the receiver.
3.4.6. Stream close 3.4.4. Stream close
Streams can be terminated in the following ways: Streams can be terminated in the following ways:
Normal termination: Normal stream termination occurs when both Normal termination: Normal stream termination occurs when both
sender and recipient have half-closed the stream by sending a client and server have half-closed the stream by sending a frame
frame containing a FINAL flag (Section 3.3.1). containing a FINAL flag (Section 3.3.1).
Half-close on unidirectional stream: A stream that only has frames Half-close on unidirectional stream: A stream that only has frames
sent in one direction can be tentatively considered to be closed sent in one direction can be tentatively considered to be closed
once a frame containing a FINAL flag is sent. The active sender once a frame containing a FINAL flag is sent. The active sender
on the stream MUST be prepared to receive frames after closing the on the stream MUST be prepared to receive frames after closing the
stream. stream.
Abrupt termination: Either the peer can send a RST_STREAM control Abrupt termination: Either peer can send a RST_STREAM control frame
frame at any time to terminate an active stream. RST_STREAM at any time to terminate an active stream. RST_STREAM contains an
contains an error code to indicate the reason for termination. A error code to indicate the reason for termination. A RST_STREAM
RST_STREAM indicates that the sender will transmit no further data indicates that the sender will transmit no further data on the
on the stream and that the receiver is requested to cease stream and that the receiver is advised to cease transmission on
transmission. it.
The sender of a RST_STREAM frame MUST allow for frames that have The sender of a RST_STREAM frame MUST allow for frames that have
already been sent by the peer prior to the RST_STREAM being already been sent by the peer prior to the RST_STREAM being
processed. If in-transit frames alter session state, these frames processed. If in-transit frames alter connection state, these
cannot be safely discarded. See Stream Error Handling frames cannot be safely discarded. See Stream Error Handling
(Section 3.5.2) for more details. (Section 3.5.2) for more details.
TCP connection teardown: If the TCP connection is torn down while TCP connection teardown: If the TCP connection is torn down while
un-closed streams exist, then the endpoint must assume that the un-closed streams exist, then the endpoint MUST assume that the
stream was abnormally interrupted and may be incomplete. stream was abnormally interrupted and may be incomplete.
If an endpoint receives a data frame after the stream is closed, it
MAY send a RST_STREAM to the sender with the status PROTOCOL_ERROR.
3.5. Error Handling 3.5. Error Handling
HTTP/2.0 framing permits two classes of error: HTTP/2.0 framing permits two classes of error:
o An error condition that renders the entire session unusable is a o An error condition that renders the entire connection unusable is
session error. a connection error.
o An error in an individual stream is a stream error. o An error in an individual stream is a stream error.
3.5.1. Session Error Handling 3.5.1. Connection Error Handling
A session error is any error which prevents further processing of the A connection error is any error which prevents further processing of
framing layer or which corrupts any session state. the framing layer or which corrupts any connection state.
An endpoint that encounters a session error MUST first send a GOAWAY An endpoint that encounters a connection error MUST first send a
(Section 3.7.7) frame with the stream identifier of the last stream GOAWAY (Section 3.8.7) frame with the stream identifier of the last
that it successfully received from its peer. The GOAWAY frame stream that it successfully received from its peer. The GOAWAY frame
includes an error code that indicates why the session is terminating. includes an error code that indicates why the connection is
After sending the GOAWAY frame, the endpoint MUST close the TCP terminating. After sending the GOAWAY frame, the endpoint MUST close
connection. the TCP connection.
It is possible that the GOAWAY will not be reliably received by the It is possible that the GOAWAY will not be reliably received by the
receiving endpoint. In the event of a session error, GOAWAY only receiving endpoint. In the event of a connection error, GOAWAY only
provides a best-effort attempt to communicate with the peer about why provides a best-effort attempt to communicate with the peer about why
the session is going down. the connection is being terminated.
An endpoint can end a session at any time. In particular, an An endpoint can end a connection at any time. In particular, an
endpoint MAY choose to treat a stream error as a session error if the endpoint MAY choose to treat a stream error as a connection error if
error is recurrent. Endpoints SHOULD send a GOAWAY frame when ending the error is recurrent. Endpoints SHOULD send a GOAWAY frame when
a session, as long as circumstances permit it. ending a connection, as long as circumstances permit it.
3.5.2. Stream Error Handling 3.5.2. Stream Error Handling
A stream error is an error related to a specific stream identifier A stream error is an error related to a specific stream identifier
that does not affect processing of other streams at the framing that does not affect processing of other streams at the framing
layer. layer.
An endpoint that detects a stream error sends a RST_STREAM An endpoint that detects a stream error sends a RST_STREAM
(Section 3.7.3) frame that contains the stream identifier of the (Section 3.8.3) frame that contains the stream identifier of the
stream where the error occurred. The RST_STREAM frame includes an stream where the error occurred. The RST_STREAM frame includes an
error code that indicates the type of error. error code that indicates the type of error.
A RST_STREAM is the last frame that an endpoint can send on a stream. A RST_STREAM is the last frame that an endpoint can send on a stream.
The peer that sends the RST_STREAM frame MUST be prepared to receive The peer that sends the RST_STREAM frame MUST be prepared to receive
any frames that were sent or enqueued for sending by the remote peer. any frames that were sent or enqueued for sending by the remote peer.
These frames can be ignored, except where they modify session state These frames can be ignored, except where they modify connection
(such as the header compression state). state (such as the state maintained for header compression
(Section 3.7)).
An endpoint SHOULD NOT send more than one RST_STREAM frame for any Normally, an endpoint SHOULD NOT send more than one RST_STREAM frame
stream. An endpoint MAY send additional RST_STREAM frames if it for any stream. However, an endpoint MAY send additional RST_STREAM
receives frames on a closed stream after more than a round trip time. frames if it receives frames on a closed stream after more than a
This behaviour is permitted to deal with misbehaving implementations round trip time. This behavior is permitted to deal with misbehaving
where treating this as a session error is inappropriate. implementations.
An endpoint MUST NOT send a RST_STREAM in response to an RST_STREAM An endpoint MUST NOT send a RST_STREAM in response to an RST_STREAM
frame. This could trigger infinite loops of RST_STREAM frames. frame, to avoid looping.
3.5.3. Error Codes 3.5.3. Error Codes
Error codes are 32-bit fields that are used in RST_STREAM and GOAWAY Error codes are 32-bit fields that are used in RST_STREAM and GOAWAY
frames to convey the reasons for the stream or session error. frames to convey the reasons for the stream or connection error.
Error codes share a common code space. Some error codes only apply Error codes share a common code space. Some error codes only apply
to specific conditions and have no defined semantics in certain frame to specific conditions and have no defined semantics in certain frame
types. types.
The following error codes are defined: The following error codes are defined:
NO_ERROR (0): The associated condition is not as a result of an NO_ERROR (0): The associated condition is not as a result of an
error. For example, a GOAWAY might include this code to indicate error. For example, a GOAWAY might include this code to indicate
graceful shutdown of a session. graceful shutdown of a connection.
PROTOCOL_ERROR (1): An unspecific protocol error was detected. This PROTOCOL_ERROR (1): The endpoint detected an unspecific protocol
error is for use when a more specific error code is not available. error. This error is for use when a more specific error code is
not available.
INTERNAL_ERROR (2): The implementation encountered an unexpected INTERNAL_ERROR (2): The endpoint encountered an unexpected internal
internal error. error.
FLOW_CONTROL_ERROR (3): The endpoint detected that its peer violated FLOW_CONTROL_ERROR (3): The endpoint detected that its peer violated
the flow control protocol. the flow control protocol.
INVALID_STREAM (4): A frame was received for an inactive stream. INVALID_STREAM (4): The endpoint received a frame for an inactive
stream.
STREAM_CLOSED (5): The endpoint received a frame after a stream was STREAM_CLOSED (5): The endpoint received a frame after a stream was
half-closed. half-closed.
FRAME_TOO_LARGE (6): The endpoint received a frame that was larger FRAME_TOO_LARGE (6): The endpoint received a frame that was larger
than the maximum size that it supports. than the maximum size that it supports.
REFUSED_STREAM (7): Indicates that the stream was refused before any REFUSED_STREAM (7): The endpoint is refusing the stream before
processing has been done on the stream. processing its payload.
CANCEL (8): Used by the creator of a stream to indicate that the CANCEL (8): Used by the creator of a stream to indicate that the
stream is no longer needed. stream is no longer needed.
COMPRESSION_ERROR (9): The endpoint is unable to maintain the
compression context for the connection.
3.6. Stream Flow Control 3.6. Stream Flow Control
Multiplexing streams introduces contention for access to the shared Using streams for multiplexing introduces contention over use of the
TCP connection. Stream contention can result in streams being TCP connection, resulting in blocked streams. A flow control scheme
blocked by other streams. A flow control scheme ensures that streams ensures that streams on the same connection do not destructively
do not destructively interfere with other streams on the same TCP interfere with each other.
connection.
HTTP/2.0 provides for flow control through use of the WINDOW_UPDATE
(Section 3.8.9) frame type.
3.6.1. Flow Control Principles 3.6.1. Flow Control Principles
Experience with TCP congestion control has shown that algorithms can Experience with TCP congestion control has shown that algorithms can
evolve over time to become more sophisticated without requiring evolve over time to become more sophisticated without requiring
protocol changes. TCP congestion control and its evolution is protocol changes. TCP congestion control and its evolution is
clearly different from HTTP/2.0 flow control, though the evolution of clearly different from HTTP/2.0 flow control, though the evolution of
TCP congestion control algorithms shows that a similar approach could TCP congestion control algorithms shows that a similar approach could
be feasible for HTTP/2.0 flow control. be feasible for HTTP/2.0 flow control.
HTTP/2.0 stream flow control aims to allow for future improvements to HTTP/2.0 stream flow control aims to allow for future improvements to
flow control algorithms without requiring protocol changes. Flow flow control algorithms without requiring protocol changes. Flow
control in HTTP/2.0 has the following characteristics: control in HTTP/2.0 has the following characteristics:
1. Flow control is hop-by-hop, not end-to-end. 1. Flow control is hop-by-hop, not end-to-end.
2. Flow control is based on window update messages. Receivers 2. Flow control is based on window update frames. Receivers
advertise how many octets they are prepared to receive on a advertise how many octets they are prepared to receive on a
stream. This is a credit-based scheme. stream. This is a credit-based scheme.
3. Flow control is directional with overall control provided by the 3. Flow control is directional with overall control provided by the
receiver. A receiver MAY choose to set any window size that it receiver. A receiver MAY choose to set any window size that it
desires for each stream and for the entire connection. A sender desires for each stream and for the entire connection. A sender
MUST respect flow control limits imposed by a receiver. Clients, MUST respect flow control limits imposed by a receiver. Clients,
servers and intermediaries all independently advertise their flow servers and intermediaries all independently advertise their flow
control preferences as a receiver and abide by the flow control control preferences as a receiver and abide by the flow control
limits set by their peer when sending. limits set by their peer when sending.
skipping to change at page 17, line 27 skipping to change at page 18, line 35
frame. Of the frames specified in this document, only data frame. Of the frames specified in this document, only data
frames are subject to flow control; all other frame types do not frames are subject to flow control; all other frame types do not
consume space in the advertised flow control window. This consume space in the advertised flow control window. This
ensures that important control frames are not blocked by flow ensures that important control frames are not blocked by flow
control. control.
6. Flow control can be disabled by a receiver. A receiver can 6. Flow control can be disabled by a receiver. A receiver can
choose to either disable flow control for a stream or connection choose to either disable flow control for a stream or connection
by declaring an infinite flow control limit. by declaring an infinite flow control limit.
7. HTTP/2.0 standardizes only the format of the window update 7. HTTP/2.0 standardizes only the format of the window update frame
message (Section 3.7.9). This does not stipulate how a receiver (Section 3.8.9). This does not stipulate how a receiver decides
decides when to send this message or the value that it sends. when to send this frame or the value that it sends. Nor does it
Nor does it specify how a sender chooses to send packets. specify how a sender chooses to send packets. Implementations
Implementations are able to select any algorithm that suits their are able to select any algorithm that suits their needs.
needs.
Implementations are also responsible for managing how requests and Implementations are also responsible for managing how requests and
responses are sent based on priority; choosing how to avoid head of responses are sent based on priority; choosing how to avoid head of
line blocking for requests; and managing the creation of new streams. line blocking for requests; and managing the creation of new streams.
Algorithm choices for these could interact with any flow control Algorithm choices for these could interact with any flow control
algorithm. algorithm.
3.6.2. Appropriate Use of Flow Control 3.6.2. Appropriate Use of Flow Control
Flow control is defined to protect deployments (client, server or Flow control is defined to protect endpoints (client, server or
intermediary) that are operating under constraints. For example, a intermediary) that are operating under resource constraints. For
proxy must share memory between many connections. Flow control example, a proxy needs to share memory between many connections, and
addresses cases where the receiver is unable process data on one also might have a slow upstream connection and a fast downstream one.
stream, yet wants to be continue to process other streams. Flow control addresses cases where the receiver is unable process
data on one stream, yet wants to continue to process other streams in
the same connection.
Deployments that do not rely on this capability SHOULD disable flow Deployments that do not require this capability SHOULD disable flow
control for data that is being received. Note that flow control control for data that is being received. Note that flow control
cannot be disabled for sending. Sending data is always subject to cannot be disabled for sending. Sending data is always subject to
the flow control window advertised by the receiver. the flow control window advertised by the receiver.
Deployments with constrained resources (for example, memory), MAY Deployments with constrained resources (for example, memory) MAY
employ flow control to limit the amount of memory a peer can consume. employ flow control to limit the amount of memory a peer can consume.
This can lead to suboptimal use of available network resources if Note, however, that this can lead to suboptimal use of available
flow control is enabled without knowledge of the bandwidth-delay network resources if flow control is enabled without knowledge of the
product (see [RFC1323]). bandwidth-delay product (see [RFC1323]).
Implementation of flow control in full awareness of the current Even with full awareness of the current bandwidth-delay product,
bandwidth-delay product is difficult, but it can ensure that implementation of flow control is difficult. However, it can ensure
constrained resources are protected without any reduction in that constrained resources are protected without any reduction in
connection utilization. connection utilization.
3.7. Frame Types 3.7. Header Blocks
3.7.1. DATA Frames The header block is found in the HEADERS, HEADERS+PRIORITY and
PUSH_PROMISE frames. The header block consists of a set of header
fields, which are name-value pairs. Headers are compressed using
black magic.
DATA frames (type=0) are used to convey HTTP message bodies. The Compression of header fields is a work in progress, as is the format
payload of a data frame contains either a request or response body. of this block.
No frame-specific flags are defined for DATA frames. The contents of header blocks MUST be processed by the compression
context, even if stream has been reset or the frame is discarded. If
header blocks cannot be processed, the receiver MUST treat the
connection with a connection error (Section 3.5.1) of type
COMPRESSION_ERROR.
3.7.2. HEADERS+PRIORITY 3.8. Frame Types
The HEADERS+PRIORITY frame (type=1) allows the sender to set header This specification defines a number of frame types, each identified
fields and stream priority at the same time. This MUST be used for by a unique 8-bit type code. Each frame type serves a distinct
each stream that is created. purpose either in the establishment and management of the connection
as a whole, or of individual streams.
The transmission of specific frame types can alter the state of a
connection. If endpoints fail to maintain a synchronized view of the
connection state, successful communication within the connection will
no longer be possible. Therefore, it is important that endpoints
have a shared comprehension of how the state is affected by the use
any given frame. Accordingly, while it is expected that new frame
types will be introduced by extensions to this protocol, only frames
defined by this document are permitted to alter the connection state.
3.8.1. DATA Frames
DATA frames (type=0x0) convey arbitrary, variable-length sequences of
octets associated with a stream. One or more DATA frames are used,
for instance, to carry HTTP request or response payloads.
The DATA frame does not define any type-specific flags.
DATA frames MUST be associated with a stream. If a DATA frame is
received whose stream identifier field is 0x0, the recipient MUST
respond with a connection error (Section 3.5.1) of type
PROTOCOL_ERROR.
3.8.2. HEADERS+PRIORITY
The HEADERS+PRIORITY frame (type=0x1) allows the sender to set header
fields and stream priority at the same time.
0 1 2 3 0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|X| Priority (31) | |X| Priority (31) |
+-+-------------------------------------------------------------+ +-+-------------------------------------------------------------+
| Header Block (*) ... | Header Block (*) ...
+---------------------------------------------------------------+ +---------------------------------------------------------------+
HEADERS+PRIORITY Frame Payload HEADERS+PRIORITY Frame Payload
The HEADERS+PRIORITY frame is identical to the HEADERS frame The HEADERS+PRIORITY frame is identical to the HEADERS frame
(Section 3.7.8), with a 32-bit field containing priority included (Section 3.8.8), preceded by a single reserved bit and a 31-bit
before the header block. priority; see Section 3.4.2.
The most significant bit of the priority is reserved. The 31-bit HEADERS+PRIORITY uses the same flags as the HEADERS frame, except
priority indicates the priority for the stream, as assigned by the that a HEADERS+PRIORITY frame with a CONTINUES bit MUST be followed
sender, see Section 3.4.2. by another HEADERS+PRIORITY frame. See HEADERS frame (Section 3.8.8)
for any flags.
3.7.3. RST_STREAM HEADERS+PRIORITY frames MUST be associated with a stream. If a
HEADERS+PRIORITY frame is received whose stream identifier field is
0x0, the recipient MUST respond with a connection error
(Section 3.5.1) of type PROTOCOL_ERROR.
The RST_STREAM frame (type=3) allows for abnormal termination of a The HEADERS+PRIORITY frame modifies the connection state as defined
stream. When sent by the creator of a stream, it indicates the in Section 3.7.
creator wishes to cancel the stream. When sent by the recipient of a
stream, it indicates an error or that the recipient did not want to 3.8.3. RST_STREAM
accept the stream, so the stream should be closed.
The RST_STREAM frame (type=0x3) allows for abnormal termination of a
stream. When sent by the initiator of a stream, it indicates that
they wish to cancel the stream. When sent by the receiver of a
stream, it indicates that either the receiver is rejecting the
stream, requesting that the stream be cancelled or that an error
condition has occurred.
0 1 2 3 0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| Error Code (32) | | Error Code (32) |
+---------------------------------------------------------------+ +---------------------------------------------------------------+
RST_STREAM Frame Payload RST_STREAM Frame Payload
The RST_STREAM frame does not define any valid flags. The RST_STREAM frame contains a single unsigned, 32-bit integer
identifying the error code (Section 3.5.3). The error code indicates
why the stream is being terminated.
The RST_STREAM frame contains a single 32-bit error code No type-flags are defined.
(Section 3.5.3). The error code indicates why the stream is being
terminated.
After receiving a RST_STREAM on a stream, the recipient must not send The RST_STREAM frame fully terminates the referenced stream and
additional frames for that stream, and the stream moves into the causes it to enter the closed state. After receiving a RST_STREAM on
closed state. a stream, the receiver MUST NOT send additional frames for that
stream. However, after sending the RST_STREAM, the sending endpoint
MUST be prepared to receive and process additional frames sent on the
stream that might have been sent by the peer prior to the arrival of
the RST_STREAM.
3.7.4. SETTINGS RST_STREAM frames MUST be associated with a stream. If a RST_STREAM
frame is received whose stream identifier field is 0x0 the recipient
MUST respond with a connection error (Section 3.5.1) of type
PROTOCOL_ERROR.
A SETTINGS frame (type=4) contains a set of id/value pairs for 3.8.4. SETTINGS
communicating configuration data about how the two endpoints may
communicate. SETTINGS frames MUST be sent at the start of a session,
but they can be sent at any other time by either endpoint. Settings
are declarative, not negotiated, each peer indicates their own
configuration.
[[anchor17: Note that persistence of settings is under discussion in The SETTINGS frame (type=0x4) conveys configuration parameters that
the WG and might be removed in a future version of this document.]] affect how endpoints communicate. The parameters are either
constraints on peer behavior or preferences.
When the server is the sender, the sender can request that SETTINGS frames MUST be sent at the start of a connection, and MAY be
configuration data be persisted by the client across HTTP/2.0 sent at any other time by either endpoint over the lifetime of the
sessions and returned to the server in future communications. connection.
Clients persist settings on a per origin basis (see [RFC6454] for a Implementations MUST support all of the settings defined by this
definition of web origins). That is, when a client connects to a specification and MAY support additional settings defined by
server, and the server persists settings within the client, the extensions. Unsupported or unrecognized settings MUST be ignored.
client SHOULD return the persisted settings on future connections to New settings MUST NOT be defined or implemented in a way that
the same origin AND IP address and TCP port. Clients MUST NOT requires endpoints to understand then in order to communicate
request servers to use the persistence features of the SETTINGS successfully.
frames, and servers MUST ignore persistence related flags sent by a
client.
Valid frame-specific flags for the SETTINGS frame are: A SETTINGS frame is not required to include every defined setting;
senders can include only those parameters for which it has accurate
values and a need to convey. When multiple parameters are sent, they
SHOULD be sent in order of numerically lowest ID to highest ID. A
single SETTINGS frame MUST NOT contain multiple values for the same
ID. If the receiver of a SETTINGS frame discovers multiple values
for the same ID, it MUST ignore all values for that ID except the
first one.
Over the lifetime of a connection, an endpoint MAY send multiple
SETTINGS frames containing previously unspecified parameters or new
values for parameters whose values have already been established.
Only the most recent value provided setting value applies.
The SETTINGS frame defines the following flag:
CLEAR_PERSISTED (0x2): Bit 2 being set indicates a request to clear CLEAR_PERSISTED (0x2): Bit 2 being set indicates a request to clear
any previously persisted settings before processing the settings. any previously persisted settings before processing the settings.
Clients MUST NOT set this flag. Clients MUST NOT set this flag.
SETTINGS frames always apply to a session, never a single stream. SETTINGS frames always apply to a connection, never a single stream.
The stream identifier for a settings frame MUST be zero. The stream identifier for a settings frame MUST be zero. If an
endpoint receives a SETTINGS frame whose stream identifier field is
anything other than 0x0, the endpoint MUST respond with a connection
error (Section 3.5.1) of type PROTOCOL_ERROR.
3.8.4.1. Setting Format
The payload of a SETTINGS frame consists of zero or more settings.
Each setting consists of an 8-bit flags field specifying per-item
instructions, an unsigned 24-bit setting identifier, and an unsigned
32-bit value.
0 1 2 3 0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|SettingFlags(8)| Setting Identifier (24) | |SettingFlags(8)| Setting Identifier (24) |
+---------------+-----------------------------------------------+ +---------------+-----------------------------------------------+
| Value (32) | | Value (32) |
+---------------------------------------------------------------+ +---------------------------------------------------------------+
Setting Format
SETTINGS ID/Value Pair Two flags are defined for the 8-bit flags field:
The payload of a SETTINGS frame contains zero or more settings. Each PERSIST_VALUE (0x1): Bit 1 (the least significant bit) being set
setting is comprised of the following indicates a request from the server to the client to persist this
setting. A client MUST NOT set this flag.
Settings Flags: An 8-bit flags field containing per-setting PERSISTED (0x2): Bit 2 being set indicates that this setting is a
instructions. The following flags are valid: persisted setting being returned by the client to the server.
This also indicates that this setting is not a client setting, but
a value previously set by the server. A server MUST NOT set this
flag.
PERSIST_VALUE (0x1): Bit 1 (the least significant bit) being set 3.8.4.2. Setting Persistence
indicates a request from the server to the client to persist
this setting. A client MUST NOT set this flag.
PERSISTED (0x2): Bit 2 being set indicates that this setting is a [[anchor12: Note that persistence of settings is under discussion in
persisted setting being returned by the client to the server. the WG and might be removed in a future version of this document.]]
This also indicates that this setting is not a client setting,
but a value previously set by the server. A server MUST NOT
set this flag.
All other settings flags are reserved. A server endpoint can request that configuration parameters sent to a
client in a SETTINGS frame are to be persisted by the client across
HTTP/2.0 connections and returned to the server in any new SETTINGS
frame the client sends to the server in the current connection or any
future connections.
Setting Identifier: A 24-bit field that identifies the setting. Persistence is requested on a per-setting basis by setting the
PERSIST_VALUE flag (0x1).
Value: A 32-bit value for the setting. Client endpoints are not permitted to make such requests. Servers
MUST ignore any attempt by clients to request that a server persist
configuration parameters.
The following settings are defined: Persistence of configuration parameters is done on a per-origin basis
(see [RFC6454]). That is, when a client establishes a connection
with a server, and the server requests that the client maintain
persistent settings, the client SHOULD return the persisted settings
on all future connections to the same origin, IP address and TCP
port.
SETTINGS_UPLOAD_BANDWIDTH (1): allows the sender to send its Whenever the client sends a SETTINGS frame in the current connection,
expected upload bandwidth on this channel. This number is an or establishes a new connection with the same origin, persisted
estimate. The value should be the integral number of kilobytes configuration parameters are sent with the PERSISTED flag (0x2) set
per second that the sender predicts as an expected maximum upload for each persisted parameter.
channel capacity.
SETTINGS_DOWNLOAD_BANDWIDTH (2): allows the sender to send its Persisted settings accumulate until the server requests that all
expected download bandwidth on this channel. This number is an previously persisted settings are to be cleared by setting the
estimate. The value should be the integral number of kilobytes CLEAR_PERSISTED (0x2) flag on the SETTINGS frame.
per second that the sender predicts as an expected maximum
download channel capacity.
SETTINGS_ROUND_TRIP_TIME (3): allows the sender to send its expected For example, if the server sends IDs 1, 2, and 3 with the
round-trip-time on this channel. The round trip time is defined FLAG_SETTINGS_PERSIST_VALUE in a first SETTINGS frame, and then sends
as the minimum amount of time to send a control frame from this IDs 4 and 5 with the FLAG_SETTINGS_PERSIST_VALUE in a subsequent
client to the remote and receive a response. The value is SETTINGS frame, the client will return values for all 5 settings (1,
represented in milliseconds. 2, 3, 4, and 5 in this example) to the server.
SETTINGS_MAX_CONCURRENT_STREAMS (4): allows the sender to inform the 3.8.4.3. Defined Settings
remote endpoint the maximum number of concurrent streams which it
will allow. This limit is directional: it applies to the number
of streams that the sender permits the receiver to create. By
default there is no limit. For implementers it is recommended
that this value be no smaller than 100, so as to not unnecessarily
limit parallelism.
SETTINGS_CURRENT_CWND (5): allows the sender to inform the remote The following settings are defined:
endpoint of the current TCP CWND value.
SETTINGS_DOWNLOAD_RETRANS_RATE (6): allows the sender to inform the SETTINGS_UPLOAD_BANDWIDTH (1): indicates the sender's estimated
remote endpoint the retransmission rate (bytes retransmitted / upload bandwidth for this connection. The value is an the
total bytes transmitted). integral number of kilobytes per second that the sender predicts
as an expected maximum upload channel capacity.
SETTINGS_INITIAL_WINDOW_SIZE (7): allows the sender to inform the SETTINGS_DOWNLOAD_BANDWIDTH (2): indicates the sender's estimated
remote endpoint the initial window size (in bytes) for new download bandwidth for this connection. The value is an integral
streams. number of kilobytes per second that the sender predicts as an
expected maximum download channel capacity.
SETTINGS_FLOW_CONTROL_OPTIONS (10): This setting allows an endpoint SETTINGS_ROUND_TRIP_TIME (3): indicates the sender's estimated
to indicate that streams directed to them will not be subject to round-trip-time for this connection. The round trip time is
flow control. The least significant bit (0x1) is set to indicate defined as the minimum amount of time to send a control frame from
that new streams are not flow controlled. Bit 2 (0x2) is set to this client to the remote and receive a response. The value is
indicate that the session is not flow controlled. All other bits represented in milliseconds.
are reserved.
This setting applies to all streams, including existing streams. SETTINGS_MAX_CONCURRENT_STREAMS (4): indicates the maximum number of
concurrent streams that the sender will allow. This limit is
directional: it applies to the number of streams that the sender
permits the receiver to create. By default there is no limit. It
is recommended that this value be no smaller than 100, so as to
not unnecessarily limit parallelism.
These bits cannot be cleared once set, see Section 3.7.9.4. SETTINGS_CURRENT_CWND (5): indicates the sender's current TCP CWND
value.
The message is intentionally extensible for future information which SETTINGS_DOWNLOAD_RETRANS_RATE (6): indicates the sender's
may improve client-server communications. The sender does not need retransmission rate (bytes retransmitted / total bytes
to send every type of ID/value. It must only send those for which it transmitted).
has accurate values to convey. When multiple ID/value pairs are
sent, they should be sent in order of lowest id to highest id. A
single SETTINGS frame MUST not contain multiple values for the same
ID. If the recipient of a SETTINGS frame discovers multiple values
for the same ID, it MUST ignore all values except the first one.
A server may send multiple SETTINGS frames containing different ID/ SETTINGS_INITIAL_WINDOW_SIZE (7): indicates the sender's initial
Value pairs. When the same ID/Value is sent twice, the most recent stream window size (in bytes) for new streams.
value overrides any previously sent values. If the server sends IDs
1, 2, and 3 with the FLAG_SETTINGS_PERSIST_VALUE in a first SETTINGS
frame, and then sends IDs 4 and 5 with the
FLAG_SETTINGS_PERSIST_VALUE, when the client returns the persisted
state on its next SETTINGS frame, it SHOULD send all 5 settings (1,
2, 3, 4, and 5 in this example) to the server.
3.7.5. PUSH_PROMISE SETTINGS_FLOW_CONTROL_OPTIONS (10): indicates that streams directed
to the sender will not be subject to flow control. The least
significant bit (0x1) is set to indicate that new streams are not
flow controlled. All other bits are reserved.
The PUSH_PROMISE frame (type=5) allows the sender to signal a promise This setting applies to all streams, including existing streams.
to create a stream and serve the referenced resource. Minimal data
allowing the receiver to understand which resource(s) are to be
pushed are to be included.
PUSH_PROMISE frames are sent on an existing stream. They declare the These bits cannot be cleared once set, see Section 3.8.9.4.
intent to use another stream for the pushing of a resource. The
PUSH_PROMISE allows the client an opportunity to reject pushed 3.8.5. PUSH_PROMISE
resources.
The PUSH_PROMISE frame (type=0x5) is used to notify the peer endpoint
in advance of streams the sender intends to initiate. The
PUSH_PROMISE frame includes the unsigned 31-bit identifier of the
stream the endpoint plans to create along with a minimal set of
headers that provide additional context for the stream. Section 4.3
contains a thorough description of the use of PUSH_PROMISE frames.
0 1 2 3 0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|X| Promised-Stream-ID (31) | |X| Promised-Stream-ID (31) |
+-+-------------------------------------------------------------+ +-+-------------------------------------------------------------+
| Header Block (*) ... | Header Block (*) ...
+---------------------------------------------------------------+ +---------------------------------------------------------------+
PUSH_PROMISE Payload Format PUSH_PROMISE Payload Format
There are no frame-specific flags for the PUSH_PROMISE frame. The payload of a PUSH_PROMISE includes a "Promised-Stream-ID". This
unsigned 31-bit integer identifies the stream the endpoint intends to
start sending frames for. The promised stream identifier MUST be a
valid choice for the next stream sent by the sender (see new stream
identifier (Section 3.4.1)).
The body of a PUSH_PROMISE includes a "Promised-Stream-ID". This 31- PUSH_PROMISE frames MUST be associated with an existing stream. If
bit identifier indicates the stream on which the resource will be the stream identifier field specifies the value 0x0, a recipient MUST
pushed. The promised stream identifier MUST be a valid choice for respond with a connection error (Section 3.5.1) of type
the next stream sent by the sender (see new stream identifier PROTOCOL_ERROR.
(Section 3.4.1)).
There is no requirement that the streams referred to by this frame The state of promised streams is bound to the state of the original
are created in the order referenced. The PUSH_PROMISE reserves associated stream on which the PUSH_PROMISE frame were sent. If the
stream identifiers for later use; these reserved identifiers can be originating stream state changes to fully closed, all associated
used as prioritization needs dictate. promised streams fully close as well. [[anchor13: Ed. Note: We need
clarification on this point. How synchronized are the lifecycles of
streams and associated promised streams?]]
The PUSH_PROMISE also includes a header block (Section 3.7.10), which PUSH_PROMISE uses the same flags as the HEADERS frame, except that a
describes the resource that will be pushed. PUSH_PROMISE frame with a CONTINUES bit MUST be followed by another
PUSH_PROMISE frame. See HEADERS frame (Section 3.8.8) for any flags.
3.7.6. PING Promised streams are not required to be used in order promised. The
PUSH_PROMISE only reserves stream identifiers for later use.
The PING frame (type=6) is a mechanism for measuring a minimal round- Recipients of PUSH_PROMISE frames can choose to reject promised
trip time from the sender. PING frames can be sent from the client streams by returning a RST_STREAM referencing the promised stream
or the server. identifier back to the sender of the PUSH_PROMISE.
Recipients of a PING frame send an identical frame to the sender as The PUSH_PROMISE frame modifies the connection state as defined in
soon as possible. PING should take highest priority if there is Section 3.7.
other data waiting to be sent.
The PING frame defines a frame-specific flag: 3.8.6. PING
PONG (0x2): Bit 2 being set indicates that this ping frame is a ping The PING frame (type=0x6) is a mechanism for measuring a minimal
response. An endpoint MUST set this flag in ping responses. An round-trip time from the sender, as well as determining whether an
endpoint MUST NOT respond to ping frames containing this flag. idle connection is still functional. PING frames can be sent from
any endpoint.
The payload of a PING frame contains any value. A PING response MUST PING frames consist of an arbitrary, variable-length sequence of
contain the contents of the PING request. octets. Receivers of a PING send a response PING frame with the PONG
flag set and precisely the same sequence of octets back to the sender
as soon as possible.
3.7.7. GOAWAY Processing of PING frames SHOULD be performed with the highest
priority if there are additional frames waiting to be processed.
The GOAWAY frame (type=7) informs the remote side of the connection The PING frame defines one type-specific flag:
to stop creating streams on this session. It can be sent from the
client or the server. Once sent, the sender will ignore frames sent PONG (0x2): Bit 2 being set indicates that this PING frame is a PING
on new streams for the remainder of the session. Recipients of a response. An endpoint MUST set this flag in PING responses. An
GOAWAY frame MUST NOT open additional streams on the session, endpoint MUST NOT respond to PING frames containing this flag.
although a new session can be established for new streams. The
purpose of this message is to allow an endpoint to gracefully stop PING frames are not associated with any individual stream. If a PING
accepting new streams (perhaps for a reboot or maintenance), while frame is received with a stream identifier field value other than
still finishing processing of previously established streams. 0x0, the recipient MUST respond with a connection error
(Section 3.5.1) of type PROTOCOL_ERROR.
3.8.7. GOAWAY
The GOAWAY frame (type=0x7) informs the remote peer to stop creating
streams on this connection. It can be sent from the client or the
server. Once sent, the sender will ignore frames sent on new streams
for the remainder of the connection. Receivers of a GOAWAY frame
MUST NOT open additional streams on the connection, although a new
connection can be established for new streams. The purpose of this
frame is to allow an endpoint to gracefully stop accepting new
streams (perhaps for a reboot or maintenance), while still finishing
processing of previously established streams.
There is an inherent race condition between an endpoint starting new There is an inherent race condition between an endpoint starting new
streams and the remote sending a GOAWAY message. To deal with this streams and the remote sending a GOAWAY frame. To deal with this
case, the GOAWAY contains the stream identifier of the last stream case, the GOAWAY contains the stream identifier of the last stream
which was processed on the sending endpoint in this session. If the which was processed on the sending endpoint in this connection. If
receiver of the GOAWAY used streams that are newer than the indicated the receiver of the GOAWAY used streams that are newer than the
stream identifier, they were not processed by the sender and the indicated stream identifier, they were not processed by the sender
receiver may treat the streams as though they had never been created and the receiver may treat the streams as though they had never been
at all (hence the receiver may want to re-create the streams later on created at all (hence the receiver may want to re-create the streams
a new session). later on a new connection).
Endpoints should always send a GOAWAY message before closing a Endpoints should always send a GOAWAY frame before closing a
connection so that the remote can know whether a stream has been connection so that the remote can know whether a stream has been
partially processed or not. (For example, if an HTTP client sends a partially processed or not. (For example, if an HTTP client sends a
POST at the same time that a server closes a connection, the client POST at the same time that a server closes a connection, the client
cannot know if the server started to process that POST request if the cannot know if the server started to process that POST request if the
server does not send a GOAWAY frame to indicate where it stopped server does not send a GOAWAY frame to indicate where it stopped
working). working).
After sending a GOAWAY message, the sender can ignore frames for new After sending a GOAWAY frame, the sender can ignore frames for new
streams. streams.
[[anchor18: Issue: session state that is established by those [[anchor14: Issue: connection state that is established by those
"ignored" messages cannot be ignored without the state in the two "ignored" frames cannot be ignored without the state in the two peers
peers becoming unsynchronized.]] becoming unsynchronized.]]
0 1 2 3 0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|X| Last-Stream-ID (31) | |X| Last-Stream-ID (31) |
+-+-------------------------------------------------------------+ +-+-------------------------------------------------------------+
| Error Code (32) | | Error Code (32) |
+---------------------------------------------------------------+ +---------------------------------------------------------------+
GOAWAY Payload Format GOAWAY Payload Format
The GOAWAY frame does not define any valid flags. The GOAWAY frame does not define any type-specific flags.
The GOAWAY frame applies to the session, not a specific stream. The The GOAWAY frame applies to the connection, not a specific stream.
stream identifier MUST be zero. The stream identifier MUST be zero.
The GOAWAY frame contains an identifier of the last stream that the The last stream identifier in the GOAWAY frame contains the highest
sender of the GOAWAY is prepared to act upon, which can include numbered stream identifier for which the sender of the GOAWAY frame
processing and replies. This allows an endpoint to discover what has received frames on and might have taken some action on. All
streams might have had some effect or what might be safe to streams up to and including the identified stream might have been
automatically retry. If no streams were acted upon, the last stream processed in some way. The last stream identifier is set to 0 if no
ID MUST be 0. streams were processed.
The GOAWAY frame contains a 32-bit error code (Section 3.5.3) that Note: In this case, "processed" means that some data from the
contains the reason for closing the session. stream was passed to some higher layer of software that might have
taken some action as a result.
3.7.8. HEADERS On streams with lower or equal numbered identifiers that do not close
completely prior to the connection being closed, re-attempting
requests, transactions, or any protocol activity is not possible
(with the exception of idempotent actions like HTTP GET, PUT, or
DELETE). Any protocol activity that uses higher numbered streams can
be safely retried using a new connection.
The HEADERS frame (type=8) provides header fields for a stream. It Activity on streams numbered lower or equal to the last stream
may be optionally sent on an existing stream at any time. Specific identifier might still complete successfully. The sender of a GOAWAY
application of the headers in this frame is application-dependent. frame gracefully shut down a connection by sending a GOAWAY frame,
maintaining the connection in an open state until all in-progress
streams complete.
No frame-specific flags are defined for the HEADERS frame. The last stream ID MUST be 0 if no streams were acted upon.
The body of a HEADERS frame contains a Headers Block The GOAWAY frame also contains a 32-bit error code (Section 3.5.3)
(Section 3.7.10). that contains the reason for closing the connection.
3.7.9. WINDOW_UPDATE 3.8.8. HEADERS
The WINDOW_UPDATE frame (type=9) is used to implement flow control in The HEADERS frame (type=0x8) provides header fields for a stream.
HTTP/2.0. Any number of HEADERS frames can may be sent on an existing stream at
any time.
Flow control in HTTP/2.0 operates at two levels: on each individual Additional type-specific flags for the HEADERS frame are:
stream and on the entire session.
Flow control in HTTP/2.0 is hop by hop, that is, only between the two CONTINUES (0x2): The CONTINUES bit indicates that this frame does
endpoints of a HTTP/2.0 connection. Intermediaries do not forward not contain the entire payload necessary to provide a complete set
WINDOW_UPDATE messages between dependent sessions. However, of headers.
throttling of data transfer by any recipient can indirectly cause the
propagation of flow control information toward the original sender. The payload for a complete set of headers is provided by a
sequence of HEADERS frames, terminated by a HEADERS frame without
the CONTINUES bit. Once the sequence terminates, the payload of
all HEADERS frames are concatenated and interpreted as a single
block.
A HEADERS frame that includes a CONTINUES bit MUST be followed by
a HEADERS frame for the same stream. A receiver MUST treat the
receipt of any other type of frame or a frame on a different
stream as a connection error (Section 3.5.1) of type
PROTOCOL_ERROR.
The payload of a HEADERS frame contains a Headers Block
(Section 3.7).
The HEADERS frame is associated with an existing stream. If a
HEADERS frame is received with a stream identifier of 0x0, the
recipient MUST respond with a stream error (Section 3.5.2) of type
PROTOCOL_ERROR.
The HEADERS frame changes the connection state as defined in
Section 3.7.
3.8.9. WINDOW_UPDATE
The WINDOW_UPDATE frame (type=0x9) is used to implement flow control.
Flow control operates at two levels: on each individual stream and on
the entire connection.
Both types of flow control are hop by hop; that is, only between the
two endpoints. Intermediaries do not forward WINDOW_UPDATE frames
between dependent connections. However, throttling of data transfer
by any receiver can indirectly cause the propagation of flow control
information toward the original sender.
Flow control only applies to frames that are identified as being Flow control only applies to frames that are identified as being
subject to flow control. Of the frames defined in this document, subject to flow control. Of the frame types defined in this
only data frames are subject to flow control. Receivers MUST either document, this includes only DATA frame. Frames that are exempt from
buffer or process all other frames, terminate the corresponding flow control MUST be accepted and processed, unless the receiver is
stream, or terminate the session. The stream or session is unable to assign resources to handling the frame. A receiver MAY
terminated with a FLOW_CONTROL_ERROR code. respond with a stream error (Section 3.5.2) or connection error
(Section 3.5.1) of type FLOW_CONTROL_ERROR if it is unable accept a
frame.
Valid flags for the WINDOW_UPDATE frame are: The following additional flags are defined for the WINDOW_UPDATE
frame:
END_FLOW_CONTROL (0x2): Bit 2 being set indicates that flow control END_FLOW_CONTROL (0x2): Bit 2 being set indicates that flow control
for the identified stream or session is ended and subsequent for the identified stream or connection has been ended; subsequent
frames do not need to be flow controlled. frames do not need to be flow controlled.
The WINDOW_UPDATE frame can be stream related or session related. The WINDOW_UPDATE frame can be specific to a stream or to the entire
The stream identifier in the WINDOW_UPDATE frame header identifies connection. In the former case, the frame's stream identifier
the affected stream, or includes a value of 0 to indicate that the indicates the affected stream; in the latter, the value "0" indicates
session flow control window is updated. that the entire connection is the subject of the frame.
The payload of a WINDOW_UPDATE frame contains a 32-bit value. This The payload of a WINDOW_UPDATE frame is a 32-bit value indicating the
value is the additional number of bytes that the sender can transmit additional number of bytes that the sender can transmit in addition
in addition to the existing flow control window. The legal range for to the existing flow control window. The legal range for this field
this field is 1 to 2^31 - 1 (0x7fffffff) bytes; the most significant is 1 to 2^31 - 1 (0x7fffffff) bytes; the most significant bit of this
bit of this value is reserved. value is reserved.
3.7.9.1. The Flow Control Window 3.8.9.1. The Flow Control Window
Flow control in HTTP/2.0 is implemented by a flow control window kept Flow control in HTTP/2.0 is implemented using a window kept by each
by the sender of each stream. The flow control window is a simple sender on every stream. The flow control window is a simple integer
integer value that indicates how many bytes of data the sender is value that indicates how many bytes of data the sender is permitted
permitted to transmit. The flow control window size is a measure of to transmit; as such, its size is a measure of the buffering
the buffering capability of the recipient. capability of the receiver.
Two flow control windows apply to the sending of every message: the Two flow control windows are applicable; the stream flow control
stream flow control window and the session flow control window. The window and the connection flow control window. The sender MUST NOT
sender MUST NOT send a flow controlled frame with a length that send a flow controlled frame with a length that exceeds the space
exceeds the space available in either of the flow control windows available in either of the flow control windows advertised by the
advertised by the receiver. Frames with zero length with the FINAL receiver. Frames with zero length with the FINAL flag set (for
flag set (for example, an empty data frame) MAY be sent if there is example, an empty data frame) MAY be sent if there is no available
no available space in either flow control window. space in either flow control window.
For flow control calculations, the 8 byte frame header is not For flow control calculations, the 8 byte frame header is not
counted. counted.
After sending a flow controlled frame, the sender reduces the space After sending a flow controlled frame, the sender reduces the space
available in both windows by the length of the transmitted frame. available in both windows by the length of the transmitted frame.
The receiver of a message sends a WINDOW_UPDATE frame as it consumes The receiver of a frame sends a WINDOW_UPDATE frame as it consumes
data and frees up space in flow control windows. Separate data and frees up space in flow control windows. Separate
WINDOW_UPDATE messages are sent for the stream and session level flow WINDOW_UPDATE frames are sent for the stream and connection level
control windows. flow control windows.
A sender that receives a WINDOW_UPDATE frame updates the A sender that receives a WINDOW_UPDATE frame updates the
corresponding window by the amount specified in the frame. corresponding window by the amount specified in the frame.
A sender MUST NOT allow a flow control window to exceed 2^31 - 1 A sender MUST NOT allow a flow control window to exceed 2^31 - 1
bytes. If a sender receives a WINDOW_UPDATE that causes a flow bytes. If a sender receives a WINDOW_UPDATE that causes a flow
control window to exceed this maximum it MUST terminate either the control window to exceed this maximum it MUST terminate either the
stream or the session, as appropriate. For streams, the sender sends stream or the connection, as appropriate. For streams, the sender
a RST_STREAM with the error code of FLOW_CONTROL_ERROR code; for the sends a RST_STREAM with the error code of FLOW_CONTROL_ERROR code;
session, a GOAWAY message with a FLOW_CONTROL_ERROR code. for the connection, a GOAWAY frame with a FLOW_CONTROL_ERROR code.
Flow controlled frames from the sender and WINDOW_UPDATE frames from Flow controlled frames from the sender and WINDOW_UPDATE frames from
the receiver are completely asynchronous with respect to each other. the receiver are completely asynchronous with respect to each other.
This property allows a receiver to aggressively update the window This property allows a receiver to aggressively update the window
size kept by the sender to prevent streams from stalling. size kept by the sender to prevent streams from stalling.
3.7.9.2. Initial Flow Control Window Size 3.8.9.2. Initial Flow Control Window Size
When a HTTP/2.0 connection is first established, new streams are When a HTTP/2.0 connection is first established, new streams are
created with an initial flow control window size of 65535 bytes. The created with an initial flow control window size of 65535 bytes. The
session flow control window is 65536 bytes. Both endpoints can connection flow control window is 65536 bytes. Both endpoints can
adjust the initial window size for new streams by including a value adjust the initial window size for new streams by including a value
for SETTINGS_INITIAL_WINDOW_SIZE in the SETTINGS frame that forms for SETTINGS_INITIAL_WINDOW_SIZE in the SETTINGS frame that forms
part of the session header. part of the connection header.
Prior to receiving a SETTINGS frame that sets a value for Prior to receiving a SETTINGS frame that sets a value for
SETTINGS_INITIAL_WINDOW_SIZE, a client can only use the default SETTINGS_INITIAL_WINDOW_SIZE, a client can only use the default
initial window size when sending flow controlled frames. Similarly, initial window size when sending flow controlled frames. Similarly,
the session flow control window is set to the default initial window the connection flow control window is set to the default initial
size until a WINDOW_UPDATE message is received. window size until a WINDOW_UPDATE frame is received.
A SETTINGS frame can alter the initial flow control window size for A SETTINGS frame can alter the initial flow control window size for
all current streams. When the value of SETTINGS_INITIAL_WINDOW_SIZE all current streams. When the value of SETTINGS_INITIAL_WINDOW_SIZE
changes, a receiver MUST adjust the size of all flow control windows changes, a receiver MUST adjust the size of all flow control windows
that it maintains by the difference between the new value and the old that it maintains by the difference between the new value and the old
value. value.
A change to SETTINGS_INITIAL_WINDOW_SIZE could cause the available A change to SETTINGS_INITIAL_WINDOW_SIZE could cause the available
space in a flow control window to become negative. A sender MUST space in a flow control window to become negative. A sender MUST
track the negative flow control window and not send new flow track the negative flow control window, and MUST NOT send new flow
controlled frames until it receives WINDOW_UPDATE messages that cause controlled frames until it receives WINDOW_UPDATE frames that cause
the flow control window to become positive. the flow control window to become positive.
For example, if the server sets the initial window size to be 16KB, For example, if the server sets the initial window size to be 16KB,
and the client sends 64KB immediately on connection establishment, and the client sends 64KB immediately on connection establishment,
the client will recalculate the available flow control window to be the client will recalculate the available flow control window to be
-48KB on receipt of the SETTINGS frame. The client retains a -48KB on receipt of the SETTINGS frame. The client retains a
negative flow control window until WINDOW_UPDATE frames restore the negative flow control window until WINDOW_UPDATE frames restore the
window to being positive, after which the client can resume sending. window to being positive, after which the client can resume sending.
3.7.9.3. Reducing the Stream Window Size 3.8.9.3. Reducing the Stream Window Size
A receiver that wishes to use a smaller flow control window than the A receiver that wishes to use a smaller flow control window than the
current size sends a new SETTINGS frame. However, the receiver MUST current size can send a new SETTINGS frame. However, the receiver
be prepared to receive data that exceeds this window size, since the MUST be prepared to receive data that exceeds this window size, since
sender might send data that exceeds the lower limit prior to the sender might send data that exceeds the lower limit prior to
processing the SETTINGS frame. processing the SETTINGS frame.
A receiver has two options for handling streams that exceed flow A receiver has two options for handling streams that exceed flow
control limits: control limits:
1. The receiver can immediately send RST_STREAM with 1. The receiver can immediately send RST_STREAM with
FLOW_CONTROL_ERROR error code for the affected streams. FLOW_CONTROL_ERROR error code for the affected streams.
2. The receiver can accept the streams and tolerate the resulting 2. The receiver can accept the streams and tolerate the resulting
head of line blocking, sending WINDOW_UPDATE messages as it head of line blocking, sending WINDOW_UPDATE frames as it
consumes data. consumes data.
If a receiver decides to accept streams, both sides must recompute If a receiver decides to accept streams, both sides MUST recompute
the available flow control window based on the initial window size the available flow control window based on the initial window size
sent in the SETTINGS. sent in the SETTINGS.
3.7.9.4. Ending Flow Control 3.8.9.4. Ending Flow Control
After a recipient reads in a frame that marks the end of a stream After a receiver reads in a frame that marks the end of a stream (for
(for example, a data stream with a FINAL flag set), it ceases example, a data stream with a FINAL flag set), it MUST cease
transmission of WINDOW_UPDATE frames. A sender is not required to transmission of WINDOW_UPDATE frames for that stream. A sender is
maintain the available flow control window for streams that it is no not obligated to maintain the available flow control window for
longer sending on. streams that it is no longer sending on.
Flow control can be disabled for all streams or the session using the Flow control can be disabled for all streams or the connection using
SETTINGS_FLOW_CONTROL_OPTIONS setting. An implementation that does the SETTINGS_FLOW_CONTROL_OPTIONS setting. An implementation that
not wish to perform flow control can use this in the initial SETTINGS does not wish to perform flow control can use this in the initial
exchange. SETTINGS exchange.
Flow control can be disabled for an individual stream or the overall Flow control can be disabled for an individual stream or the overall
session by sending a WINDOW_UPDATE with the END_FLOW_CONTROL flag connection by sending a WINDOW_UPDATE with the END_FLOW_CONTROL flag
set. The payload of a WINDOW_UPDATE frame that has the set. The payload of a WINDOW_UPDATE frame that has the
END_FLOW_CONTROL flag set is ignored. END_FLOW_CONTROL flag set is ignored.
Flow control cannot be enabled again once disabled. Any attempt to Flow control cannot be enabled again once disabled. Any attempt to
re-enable flow control - by sending a WINDOW_UPDATE or by clearing re-enable flow control - by sending a WINDOW_UPDATE or by clearing
the bits on the SETTINGS_FLOW_CONTROL_OPTIONS setting - MUST be the bits on the SETTINGS_FLOW_CONTROL_OPTIONS setting - MUST be
rejected with a FLOW_CONTROL_ERROR error code. rejected with a FLOW_CONTROL_ERROR error code.
3.7.10. Header Block
The header block is found in the HEADERS, HEADERS+PRIORITY and
PUSH_PROMISE frames. The header block consists of a set of header
fields, which are name-value pairs. Headers are compressed using
black magic.
Compression of header fields is a work in progress, as is the format
of this block.
4. HTTP Message Exchanges 4. HTTP Message Exchanges
HTTP/2.0 is intended to be as compatible as possible with current HTTP/2.0 is intended to be as compatible as possible with current
web-based applications. This means that, from the perspective of the web-based applications. This means that, from the perspective of the
server business logic or application API, the features of HTTP are server business logic or application API, the features of HTTP are
unchanged. To achieve this, all of the application request and unchanged. To achieve this, all of the application request and
response header semantics are preserved, although the syntax of response header semantics are preserved, although the syntax of
conveying those semantics has changed. Thus, the rules from HTTP/1.1 conveying those semantics has changed. Thus, the rules from HTTP/1.1
([HTTP-p1], [HTTP-p2], [HTTP-p4], [HTTP-p5], [HTTP-p6], and ([HTTP-p1], [HTTP-p2], [HTTP-p4], [HTTP-p5], [HTTP-p6], and
[HTTP-p7]) apply with the changes in the sections below. [HTTP-p7]) apply with the changes in the sections below.
4.1. Connection Management 4.1. Connection Management
Clients SHOULD NOT open more than one HTTP/2.0 session to a given Clients SHOULD NOT open more than one HTTP/2.0 connection to a given
origin ([RFC6454]) concurrently. origin ([RFC6454]) concurrently.
Note that it is possible for one HTTP/2.0 session to be finishing Note that it is possible for one HTTP/2.0 connection to be finishing
(e.g. a GOAWAY message has been sent, but not all streams have (e.g. a GOAWAY frame has been sent, but not all streams have
finished), while another HTTP/2.0 session is starting. finished), while another HTTP/2.0 connection is starting.
4.1.1. Use of GOAWAY
HTTP/2.0 provides a GOAWAY message which can be used when closing a
connection from either the client or server. Without a server GOAWAY
message, HTTP has a race condition where the client sends a request
just as the server is closing the connection, and the client cannot
know if the server received the stream or not. By using the last-
stream-id in the GOAWAY, servers can indicate to the client if a
request was processed or not.
Note that some servers will choose to send the GOAWAY and immediately
terminate the connection without waiting for active streams to
finish. The client will be able to determine this because HTTP/2.0
streams are deterministically closed. This abrupt termination will
force the client to heuristically decide whether to retry the pending
requests. Clients always need to be capable of dealing with this
case because they must deal with accidental connection termination
cases, which are the same as the server never having sent a GOAWAY.
More sophisticated servers will use GOAWAY to implement a graceful
teardown. They will send the GOAWAY and provide some time for the
active streams to finish before terminating the connection.
If a HTTP/2.0 client closes the connection, it should also send a
GOAWAY message. This allows the server to know if any server-push
streams were received by the client.
If the endpoint closing the connection has not received frames on any
stream, the GOAWAY will contain a last-stream-id of 0.
4.2. HTTP Request/Response 4.2. HTTP Request/Response
4.2.1. HTTP Header Fields and HTTP/2.0 Headers 4.2.1. HTTP Header Fields and HTTP/2.0 Headers
At the application level, HTTP uses name-value pairs in its header At the application level, HTTP uses name-value pairs in its header
fields. Because HTTP/2.0 merges the existing HTTP header fields with fields. Because HTTP/2.0 merges the existing HTTP header fields with
HTTP/2.0 headers, there is a possibility that some HTTP applications HTTP/2.0 headers, there is a possibility that some HTTP applications
already use a particular header field name. To avoid any conflicts, already use a particular header field name. To avoid any conflicts,
all header fields introduced for layering HTTP over HTTP/2.0 are all header fields introduced for layering HTTP over HTTP/2.0 are
skipping to change at page 30, line 10 skipping to change at page 33, line 29
The client initiates a request by sending a HEADERS+PRIORITY frame. The client initiates a request by sending a HEADERS+PRIORITY frame.
Requests that do not contain a body MUST set the FINAL flag, Requests that do not contain a body MUST set the FINAL flag,
indicating that the client intends to send no further data on this indicating that the client intends to send no further data on this
stream, unless the server intends to push resources (see stream, unless the server intends to push resources (see
Section 4.3). HEADERS+PRIORITY frame does not contain the FINAL flag Section 4.3). HEADERS+PRIORITY frame does not contain the FINAL flag
for requests that contain a body. The body of a request follows as a for requests that contain a body. The body of a request follows as a
series of DATA frames. The last DATA frame sets the FINAL flag to series of DATA frames. The last DATA frame sets the FINAL flag to
indicate the end of the body. indicate the end of the body.
The header fields included in the HEADERS+PRIORITY frame contain all The header fields included in the HEADERS+PRIORITY frame contain all
of the HTTP header fields that are associated with an HTTP request. of the HTTP header fields associated with an HTTP request. The
The header block in HTTP/2.0 is mostly unchanged from today's HTTP definitions of these headers are largely unchanged relative to
header block, with the following differences: HTTP/1.1, with a few notable exceptions:
The following fields that are carried in the request line in
HTTP/1.1 ([HTTP-p1], Section 3.1.1) are defined as special-valued
name-value pairs:
":method": the HTTP method for this request (e.g. "GET", "POST",
"HEAD", etc) ([HTTP-p2], Section 4)
":path": ":path" - the request-target for this URI with "/"
prefixed (see [HTTP-p1], Section 3.1.1). For example, for
"http://www.google.com/search?q=dogs" the path would be
"/search?q=dogs". [[anchor26: what forms of the HTTPbis
request-target are allowed here?]]
These header fields MUST be present in HTTP requests.
In addition, the following two name-value pairs MUST be present in
every request:
":host": the host and optional port portions (see [RFC3986],
Section 3.2) of the URI for this request (e.g. "www.google.com:
1234"). This header field is the same as the HTTP 'Host'
header field ([HTTP-p1], Section 5.4).
":scheme": the scheme portion of the URI for this request (e.g.
"https")
All header field names starting with ":" (whether defined in this o The HTTP/1.1 request-line has been split into two separate header
document or future extensions to this document) MUST appear before fields named :method and :path, whose values specify the HTTP
any other header fields. method for the request and the request-target, respectively. The
HTTP-version component of the request-line is removed entirely
from the headers.
Header field names MUST be all lowercase. o The host and optional port portions of the request URI (see
[RFC3986], Section 3.2), is specified using the new :host header
field. [[anchor21: Ed. Note: it needs to be clarified whether or
not this replaces the existing HTTP/1.1 Host header.]]
The Connection, Host, Keep-Alive, Proxy-Connection, and Transfer- o A new :scheme header field has been added to specify the scheme
Encoding header fields are not valid and MUST not be sent. portion of the request-target (e.g. "https")
User-agents MUST support gzip compression. Regardless of the o All header field names MUST be lowercased, and the definitions of
Accept-Encoding sent by the user-agent, the server may always send all header field names defined by HTTP/1.1 are updated to be all
content encoded with gzip or deflate encoding. [[anchor27: Still lowercase.
valid?]]
If a server receives a request where the sum of the data frame o The Connection, Host, Keep-Alive, Proxy-Connection, and Transfer-
payload lengths does not equal the size of the Content-Length Encoding header fields are no longer valid and MUST not be sent.
header field, the server MUST return a 400 (Bad Request) error.
Although POSTs are inherently chunked, POST requests SHOULD also All HTTP Requests MUST include the ":method", ":path", ":host", and
be accompanied by a Content-Length header field. First, it ":scheme" header fields.
informs the server of how much data to expect, which the server
can used to track overall progress and provide appropriate user
feedback. More importantly, some HTTP server implementations fail
to correctly process requests that omit the Content-Length header
field. Many existing clients send a Content-Length header field,
which caused server implementations have come to depend upon its
presence.
The user-agent is free to prioritize requests as it sees fit. If the Header fields whose names begin with ":" (whether defined in this
user-agent cannot make progress without receiving a resource, it document or future extensions to this document) MUST appear before
should attempt to raise the priority of that resource. Resources any other header fields.
such as images, SHOULD generally use the lowest priority.
If a client sends a HEADERS+PRIORITY frame that omits a mandatory If a client sends a HEADERS+PRIORITY frame that omits a mandatory
header, the server MUST reply with a HTTP 400 Bad Request reply. header, the server MUST reply with a HTTP 400 Bad Request reply.
[[anchor28: Ed: why PROTOCOL_ERROR on missing ":status" in the [[anchor22: Ed: why PROTOCOL_ERROR on missing ":status" in the
response, but HTTP 400 here?]] response, but HTTP 400 here?]]
If the server receives a data frame prior to a HEADERS or HEADERS+ If a server receives a request where the sum of the data frame
PRIORITY frame the server MUST treat this as a stream error payload lengths does not equal the size of the Content-Length header
(Section 3.5.2) of type PROTOCOL_ERROR. field, the server MUST return a 400 (Bad Request) error.
Although POSTs are inherently chunked, POST requests SHOULD also be
accompanied by a Content-Length header field. First, it informs the
server of how much data to expect, which the server can use to track
overall progress and provide appropriate user feedback. More
importantly, some HTTP server implementations fail to correctly
process requests that omit the Content-Length header field. Many
existing clients send a Content-Length header field, and some server
implementations have come to depend upon its presence.
A client provides priority in requests as a hint to the server. A
server SHOULD attempt to provide responses to higher priority
requests before lower priority requests. A server could send lower
priority responses during periods that higher priority responses are
unavailable to ensure better utilization of a connection.
If the server receives a data frame prior to a HEADERS+PRIORITY frame
the server MUST treat this as a stream error (Section 3.5.2) of type
PROTOCOL_ERROR.
4.2.3. Response 4.2.3. Response
The server responds to a client request with a HEADERS frame. The server responds to a client request using the same stream
Symmetric to the client's upload stream, server will send any identifier that was used by the request. An HTTP response begins
response body in a series of DATA frames. The last data frame will with a HEADERS frame. An HTTP response body consists of a series of
contain the FINAL flag to indicate the end of the stream and the end DATA frames. The last data frame contains a FINAL flag to indicate
of the response. A response that contains no body (such as a 204 or the end of the response. A response that contains no body (such as a
304 response) consists only of a HEADERS frame that contains the 204 or 304 response) consists only of a HEADERS frame that contains
FINAL flag to indicate no further data will be sent on the stream. the FINAL flag to indicate no further data will be sent on the
stream.
The response status line is unfolded into name-value pairs like The response status line is unfolded into name-value pairs like
other HTTP header fields and must be present: other HTTP header fields and must be present:
":status": The HTTP response status code (e.g. "200" or "200 OK") ":status": The HTTP response status code (e.g. "200" or "200 OK")
All header field names starting with ":" (whether defined in this All header field names starting with ":" (whether defined in this
document or future extensions to this document) MUST appear before document or future extensions to this document) MUST appear before
any other header fields. any other header fields.
skipping to change at page 32, line 18 skipping to change at page 35, line 27
Encoding header fields are not valid and MUST not be sent. Encoding header fields are not valid and MUST not be sent.
Responses MAY be accompanied by a Content-Length header field for Responses MAY be accompanied by a Content-Length header field for
advisory purposes. This allows clients to learn the full size of advisory purposes. This allows clients to learn the full size of
an entity prior to receiving all the data frames. This can help an entity prior to receiving all the data frames. This can help
in, for example, reporting progress. in, for example, reporting progress.
If a client receives a response where the sum of the data frame If a client receives a response where the sum of the data frame
payload length does not equal the size of the Content-Length payload length does not equal the size of the Content-Length
header field, the client MUST ignore the content length header header field, the client MUST ignore the content length header
field. [[anchor29: Ed: See field. [[anchor23: Ed: See
<https://github.com/http2/http2-spec/issues/46>.]] <https://github.com/http2/http2-spec/issues/46>.]]
If a client receives a response with an absent or duplicated status If a client receives a response with an absent or duplicated status
header, the client MUST treat this as a stream error (Section 3.5.2) header, the client MUST treat this as a stream error (Section 3.5.2)
of type PROTOCOL_ERROR. of type PROTOCOL_ERROR.
If the client receives a data frame prior to a HEADERS or HEADERS+ If the client receives a data frame prior to a HEADERS frame the
PRIORITY frame the client MUST treat this as a stream error client MUST treat this as a stream error (Section 3.5.2) of type
(Section 3.5.2) of type PROTOCOL_ERROR. PROTOCOL_ERROR.
Clients MUST support gzip compression. Regardless of the value of
the Accept-Encoding header field, a server MAY send responses with
gzip or deflate encoding. A compressed response MUST still bear an
appropriate Content-Encoding header field.
4.3. Server Push Transactions 4.3. Server Push Transactions
HTTP/2.0 enables a server to send multiple replies to a client for a HTTP/2.0 enables a server to send multiple replies to a client for a
single request. The rationale for this feature is that sometimes a single request. The rationale for this feature is that sometimes a
server knows that it will need to send multiple resources in response server knows that it will need to send multiple resources in response
to a single request. Without server push features, the client must to a single request. Without server push features, the client must
first download the primary resource, then discover the secondary first download the primary resource, then discover the secondary
resource(s), and request them. Pushing of resources avoids the resource(s), and request them.
round-trip delay, but also creates a potential race where a server
can be pushing content which a user-agent is in the process of
requesting. The following mechanics attempt to prevent the race
condition while enabling the performance benefit.
Server push is an optional feature. Server push can be disabled by Server push is an optional feature. The
clients that do not wish to receive pushed resources by advertising a SETTINGS_MAX_CONCURRENT_STREAMS setting from the client limits the
SETTINGS_MAX_CONCURRENT_STREAMS SETTING (Section 3.7.4) of zero. number of resources that can be concurrently pushed by a server.
This prevents servers from creating the streams necessary to push Server push can be disabled by clients that do not wish to receive
resources. pushed resources by advertising a SETTINGS_MAX_CONCURRENT_STREAMS
SETTING (Section 3.8.4) of zero. This prevents servers from creating
the streams necessary to push resources.
Browsers receiving a pushed response MUST validate that the server is Clients receiving a pushed response MUST validate that the server is
authorized to push the resource using the same-origin policy authorized to push the resource using the same-origin policy
([RFC6454], Section 3). For example, a HTTP/2.0 connection to ([RFC6454], Section 3). For example, a HTTP/2.0 connection to
"example.com" is generally [[anchor30: Ed: weaselly use of "example.com" is generally [[anchor24: Ed: weaselly use of
"generally", needs better definition]] not permitted to push a "generally", needs better definition]] not permitted to push a
response for "www.example.org". response for "www.example.org".
A client that accepts pushed resources caches those resources as A client that accepts pushed resources caches those resources as
though they were responses to GET requests. though they were responses to GET requests.
Pushing of resources avoids the round-trip delay, but also creates a
potential race where a server can be pushing content which a client
is in the process of requesting. The PUSH_PROMISE frame reduces the
chances of this condition occurring, while retaining the performance
benefit.
Pushed responses are associated with a request at the HTTP/2.0 Pushed responses are associated with a request at the HTTP/2.0
framing layer. The PUSH_PROMISE includes a stream identifier for an framing layer. The PUSH_PROMISE is sent on the stream for the
associated request/response exchange that supplies request header associated request, which allows a receiver to correlate the pushed
fields. The pushed stream inherits all of the request header fields resource with a request. The pushed stream inherits all of the
from the associated stream with the exception of resource request header fields from the associated stream with the exception
identification header fields (":host", ":scheme", and ":path"), which of resource identification header fields (":host", ":scheme", and
are provided as part of the PUSH_PROMISE frame. Pushed resources ":path"), which are provided as part of the PUSH_PROMISE frame.
always have an associated ":method" of "GET". A cache MUST store
these inherited and implied request header fields with the cached
resource.
Implementation note: With server push, it is theoretically possible Pushed resources always have an associated ":method" of "GET". A
for servers to push unreasonable amounts of content or resources to cache MUST store these inherited and implied request header fields
the user-agent. Browsers MUST implement throttles to protect against with the cached resource.
unreasonable push attacks. [[anchor31: Ed: insufficiently specified
to implement; would like to remove]]
4.3.1. Server implementation 4.3.1. Server implementation
A server pushes resources in association with a request from the A server pushes resources in association with a request from the
client. Prior to closing the response stream, the server sends a client. Prior to closing the response stream, the server sends a
PUSH_PROMISE for each resource that it intends to push. The PUSH_PROMISE for each resource that it intends to push. The
PUSH_PROMISE includes header fields that allow the client to identify PUSH_PROMISE includes header fields that allow the client to identify
the resource (":scheme", ":host", and ":port"). the resource (":scheme", ":host", and ":path").
A server can push multiple resources in response to a request, but A server can push multiple resources in response to a request, but
these can only be sent while the response stream remains open. A all pushed resources MUST be promised on the response stream for the
server MUST NOT send a PUSH_PROMISE on a half-closed stream. associated request. A server cannot send a PUSH_PROMISE on a new
stream or a half-closed stream.
The server SHOULD include any header fields in a PUSH_PROMISE that The server SHOULD include any header fields in a PUSH_PROMISE that
would allow a cache to determine if the resource is already cached would allow a cache to determine if the resource is already cached
(see [HTTP-p6], Section 4). (see [HTTP-p6], Section 4).
After sending a PUSH_PROMISE, the server commences transmission of a After sending a PUSH_PROMISE, the server commences transmission of a
pushed resource. A pushed resource uses a server-initiated stream. pushed resource. A pushed resource uses a server-initiated stream.
The server sends frames on this stream in the same order as an HTTP The server sends frames on this stream in the same order as an HTTP
response (Section 4.2.3): a HEADERS frame followed by DATA frames. response (Section 4.2.3): a HEADERS frame followed by DATA frames.
skipping to change at page 34, line 26 skipping to change at page 37, line 41
4.3.2. Client implementation 4.3.2. Client implementation
When fetching a resource the client has 3 possibilities: When fetching a resource the client has 3 possibilities:
1. the resource is not being pushed 1. the resource is not being pushed
2. the resource is being pushed, but the data has not yet arrived 2. the resource is being pushed, but the data has not yet arrived
3. the resource is being pushed, and the data has started to arrive 3. the resource is being pushed, and the data has started to arrive
When a HEADERS+PRIORITY frame that contains an A client SHOULD NOT issue GET requests for a resource that has been
Associated-To-Stream-ID is received, the client MUST NOT[[anchor34: promised. A client is instead advised to wait for the pushed
SHOULD NOT?]] issue GET requests for the resource in the pushed resource to arrive.
stream, and instead wait for the pushed stream to arrive.
A server MUST NOT push a resource with an Associated-To-Stream-ID of
0. Clients MUST treat this as a session error (Section 3.5.1) of
type PROTOCOL_ERROR.
When a client receives a PUSH_PROMISE frame from the server without a When a client receives a PUSH_PROMISE frame from the server without a
the ":host", ":scheme", and ":path" header fields, it MUST treat this the ":host", ":scheme", and ":path" header fields, it MUST treat this
as a stream error (Section 3.5.2) of type PROTOCOL_ERROR. as a stream error (Section 3.5.2) of type PROTOCOL_ERROR.
To cancel individual server push streams, the client can issue a To cancel individual server push streams, the client can issue a
stream error (Section 3.5.2) of type CANCEL. Upon receipt, the stream error (Section 3.5.2) of type CANCEL. After receiving a
server ceases transmission of the pushed data. PUSH_PROMISE frame, the client is able to cancel the pushed resource
before receiving any frames on the promised stream. The server
ceases transmission of the pushed resource; if the server has not
commenced transmission, it does not start.
To cancel all server push streams related to a request, the client To cancel all server push streams related to a request, the client
may issue a stream error (Section 3.5.2) of type CANCEL on the may issue a stream error (Section 3.5.2) of type CANCEL on the
associated-stream-id. By cancelling that stream, the server MUST associated-stream-id. By cancelling that stream, the server MUST
immediately stop sending frames for any streams with immediately stop sending frames for any streams with
in-association-to for the original stream. [[anchor35: Ed: Triggering in-association-to for the original stream. [[anchor27: Ed: Triggering
side-effects on stream reset is going to be problematic for the side-effects on stream reset is going to be problematic for the
framing layer. Purely from a design perspective, it's a layering framing layer. Purely from a design perspective, it's a layering
violation. More practically speaking, the base request stream might violation. More practically speaking, the base request stream might
already be removed. Special handling logic would be required.]] already be removed. Special handling logic would be required.]]
A client can choose to time out pushed streams if the server does not
provide the resource in a timely fashion. A stream error
(Section 3.5.2) of type CANCEL can be used to stop a timed out push.
If the server sends a HEADERS frame containing header fields that If the server sends a HEADERS frame containing header fields that
duplicate values on a previous HEADERS or PUSH_PROMISE frames on the duplicate values on a previous HEADERS or PUSH_PROMISE frames on the
same stream, the client MUST treat this as a stream error same stream, the client MUST treat this as a stream error
(Section 3.5.2) of type PROTOCOL_ERROR. (Section 3.5.2) of type PROTOCOL_ERROR.
If the server sends a HEADERS frame after sending a data frame for If the server sends a HEADERS frame after sending a data frame for
the same stream, the client MAY ignore the HEADERS frame. Ignoring the same stream, the client MAY ignore the HEADERS frame. Ignoring
the HEADERS frame after a data frame prevents handling of HTTP's the HEADERS frame after a data frame prevents handling of HTTP's
trailing header fields (Section 4.1.1 of [HTTP-p1]). trailing header fields (Section 4.1.1 of [HTTP-p1]).
skipping to change at page 35, line 44 skipping to change at page 39, line 13
relates to existing HTTP implementations. However, the ability to relates to existing HTTP implementations. However, the ability to
reuse the HTTP/2.0 framing layer is a non goal. reuse the HTTP/2.0 framing layer is a non goal.
5.2. Error handling - Framing Layer 5.2. Error handling - Framing Layer
Error handling at the HTTP/2.0 layer splits errors into two groups: Error handling at the HTTP/2.0 layer splits errors into two groups:
Those that affect an individual HTTP/2.0 stream, and those that do Those that affect an individual HTTP/2.0 stream, and those that do
not. not.
When an error is confined to a single stream, but general framing is When an error is confined to a single stream, but general framing is
in tact, HTTP/2.0 attempts to use the RST_STREAM as a mechanism to intact, HTTP/2.0 attempts to use the RST_STREAM as a mechanism to
invalidate the stream but move forward without aborting the invalidate the stream but move forward without aborting the
connection altogether. connection altogether.
For errors occurring outside of a single stream context, HTTP/2.0 For errors occurring outside of a single stream context, HTTP/2.0
assumes the entire session is hosed. In this case, the endpoint assumes the entire connection is hosed. In this case, the endpoint
detecting the error should initiate a connection close. detecting the error should initiate a connection close.
5.3. One Connection Per Domain 5.3. One Connection per Domain
HTTP/2.0 attempts to use fewer connections than other protocols have HTTP/2.0 attempts to use fewer connections than other protocols have
traditionally used. The rationale for this behavior is because it is traditionally used. The rationale for this behavior is because it is
very difficult to provide a consistent level of service (e.g. TCP very difficult to provide a consistent level of service (e.g. TCP
slow-start), prioritization, or optimal compression when the client slow-start), prioritization, or optimal compression when the client
is connecting to the server through multiple channels. is connecting to the server through multiple channels.
Through lab measurements, we have seen consistent latency benefits by Through lab measurements, we have seen consistent latency benefits by
using fewer connections from the client. The overall number of using fewer connections from the client. The overall number of
packets sent by HTTP/2.0 can be as much as 40% less than HTTP. packets sent by HTTP/2.0 can be as much as 40% less than HTTP.
skipping to change at page 37, line 10 skipping to change at page 40, line 27
A subtle but important point is that server push streams must be A subtle but important point is that server push streams must be
declared before the associated stream is closed. The reason for this declared before the associated stream is closed. The reason for this
is so that proxies have a lifetime for which they can discard is so that proxies have a lifetime for which they can discard
information about previous streams. If a pushed stream could information about previous streams. If a pushed stream could
associate itself with an already-closed stream, then endpoints would associate itself with an already-closed stream, then endpoints would
not have a specific lifecycle for when they could disavow knowledge not have a specific lifecycle for when they could disavow knowledge
of the streams which went before. of the streams which went before.
6. Security Considerations 6. Security Considerations
6.1. Use of Same-origin constraints 6.1. Server Authority and Same-Origin
This specification uses the same-origin policy ([RFC6454], Section 3) This specification uses the same-origin policy ([RFC6454], Section 3)
in all cases where verification of content is required. to determine whether an origin server is permitted to provide
content.
A server that is contacted using TLS is authenticated based on the
certificate that it offers in the TLS handshake (see [RFC2818],
Section 3). A server is considered authoritative for an "https:"
resource if it has been successfully authenticated for the domain
part of the origin of the resource that it is providing.
A server is considered authoritative for an "http:" resource if the
connection is established to a resolved IP address for the domain in
the origin of the resource.
A client MUST NOT use, in any way, resources provided by a server
that is not authoritative for those resources.
6.2. Cross-Protocol Attacks 6.2. Cross-Protocol Attacks
By utilizing TLS, we believe that HTTP/2.0 introduces no new cross- When using TLS, we believe that HTTP/2.0 introduces no new cross-
protocol attacks. TLS encrypts the contents of all transmission protocol attacks. TLS encrypts the contents of all transmission
(except the handshake itself), making it difficult for attackers to (except the handshake itself), making it difficult for attackers to
control the data which could be used in a cross-protocol attack. control the data which could be used in a cross-protocol attack.
[[anchor45: Issue: This is no longer true]]
[[anchor37: Issue: This is no longer true]]
6.3. Cacheability of Pushed Resources 6.3. Cacheability of Pushed Resources
Pushed resources do not have an associated request. In order for Pushed resources are synthesized responses without an explicit
existing HTTP cache control validations (such as the Vary header request; the request for a pushed resource is synthesized from the
field) to work, all cached resources must have a set of request request that triggered the push, plus resource identification
header fields. For this reason, caches MUST be careful to inherit information provided by the server. Request header fields are
request header fields from the associated stream for the push. This necessary for HTTP cache control validations (such as the Vary header
includes the Cookie header field. field) to work. For this reason, caches MUST inherit request header
fields from the associated stream for the push. This includes the
Cookie header field.
Caching resources that are pushed is possible, based on the guidance Caching resources that are pushed is possible, based on the guidance
provided by the origin server in the Cache-Control header field. provided by the origin server in the Cache-Control header field.
However, this can cause issues if a single server hosts more than one However, this can cause issues if a single server hosts more than one
tenant. For example, a server might offer multiple users each a tenant. For example, a server might offer multiple users each a
small portion of its URI space. small portion of its URI space.
Where multiple tenants share space on the same server, that server Where multiple tenants share space on the same server, that server
MUST ensure that tenants are not able to push representations of MUST ensure that tenants are not able to push representations of
resources that they do not have authority over. Failure to enforce resources that they do not have authority over. Failure to enforce
skipping to change at page 40, line 29 skipping to change at page 44, line 16
and semantics. and semantics.
Description: A description of the setting. This might include the Description: A description of the setting. This might include the
range of values, any applicable units and how to act upon a value range of values, any applicable units and how to act upon a value
when it is provided. when it is provided.
Specification: An optional reference for a specification that Specification: An optional reference for a specification that
defines the setting. defines the setting.
An initial set of settings registrations can be found in An initial set of settings registrations can be found in
Section 3.7.4. Section 3.8.4.3.
9. Acknowledgements 9. Acknowledgements
This document includes substantial input from the following This document includes substantial input from the following
individuals: individuals:
o Adam Langley, Wan-Teh Chang, Jim Morrison, Mark Nottingham, Alyssa o Adam Langley, Wan-Teh Chang, Jim Morrison, Mark Nottingham, Alyssa
Wilk, Costin Manolache, William Chan, Vitaliy Lvin, Joe Chan, Adam Wilk, Costin Manolache, William Chan, Vitaliy Lvin, Joe Chan, Adam
Barth, Ryan Hamilton, Gavin Peters, Kent Alstad, Kevin Lindsay, Barth, Ryan Hamilton, Gavin Peters, Kent Alstad, Kevin Lindsay,
Paul Amer, Fan Yang, Jonathan Leighton (SPDY contributors). Paul Amer, Fan Yang, Jonathan Leighton (SPDY contributors).
o Gabriel Montenegro and Willy Tarreau (Upgrade mechanism) o Gabriel Montenegro and Willy Tarreau (Upgrade mechanism)
o William Chan, Salvatore Loreto, Osama Mazahir, Gabriel Montenegro, o William Chan, Salvatore Loreto, Osama Mazahir, Gabriel Montenegro,
Jitu Padhye, Roberto Peon, Rob Trace (Flow control) Jitu Padhye, Roberto Peon, Rob Trace (Flow control)
o Mark Nottingham and Julian Reschke o Mark Nottingham, Julian Reschke, James Snell (Editorial)
10. References 10. References
10.1. Normative References 10.1. Normative References
[HTTP-p1] Fielding, R. and J. Reschke, "Hypertext Transfer Protocol [HTTP-p1] Fielding, R. and J. Reschke, "Hypertext Transfer Protocol
(HTTP/1.1): Message Syntax and Routing", (HTTP/1.1): Message Syntax and Routing",
draft-ietf-httpbis-p1-messaging-22 (work in progress), draft-ietf-httpbis-p1-messaging-22 (work in progress),
February 2013. February 2013.
[HTTP-p2] Fielding, R. and J. Reschke, "Hypertext Transfer Protocol [HTTP-p2] Fielding, R. and J. Reschke, "Hypertext Transfer Protocol
(HTTP/1.1): Semantics and Content", (HTTP/1.1): Semantics and Content",
draft-ietf-httpbis-p2-semantics-22 (work in progress), draft-ietf-httpbis-p2-semantics-22 (work in progress),
skipping to change at page 41, line 42 skipping to change at page 45, line 26
Protocol (HTTP/1.1): Authentication", Protocol (HTTP/1.1): Authentication",
draft-ietf-httpbis-p7-auth-22 (work in progress), draft-ietf-httpbis-p7-auth-22 (work in progress),
February 2013. February 2013.
[RFC0793] Postel, J., "Transmission Control Protocol", STD 7, [RFC0793] Postel, J., "Transmission Control Protocol", STD 7,
RFC 793, September 1981. RFC 793, September 1981.
[RFC2119] Bradner, S., "Key words for use in RFCs to Indicate [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate
Requirement Levels", BCP 14, RFC 2119, March 1997. Requirement Levels", BCP 14, RFC 2119, March 1997.
[RFC2818] Rescorla, E., "HTTP Over TLS", RFC 2818, May 2000.
[RFC3986] Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform [RFC3986] Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform
Resource Identifier (URI): Generic Syntax", STD 66, Resource Identifier (URI): Generic Syntax", STD 66,
RFC 3986, January 2005. RFC 3986, January 2005.
[RFC5226] Narten, T. and H. Alvestrand, "Guidelines for Writing an [RFC5226] Narten, T. and H. Alvestrand, "Guidelines for Writing an
IANA Considerations Section in RFCs", BCP 26, RFC 5226, IANA Considerations Section in RFCs", BCP 26, RFC 5226,
May 2008. May 2008.
[RFC5246] Dierks, T. and E. Rescorla, "The Transport Layer Security [RFC5246] Dierks, T. and E. Rescorla, "The Transport Layer Security
(TLS) Protocol Version 1.2", RFC 5246, August 2008. (TLS) Protocol Version 1.2", RFC 5246, August 2008.
[RFC6454] Barth, A., "The Web Origin Concept", RFC 6454, [RFC6454] Barth, A., "The Web Origin Concept", RFC 6454,
December 2011. December 2011.
[TLSNPN] Langley, A., "Transport Layer Security (TLS) Next Protocol [TLSALPN] Friedl, S., Popov, A., Langley, A., and E. Stephan,
Negotiation Extension", draft-agl-tls-nextprotoneg-04 "Transport Layer Security (TLS) Application Layer Protocol
(work in progress), May 2012. Negotiation Extension", draft-ietf-tls-applayerprotoneg-01
(work in progress), April 2013.
10.2. Informative References 10.2. Informative References
[RFC1323] Jacobson, V., Braden, B., and D. Borman, "TCP Extensions [RFC1323] Jacobson, V., Braden, B., and D. Borman, "TCP Extensions
for High Performance", RFC 1323, May 1992. for High Performance", RFC 1323, May 1992.
[TALKING] Huang, L-S., Chen, E., Barth, A., Rescorla, E., and C. [TALKING] Huang, L-S., Chen, E., Barth, A., Rescorla, E., and C.
Jackson, "Talking to Yourself for Fun and Profit", 2011, Jackson, "Talking to Yourself for Fun and Profit", 2011,
<http://w2spconf.com/2011/papers/websocket.pdf>. <http://w2spconf.com/2011/papers/websocket.pdf>.
Appendix A. Change Log (to be removed by RFC Editor before publication) Appendix A. Change Log (to be removed by RFC Editor before publication)
A.1. Since draft-ietf-httpbis-http2-01 A.1. Since draft-ietf-httpbis-http2-02
Added continuations to frames carrying header blocks.
Replaced use of "session" with "connection" to avoid confusion with
other HTTP stateful concepts, like cookies.
Removed "message".
Switched to TLS ALPN from NPN.
Editorial changes.
A.2. Since draft-ietf-httpbis-http2-01
Added IANA considerations section for frame types, error codes and Added IANA considerations section for frame types, error codes and
settings. settings.
Removed data frame compression. Removed data frame compression.
Added PUSH_PROMISE. Added PUSH_PROMISE.
Added globally applicable flags to framing. Added globally applicable flags to framing.
skipping to change at page 43, line 7 skipping to change at page 47, line 7
Restructured frame header. Removed distinction between data and Restructured frame header. Removed distinction between data and
control frames. control frames.
Altered flow control properties to include session-level limits. Altered flow control properties to include session-level limits.
Added note on cacheability of pushed resources and multiple tenant Added note on cacheability of pushed resources and multiple tenant
servers. servers.
Changed protocol label form based on discussions. Changed protocol label form based on discussions.
A.2. Since draft-ietf-httpbis-http2-00 A.3. Since draft-ietf-httpbis-http2-00
Changed title throughout. Changed title throughout.
Removed section on Incompatibilities with SPDY draft#2. Removed section on Incompatibilities with SPDY draft#2.
Changed INTERNAL_ERROR on GOAWAY to have a value of 2 <https:// Changed INTERNAL_ERROR on GOAWAY to have a value of 2 <https://
groups.google.com/forum/?fromgroups#!topic/spdy-dev/cfUef2gL3iU>. groups.google.com/forum/?fromgroups#!topic/spdy-dev/cfUef2gL3iU>.
Replaced abstract and introduction. Replaced abstract and introduction.
Added section on starting HTTP/2.0, including upgrade mechanism. Added section on starting HTTP/2.0, including upgrade mechanism.
Removed unused references. Removed unused references.
Added flow control principles (Section 3.6.1) based on <http:// Added flow control principles (Section 3.6.1) based on <http://
tools.ietf.org/html/draft-montenegro-httpbis-http2-fc-principles-01>. tools.ietf.org/html/draft-montenegro-httpbis-http2-fc-principles-01>.
A.3. Since draft-mbelshe-httpbis-spdy-00 A.4. Since draft-mbelshe-httpbis-spdy-00
Adopted as base for draft-ietf-httpbis-http2. Adopted as base for draft-ietf-httpbis-http2.
Updated authors/editors list. Updated authors/editors list.
Added status note. Added status note.
Authors' Addresses Authors' Addresses
Mike Belshe Mike Belshe
skipping to change at page 44, line 7 skipping to change at page 48, line 7
EMail: mbelshe@chromium.org EMail: mbelshe@chromium.org
Roberto Peon Roberto Peon
Google, Inc Google, Inc
EMail: fenix@google.com EMail: fenix@google.com
Martin Thomson (editor) Martin Thomson (editor)
Microsoft Microsoft
3210 Porter Drive 3210 Porter Drive
Palo Alto 94043 Palo Alto 94304
US US
EMail: martin.thomson@skype.net EMail: martin.thomson@skype.net
Alexey Melnikov (editor) Alexey Melnikov (editor)
Isode Ltd Isode Ltd
5 Castle Business Village 5 Castle Business Village
36 Station Road 36 Station Road
Hampton, Middlesex TW12 2BX Hampton, Middlesex TW12 2BX
UK UK
 End of changes. 263 change blocks. 
799 lines changed or deleted 988 lines changed or added

This html diff was produced by rfcdiff 1.41. The latest version is available from http://tools.ietf.org/tools/rfcdiff/