| Network Working Group | C. Jennings |
| Internet-Draft | Cisco |
| Intended status: Informational | February 18, 2013 |
| Expires: August 22, 2013 |
Proposed Plan for Usage of SDP and RTP
draft-jennings-rtcweb-plan-00
This draft outlines a bunch of the remaining issues in RTCWeb related to how the the W3C APIs map to various usages of RTP and the associated SDP. It proposes one possible solution to that problem and outlines several chunks of work that would need to be put into other drafts or result in new drafts being written. The underlying design guideline is to, as much as possible, re-use what is already defined in existing SDP [RFC4566] and RTP [RFC3550] specifications.
This draft is not intended to become an specification but is meant for working group discussion to help build the specifications. It is being discussed on the webrtc@ietf.org mailing list though it has topics relating to the CLUE WG, MMUSIC WG, AVT* WG, and WebRTC WG at W3C.
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at http://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."
This Internet-Draft will expire on August 22, 2013.
Copyright (c) 2013 IETF Trust and the persons identified as the document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.
This document may not be modified, and derivative works of it may not be created, and it may not be published except as an Internet-Draft.
The reoccurring theme of this draft is that SDP [RFC4566] already has a way of solving the problems being discussed at the RTCWeb WG and we not try to invent something new but rather re-use the existing methods instead.
This does results in lots of m lines but all the alternatives resulted in an nearly equivalent number of SSRC lines with a possibility of redefining most of the media level attributes. So it's really hard to see the big difference.This assumes that it is perfectly feasible to transport SDP that much larger than a single MTU. The SIP [RFC3261] usage of SDP has successfully passed over this long ago. In the cases where the SDP is passed over web mechanisms, it is easy to use compression and it is more of an optimization criteria than a limiting issue.
The key words "MUST", "MUST NOT", "REQUIRED", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119].
This draft uses the API and terminology described in [webrtc-api].
Transport-Flow: 5 Tuple representing a RTP association.
5-tuple: A collection of the following values: source IP address, source transport port, destination IP address, destination transport port and transport protocol.
PC-Track: A source of media (audio and/or video) that is contained in a PC-Stream. A PC-Track represents content comprising one or more PC-Channels.
PC-Channel: Smallest unit of a PC-Track representing inter-related media aspects such as stereo or 5.1 audio signal
PC-Stream: Represents stream of data of audio and/or video added to a Peer Connection by local or remote media source(s). A PC-Stream is made up of zero or more PC-Tracks.
m-line: An RFC4566 [RFC4566] media description identifier that starts with "m=" field and conveys following values:media type,transport port,transport protocol and media format descriptions.
m-block: An RFC4566 [RFC4566] media description that starts with an m-line and is terminated by either the next m-line or by the end of the session description.
Offer: An [RFC3264] SDP message generated by the participant who wishes to initiate a multimedia communication session. An Offer describes participants capabilities for engaging in a multimedia session.
Answer: An [RFC3264] SDP message generated by the participant in response to an Offer. An Answer describes participants capabilities in continuing with the multimedia session with in the constraints of the Offer.
This draft avoids using terms that implementors do not have a clear idea of exactly what they are - for example RTP Session.
The requirements listed here are a collection of requirements that have come from WebRTC, CLUE, and the general community that uses RTP for interactive communications based on Offer/Answer. It does not try to meet the needs of streaming usages or usages involving multicast. This list does not also try to list every possible requirement but instead outlines the ones that might influence the design.
This section outlines a set of rules for the usage of SDP and RTP that seems to deal with the various problems and issues that have been discussed. Most of these are not new and are pretty much how many systems do it today. Some of them are new, but all the items requiring new standardization work are called out in the Section 6.
Approach:
The port number that RTP is received on provides the primary mechanism for correlating it to the correct m-line. However, when the port does not uniquely male the RTP packet to the correct m-block (such as in multiplexing and other cases), the next thing that can be looked at is the PT number. Finally there are cases where SSRC can be used if that was signaled.
There are some complications when using SSRC for correlation with signaling. First, the offerer may end up receiving RTP packets before receiving the signaling with the SSRC correlation information. This is because the sender of the RTP chooses the SSRC; there is no way for the receiver to signal how some of the bits in the SSRC should be set. Numerous attempts to provide a way to do this have been made, but they have all been rejected for various reasons, so this situation is unlikely to change. The second issue is that the signaled SSRC can change, particularly in collision cases, and there is no good way to know when SSRC are changing, such that the currently signaled SSRC usage maps to the actual RTP SSRC usage. Finally SSRC does not always provide correlation information between media flows - take for example trying to look at SSRC to tell that an audio media flow and video media flow came from the same camera. The nice thing about SSRC is that they are also included in the RTP.
The proposal here is to extend the MSID draft to meet these needs: each media flow would have a unique MSID and the MSID would have some level of internal structure, which would allow various forms of correlation, including what WebRTC needs to be able to recreate the MS-Stream / MS-Track hierarchy to be the same on both sides. In addition, this work proposes creating an optional RTP header extension that could be used to carry the MSID for a media flow in the RTP packets. This is not absolutely needed for the WebRTC use cases but it helps in the case where media arrives before signaling and it helps resolve a broader category of web conferencing use cases.
The MSID consists of three things and can extended to have more. It has a device identifier, which corresponds to a unique identifier of the device that created the offer; one or more synchronization context identifiers, which is a number that helps correlate different synchronized media flows; and a media flow identifier. The synchronization identifier and flow identifier are scoped within the context of the device identifier, but the device identifier is globally unique. The suggested device identifier is a 64-bit random number. The synchronization group is an integer that is the same for all media flows that have this device identifier and are meant to be synchronized. Right now there can be more than one synchronization identifier, but the open issues suggest that one would be preferable. The flow identifier is an integer that uniquely identifies this media flow within the context of the device identifier.
An example MSID for a device identifier of 12345123451234512345, synchronization group of 1, and a media flow id of 3 would be:
When the MSID is used in an answer, the MSID also has the remote device identifier included. In the case where the device ID of the device sending the answer was 22222333334444455555, the MSID would look like:
Note: The 64 bit size for the device identifier was chosen as it allows less than a one in a million chance of collision with greater than 10,000 flows (actually it allows this probability with more like 6 million flows). Much smaller numbers could be used but 32 bits is probably too small. More discussion on the size of this and the color of the bike shed is needed.
When used in the WebRTC context, each PeerConnection should generate a unique device identifier. Each PC-Stream in the PeerConnection will get a a unique synchronization group identifier, and each PC-Track in the Peer Connection will get a unique flow identifier. Together these will be used to form the MSID. The MSID MUST be included in the SDP offer or answer so that the WebRTC connection on the remote side can form the correct structure of remote PC-Streams and PC-Tracks. If a WebRTC client receives an Offer with no MSID information and no LS group information, it MUST put all the remote PC-Tracks into a single PC-Stream. If there is LS group information but no MSID, a PC-Stream for each LS group MUST be created and the PC-Tracks put in the appropriate PC-Stream.
The W3C specs should be updated to have the ID attribute of the MS-Stream be the MSID with no flow identifier, and the ID attribute of the MS-Track be the MSID.
There are cases - such as a grid of security cameras or thumbnails in a video conference - where a receiver is willing to receive and display several media flows of video. The proposal here is to create a new media level attribute called multiple-render that includes an integer that indicates how many streams can be rendered at the same time.
As an example, a system that could display 16 thumbnails at the same time and was willing to receive H261 or H264 might offer
SDP Offer
m=video 52886 RTP/AVP 98 99
a=muliple-render:16
a=rtpmap:98 H261/90000
a=rtpmap:99 H264/90000
a=fmtp:99 profile-level-id=4de00a;
packetization-mode=0; mst-mode=NI-T;
sprop-parameter-sets={sps0},{pps0};
When combining this muliple-render feature with multiplexing, the answer will might not know all the SSRC that will send to this m-block so it is best to use payload type (PT) numbers that are unique for the SDP: the demultiplexing may have to only use the PT if the SSRC are unknown.
The receiver displays, in different windows, the video from the most recent 16 SSRC to send video to m-block.
This allows a switching MCU to know how many thumbnail type streams would be appropriate to send to this endpoint.
If SDP offer/answers are of type AVP or AVPF but contain a crypto of fingerprint attribute, they should be treated as if they were SAVP or SAVPF respectively. The Answer should have the same type as the offer but for all practical purposes the implementation should treat it as the secure variant.
If SDP offer/answers are of type AVP or SAVP, but contain a rtcp attribute, they should be treated as if they were AVPF or SAVPF respectively. The SDP Answer should have the same type as the Offer but for all practical purposes the implementation should treat it as the feedback variant.
If an SDP Offer has both a fingerprint and a crypto attribute, it means the Offerer supports both DTLS-SRTP and SDES and the answer should select one and return an Answer with only an attribute for the selected keying mechanism.
Example of a video client joining a video conference. The client can produce and receive two streams of video, one from the slides and the other of the person. The video of the person is synchronized with the audio. In addition, the client can display up to 10 thumbnails of video. The main video is simulcast at HD size and a thumbnail size.
SDP Offer - Client send simulcast video with 2 resolutions in 2 m-blocks indicated by a=group:simulcast and indicating lip-sync for the audio and video m-blocks. Also indicating it can accept 10 streams for rendering with a=multi-render
[TODO - populate proper fmtp values for thumbnail size] or
v=0
o=alice 2890844526 2890844527 IN IP4 host.atlanta.example.com
s=
c=IN IP4 host.atlanta.example.com
t=0 0
a=group:LS 1,2,3
a=group:simulcast 2,3
m=audio 49170 RTP/AVP 99
a=mid:1
a=rtpmap:99 iLBC/8000
m=video 51372 RTP/AVP 96
a=mid:2
a=rtpmap:96 H264/90000
a=fmtp:96 profile-level-id=428014;
max-fs=3600; max-mbps=108000; max-br=14000
m=video 51372 RTP/AVP 97
a=mid:3
a=multi-render:10
a=rtpmap:97 H264/90000
a=fmtp:97 profile-level-id=428014; max-fs=3600l
max-mbps=108000; max-br=14000
SDP Answer from the server indicating two video stream with the speaker and the slides. Also signaled is the lip-sync for speakers audio and video streams.
v=0
o=bob 2808844564 2808844565 IN IP4 host.biloxi.example.com
s=
c=IN IP4 host.biloxi.example.com
t=0 0
a=group:LS a,b
m=audio 49172 RTP/AVP 99
a=mid:a
a=rtpmap:99 iLBC/8000
m=video 51374 RTP/AVP 96
a=mid:b
a=content:speaker
a=rtpmap:96 H264/90000
m=video 51376 RTP/AVP 97
a=mid:c
a=content:slides
a=rtpmap:97 H264/90000
Example of a three-screen video endpoint connecting to a two-screen system which ends up selecting the left and middle screens.
SDP Offer
v=0
o=alice 2890844526 2890844527 IN IP4 host.atlanta.example.com
s=
c=IN IP4 host.atlanta.example.com
t=0 0
a=rtcp-fb
m=audio 49170 RTP/SAVPF 99
a=content:left
a=rtpmap:99 iLBC/8000
m=video 51372 RTP/SAVPF 31
a=content:main
a=rtpmap:96 H261/90000
m=video 51374 RTP/SAVPF 31
a=content:right
a=rtpmap:96 H261/90000
SDP Answer
v=0
o=bob 2808844564 2808844565 IN IP4 host.biloxi.example.com
s=
c=IN IP4 host.biloxi.example.com
t= 0 0
a=rtcp-fb
m=audio 49170 RTP/SAVPF 99
a=content:left
a=rtpmap:99 iLBC/8000
m=video 51372 RTP/SAVPF 31
a=content:main
a=rtpmap:96 H261/90000
m=video 0 RTP/SAVPF 31
a=content:right
a=rtpmap:96 H261/90000
Example of a client that supports SRTP-DTLS and SDES connecting to a client that supports SRTP-DTLS.
SDP Offer - with support for SRTP-DTLS and SDES signaled
[TODO - populate proper fmtp values for thumbnail size]
v=0
o=alice 2890844526 2890844527 IN IP4 host.atlanta.example.com
s=
c=IN IP4 host.atlanta.example.com
t=0 0
m=audio 49170 RTP/AVP 99
a=fingerprint:sha-1 99:41:49:83:4a:97:0e:1f:ef:6d
:f7:c9:c7:70:9d:1f:66:79:a8:07
a=crypto:1 AES_CM_128_HMAC_SHA1_80
inline:d0RmdmcmVCspeEc3QGZiNWpVLFJhQX1cfHAwJSoj|2^20|1:32
a=rtpmap:99 iLBC/8000
m=video 51372 RTP/AVP 31
a=fingerprint:sha-1 92:81:49:83:4a:23:0a:0f:1f:9d:f7:
c0:c7:70:9d:1f:66:79:a8:07
a=crypto:1 AES_CM_128_HMAC_SHA1_32
inline:NzB4d1BINUAvLEw6UzF3WSJ+PSdFcGdUJShpX1Zj|2^20|1:32
a=rtpmap:96 H261/90000
SDP Answer signaling only SDES
v=0
o=bob 2808844564 2808844565 IN IP4 host.biloxi.example.com
s=
c=IN IP4 host.biloxi.example.com
t=0 0
m=audio 49172 RTP/AVP 99
a=crypto:1 AES_CM_128_HMAC_SHA1_80
inline:d0RmdmcmVCspeEc3QGZiNWpVLFJhQX1cfHAwJSoj|2^20|1:32
a=rtpmap:99 iLBC/8000
m=video 51374 RTP/AVP 31
a=crypto:1 AES_CM_128_HMAC_SHA1_80
inline:d0RmdmcmVCspeEc3QGZiNWpVLFJhQX1cfHAwJSoj|2^20|1:32
a=rtpmap:96 H261/90000
This section outlines work that needs to be done in various specifications to make the proposal here actually happen.
Tasks:
TBD
This document requires no actions from IANA.
I would like to thank Suhas Nandakumar, Eric Rescorla, and Lyndsay Campbell for help with this draft.
The overall solution is complicated considerably by the fact that WebRTC allows a PC-Track to be used in more than one PC-Stream but requires only one copy of the RTP data for the track to be sent. I am not aware of any use case for this and think it should be removed. If a PC-Track needs to be synchronized with two different things, they should all go in one PC-Stream instead of two.
The following shows some examples of SDP today that any new system needs to be able to receive and work with in a backwards compatible way.
Multiple codecs accepted on same m-line.
SDP Offer
v=0
o=alice 2890844526 2890844527 IN IP4 host.atlanta.example.com
s=
c=IN IP4 host.atlanta.example.com
t=0 0
m=audio 49170 RTP/AVP 99
a=rtpmap:99 iLBC/8000
m=video 51372 RTP/AVP 31 32
a=rtpmap:31 H261/90000
a=rtpmap:32 MPV/90000
SDP Answer
v=0
o=bob 2808844564 2808844565 IN IP4 host.biloxi.example.com
s=
c=IN IP4 host.biloxi.example.com
t=0 0
m=audio 49172 RTP/AVP 99
a=rtpmap:99 iLBC/8000
m=video 51374 RTP/AVP 31 32
a=rtpmap:31 H261/90000
a=rtpmap:32 MPV/90000
This means that a sender can switch back and forth between H261 and MVP without any further signaling. The receiver MUST be capable of receiving both formats. At any point in time, only one video format is sent, thus implying that only one video is meant to be displayed.
Multiple m-blocks identified with respective "mid" grouped to indicate FEC operation.
SDP Offer
v=0
o=ali 1122334455 1122334466 IN IP4 fec.example.com
s=Raptor RTP FEC Example
t=0 0
a=group:FEC-FR S1 R1
m=video 30000 RTP/AVP 100
c=IN IP4 233.252.0.1/127
a=rtpmap:100 MP2T/90000
a=fec-source-flow: id=0
a=mid:S1
m=application 30000 RTP/AVP 110
c=IN IP4 233.252.0.2/127
a=rtpmap:110 raptorfec/90000
a=fmtp:110 raptor-scheme-id=1; Kmax=8192; T=128;
P=A; repair-window=200000
a=mid:R1
This example shows a single codec,say H.264, signaled with different settings.
SDP Offer
v=0
m=video 49170 RTP/AVP 100 99 98
a=rtpmap:98 H264/90000
a=fmtp:98 profile-level-id=42A01E; packetization-mode=0;
sprop-parameter-sets=Z0IACpZTBYmI,aMljiA==
a=rtpmap:99 H264/90000
a=fmtp:99 profile-level-id=42A01E; packetization-mode=1;
sprop-parameter-sets=Z0IACpZTBYmI,aMljiA==
a=rtpmap:100 H264/90000
a=fmtp:100 profile-level-id=42A01E; packetization-mode=2;
sprop-parameter-sets=Z0IACpZTBYmI,aMljiA==;
sprop-interleaving-depth=45; sprop-deint-buf-req=64000;
sprop-init-buf-time=102478; deint-buf-cap=128000
The SDP below shows various ways to specify resolutions for video codecs signaled.
SDP Offer
m=video 49170/2 RTP/AVP 31
a=rtpmap:31 H261/90000
a=fmtp:31 CIF=2;QCIF=1;D=1
m=video 49170/2 RTP/AVP 98 99
a=rtpmap:98 jpeg2000/27000000
a=rtpmap:99 jpeg2000/90000
a=fmtp:98 sampling=YCbCr-4:2:0;width=128;height=128
a=fmtp:99 sampling=YCbCr-4:2:0;width=128;height=128
[RFC4588] retransmission flow example.
SDP Offer
v=0
o=mascha 2980675221 2980675778 IN IP4 host.example.net
c=IN IP4 192.0.2.0
a=group:FID 1 2
a=group:FID 3 4
m=audio 49170 RTP/AVPF 96
a=rtpmap:96 AMR/8000
a=fmtp:96 octet-align=1
a=rtcp-fb:96 nack
a=mid:1
m=audio 49172 RTP/AVPF 97
a=rtpmap:97 rtx/8000
a=fmtp:97 apt=96;rtx-time=3000
a=mid:2
m=video 49174 RTP/AVPF 98
a=rtpmap:98 MP4V-ES/90000
a=rtcp-fb:98 nack
a=fmtp:98 profile-level-id=8;config=01010000012000884006682C209\
0A21F
a=mid:3
m=video 49176 RTP/AVPF 99
a=rtpmap:99 rtx/90000
a=fmtp:99 apt=98;rtx-time=3000
a=mid:4
[RFC5888] grouping semantics for Lip Synchronization between audio and video
SDP Offer
v=0
o=Laura 289083124 289083124 IN IP4 one.example.com
c=IN IP4 192.0.2.1
t=0 0
a=group:LS 1 2
m=audio 30000 RTP/AVP 0
a=mid:1
m=video 30002 RTP/AVP 31
a=mid:2
SDP Offer
m=application 50000 TCP/TLS/BFCP *
a=setup:passive
a=connection:new
a=fingerprint:SHA-1 \
4A:AD:B9:B1:3F:82:18:3B:54:02:12:DF:3E:5D:49:6B:19:E5:7C:AB
a=floorctrl:s-only
a=confid:4321
a=userid:1234
a=floorid:1 m-stream:10
a=floorid:2 m-stream:11
m=audio 50002 RTP/AVP 0
a=label:10
m=video 50004 RTP/AVP 31
a=label:11
Thought not yet defined, it's easy to imaging that BFCP over SCTP over DTLS might look like
m=application 50000 TCP/TLS/BFCP *
a=setup:passive
a=connection:new
a=fingerprint:SHA-1 \
4A:AD:B9:B1:3F:82:18:3B:54:02:12:DF:3E:5D:49:6B:19:E5:7C:AB
a=floorctrl:s-only
a=confid:4321
a=userid:1234
a=floorid:1 m-stream:10
a=floorid:2 m-stream:11
m=audio 50002 RTP/AVP 0
a=label:10
m=video 50004 RTP/AVP 31
a=label:11
[RFC5583] "depend" attribute is shown here to indicate dependency between layers represented by the individual m-blocks
SDP Offer
a=group:DDP L1 L2 L3
m=video 20000 RTP/AVP 96 97 98
a=rtpmap:96 H264/90000
a=fmtp:96 profile-level-id=4de00a; packetization-mode=0;
mst-mode=NI-T; sprop-parameter-sets={sps0},{pps0};
a=rtpmap:97 H264/90000
a=fmtp:97 profile-level-id=4de00a; packetization-mode=1;
mst-mode=NI-TC; sprop-parameter-sets={sps0},{pps0};
a=rtpmap:98 H264/90000
a=fmtp:98 profile-level-id=4de00a; packetization-mode=2;
mst-mode=I-C; init-buf-time=156320;
sprop-parameter-sets={sps0},{pps0};
a=mid:L1
m=video 20002 RTP/AVP 99 100
a=rtpmap:99 H264-SVC/90000
a=fmtp:99 profile-level-id=53000c; packetization-mode=1;
mst-mode=NI-T; sprop-parameter-sets={sps1},{pps1};
a=rtpmap:100 H264-SVC/90000
a=fmtp:100 profile-level-id=53000c; packetization-mode=2;
mst-mode=I-C; sprop-parameter-sets={sps1},{pps1};
a=mid:L2
a=depend:99 lay L1:96,97; 100 lay L1:98
m=video 20004 RTP/AVP 101
a=rtpmap:101 H264-SVC/90000
a=fmtp:101 profile-level-id=53001F; packetization-mode=1;
mst-mode=NI-T; sprop-parameter-sets={sps2},{pps2};
a=mid:L3
a=depend:101 lay L1:96,97 L2:99
SDP Offer
m=video 49170 RTP/AVP 96
a=rtpmap:96 H264/90000
a=ssrc:12345 cname:user@example.com
a=ssrc:67890 cname:user@example.com
This indicates what the sender will send. It's at best a guess because in the case of SSRC collision, it's all wrong. It does not allow one to reject a stream. It does not mean that both streams are displayed at the same time.
[RFC4796] "content" attribute is used to specify the semantics of content represented by the video streams.
SDP Offer
v=0
o=Alice 292742730 29277831 IN IP4 131.163.72.4
s=Second lecture from information technology
c=IN IP4 131.164.74.2
t=0 0
m=video 52886 RTP/AVP 31
a=rtpmap:31 H261/9000
a=content:slides
m=video 53334 RTP/AVP 31
a=rtpmap:31 H261/9000
a=content:speaker
m=video 54132 RTP/AVP 31
a=rtpmap:31 H261/9000
a=content:sl
| [RFC2119] | Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. |
| [RFC3264] | Rosenberg, J. and H. Schulzrinne, "An Offer/Answer Model with Session Description Protocol (SDP)", RFC 3264, June 2002. |
| [RFC4566] | Handley, M., Jacobson, V. and C. Perkins, "SDP: Session Description Protocol", RFC 4566, July 2006. |