network working group h. alvestrand internet-draft google ...why rtp sessions should be content...

Network Working Group H. AlvestrandInternet-Draft GoogleIntended status: Informational December 9, 2011Expires: June 11, 2012

Why RTP Sessions Should Be Content Neutral draft-alvestrand-rtp-sess-neutral-00

Abstract

This document is not intended for publication as an RFC.

It gives the underpinning arguments for why the idea that RTP sessions and MIME top level types are related is a deeply broken paradigm, and that we need to get away from it.

These arguments are solely the opinion of the listed author.

Requirements Language

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119 [RFC2119].

Status of this Memo

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

Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet- Drafts is at 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 June 11, 2012.

Copyright Notice

Copyright (c) 2011 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

Alvestrand Expires June 11, 2012 []

Internet-Draft RTP session neutrality December 2011

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.

Table of Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 2. Stuff that can be carried over RTP . . . . . . . . . . . . . . 3 3. What the network can do to "help" a flow . . . . . . . . . . . 4 4. The definition of an RTP "session" . . . . . . . . . . . . . . 5 5. Proper and improper use of RTP sessions . . . . . . . . . . . 6 6. The Pernicious Effect of SDP on the Media Type System . . . . 8 7. Corrective Actions . . . . . . . . . . . . . . . . . . . . . . 8 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 9 9. Security Considerations . . . . . . . . . . . . . . . . . . . 9 10. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 9 11. References . . . . . . . . . . . . . . . . . . . . . . . . . . 9 11.1. Normative References . . . . . . . . . . . . . . . . . . 9 11.2. Informative References . . . . . . . . . . . . . . . . . 9 Author’s Address . . . . . . . . . . . . . . . . . . . . . . . . . 10



1. Introduction

The RTP universe of functionality can, for the purposes of this argument, be reduced to two components: The RTP wire protocol (consisting of the RTP packet format, the RTCP reporting format, and the handling rules for RTP sessions), and the SDP session description language. For the purposes of this argument, the SDP functionality for describing non-RTP sessions is ignored, as is the ability to negotiate RTP sessions by other means than SDP.

This document argues that the RTP mechanisms of multiple RTP sessions make sense for a lot of purposes, but does NOT make sense for a mandated separation between different top-level MIME media types.

2. Stuff that can be carried over RTP

RTP, according to its own description an application layer framework component, is a suitable protocol for framing data that needs to travel across the network in a time-sensitive fashion, with the idea that it is going to be presented at the receiving end in a time sequence. Normally, the data (usually called "media") is streamed across the network at a rate approximately equal to the speed at which it is intended to be presented ("real time data").

Examples of data carried over RTP include:

o G.711 Audio - 64 Kbits/second, completely fixed bitrate

o GSM AMR Audio - 4.75 to 12.2 Kbits/second, variable bitrate

o OPUS audio compressed into near-incomprehensibility - 6 kbits/ second, variable bitrate

o OPUS audio carrying high fidelity music - 500 kbits/second, variable bitrate

o QQVGA (160x120) video at 15 FPS in H.264 compression - 50 Kbits/ second, variable bitrate, lots of schemes for error concealment and correction

o HD video at 1920x1080@60 in H.264 compression - 1.4 Mbits/second

o Real-time text (T.140) - very few bits/second

o RFC 4733 DTMF tone signalling - very few bits/second

Schemes designed to increase the reliability of data carried across



RTP include:

o Forward error correction (FEC)

o Duplicated streams, codec-independent

o Duplicate sending of important information within the codec (RFC 4733, for instance)

o NAK-based resends signalled over RTCP

o Stream reset requests signalled over RTCP

Some of these are only applicable to media types (in particular, "send me a new I-frame" doesn’t make sense if you don’t have I-frames). Others can be used with any type of data.

3. What the network can do to "help" a flow

The network can apply various things to help the session data arrive according to policy:

o Capacity reservation for specific flows

o Priority queueing, sending certain types of data faster than others

o Filtering or blocking certain types of communication that the managers deem inappropriate

The network can do these things in multiple ways, including so-called "deep packet inspection", but the most common techniques require being able to identify either the requested handling of the packets (DiffServ using DSCP codepoints) or recognizing the flow based on its 5-tuple (source and destination address and port + protocol), possibly correlating the 5-tuple with information carried to the router through some kind of management interface (either connected to the session setup protocol or managed via some other interface such as RSVP/IntServ), and behaving accordingly.

All techniques have limitations; DSCP requires a certain trust in the endpoints using the codepoints for "deserving traffic"; deep packet inspection requires that packets be unencrypted, and stream control requires that 5-tuples be related back to their putative purpose either by heuristics or by being connected to management protocols.



4. The definition of an RTP "session"

An RTP session is defined in RFC 3550 section 3:

"RTP Session: An association among a set of participants communicating with RTP. A participant may be involved in multiple RTP sessions at the same time. In a multimedia session, each medium is typically carried in a separate RTP session with its own RTCP packets unless the encoding itself multiplexes multiple media into a single data stream. A participant distinguishes multiple RTP sessions by reception of different sessions using different pairs of destination transport addresses, where a pair of transport addresses comprises one network address plus a pair of ports for RTP and RTCP. All participants in an RTP session may share a common destination transport address pair, as in the case of IP multicast, or the pairs may be different for each participant, as in the case of individual unicast network addresses and port pairs. In the unicast case, a participant may receive from all other participants in the session using the same pair of ports, or may use a distinct pair of ports for each.

The distinguishing feature of an RTP session is that each maintains a full, separate space of SSRC identifiers (defined next). The set of participants included in one RTP session consists of those that can receive an SSRC identifier transmitted by any one of the participants either in RTP as the SSRC or a CSRC (also defined below) or in RTCP. For example, consider a three- party conference implemented using unicast UDP with each participant receiving from the other two on separate port pairs. If each participant sends RTCP feedback about data received from one other participant only back to that participant, then the conference is composed of three separate point- to-point RTP sessions. If each participant provides RTCP feedback about its reception of one other participant to both of the other participants, then the conference is composed of one multi-party RTP session. The latter case simulates the behavior that would occur with IP multicast communication among the three participants.

The RTP framework allows the variations defined here, but a particular control protocol or application design will usually impose constraints on these variations."

An RTP session is thus characterized by:

o A single SSRC space

o A single reporting space - all participants see all RTCP messages



o Non overlapping transport addresses

As we can see here, it is not possible to tell from a single packet whether it belongs to the same session as another packet or not; if we observe two packets with the same source and destination addresses, it seems safe to assume that they belong to the same session, but for all other cases, deciding whether or not two packets or packet streams are in the same session requires knowledge of the configuration of the session.

5. Proper and improper use of RTP sessions

Section 5.2 of RFC 3550 gives the canonical statement of RTP session (mis)use:

"In RTP, multiplexing is provided by the destination transport address (network address and port number) which is different for each RTP session. For example, in a teleconference composed of audio and video media encoded separately, each medium SHOULD be carried in a separate RTP session with its own destination transport address."

This sentence makes two very important leaps of faith:

o That distinguishing sessions by destination transport address is necessary and sufficient

o That it is appropriate to give strong guidance about the distribution of media streams across RTP sessions

Both of these are shaky.

As the cost of connecting ports has increased due to NATs, firewalls and IPv4 exhaustion, there has been a strong push towards using fewer ports, and indeed fewer 5-tuples, so that it is not uncommon to see flows that can be distinguished only by source address; there have also been proposals floated for putting multiple RTP sessions across one 5-tuple [draft-westerlund-avtcore-transport-multiplexing].

The cost of ports is also one factor pushing towards multiple media types in one RTP session; however, the more important underlying challenge is that this distinction is neither necessary nor sufficient to distinguish the cases in which RTP media streams want to have differential treatment from the network, and thus need to assign streams either to the same session (to guarantee the same treatment) or to different sessions (to allow for differential treatment).



Consider the list of scenarios above, and imagine RTP being used for:

o A videoconference between 3 people who know each other well, using low end equipment and barely-sufficient bandwidth pipes

o A Berlin Philharmonic concert broadcast featuring Brahms’ "Tragic Overture"

o A point-to-point transmission of a Manchester United vs Liverpool football match

o A professor’s lecture, with "talking head" presentation simultaneous with slides, and opportunity for students to ask questions

In each of these contexts, the tradeoff between audio and video is different; in the Brahms case, the audio (which is the point of the transmission) is likely to be transmitted at higher bandwidth than the video, and if one of them has to have his bandwidth reduced, the video should be reduced in quality before the audio is. In contrast, in the football match, spectators care about seeing the action; as long as they can understand the commentator’s voice, the audio quality is "good enough".

In the lecture case, quality of the lecturer’s slides and voice is critical; video from students is almost irrelevant to the larger purpose.

A logical arrangement of media streams in RTP sessions would be to group them by importance, and send them with appropriate traffic engineering structuring; in the lecturer case, the slides and the professor’s voice would be carried in a high priority media stream, while the professor’s picture would have second priority, and voice and video from students would be made available on an "if it works, it works" basis. Someone may easily decide that the student feedback track is not worth listening to, or remove the talking head of the professor; it would be strange indeed to try to listen to the lecture without viewing the supporting material.

This illustrates two points:

o The RTP session mechanism, using the 5-tuple as the unit of differentiation, is a simple, effective and readily deployed mechanism for separating streams that require different treatment from the network in easily distinguished partitions.

o The assignment of media to such partitions is application dependent, and the decision on how to group and how to prioritize



needs to be taken by the application developer.

6. The Pernicious Effect of SDP on the Media Type System

In the list of reasons to argue against the inappropriate advice quoted above from RFC 3550, its pernicious influence on the MIME type system bears mentioning.

The MIME type system, as described in RFC 4288, consists of a two- level hierarchy: A top level media type (text, audio, video, application and so on), and a media subtype that identifies (to some level of precision) the format of the data being carried.

The system has mostly been respected, with some types (for instance PDF) forever being borderline between the various categories, but over the years, a few types have been entered into the system with their top level types being decided, not by the nature of their content, but by *the type with which their proponents wished to have them multiplexed in an RTP session*.

This includes the types that designate repair mechanism (audio/ parityfec, audio/red), timed data transfer (audio/clearmode) and that ultimate triumph of expediency over cleanliness: audio/t140c, audio/ 3gpp-tt and video/3gpp-tt: Text types registered as audio and video.

For each of these, there is a fairly natural fit in the normal MIME hierarchy (application/ for the mechanism types and text/ for the text types); the assignment of them to the "media" top level types has been done as an expediency in order to get around the stultifying results of the advice given in RFC 3550.

7. Corrective Actions

There are not many protocol changes that really need to be taken to solve this problem.

The basic mechanism of RTP is media type independent. There are some RTCP issues with dealing with RTP flows of wildly varying bandwidth, but as can be seen from the table of media types in the introduction, this issue isn’t solved by separating them; the bandwidth ranges of the types overlap.

The thing that binds most in the current protocol suite is the conservation of the inappropriate binding in the SDP media description/negotiation format, where the MIME type is represented in two pieces, one of which is tied to the RTP session rather than to



the payload type it is associated with, and there are fairly well- understood ways to get around that, such as the BUNDLE grouping extension (draft-holmberg-mmusic-sdp-bundle-negotiation). Better designed negotiation protocols would not have this problem at all.

In order to get out of the bind that SDP places us in, a change such as BUNDLE should be adopted, and the IETF should record that the advice from RFC 3550 is to be considered *advice*, not command: It is sometimes appropriate to separate media streams according to top level type, and sometimes not appropriate to do so. The application is the one that needs to make this decision.

8. IANA Considerations

This document makes no request of IANA.

Note to RFC Editor: this section may be removed on publication as an RFC.

9. Security Considerations

This note does not discuss any change that the author thinks would have any significant influence on the security of RTP traffic.

10. Acknowledgements

This note has benefited greatly from exchanges with Colin Perkins, whose unwavering support of a sharply differing viewpoint has served to inform the arguments presented in this document. Magnus Westerlund and Christer Holmberg also deserve special mention for engaging constructively in the discussion.

11. References

11.1. Normative References

[RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997.

11.2. Informative References

[I-D.holmberg-mmusic-sdp-bundle-negotiation] Holmberg, C. and H. Alvestrand, "Multiplexing Negotiation Using Session Description Protocol (SDP) Port Numbers",



draft-holmberg-mmusic-sdp-bundle-negotiation-00 (work in progress), October 2011.

[I-D.westerlund-avtcore-transport-multiplexing] Westerlund, M. and C. Perkins, "Multiple RTP Session on a Single Lower-Layer Transport", draft-westerlund-avtcore-transport-multiplexing-01 (work in progress), October 2011.

[RFC3550] Schulzrinne, H., Casner, S., Frederick, R., and V. Jacobson, "RTP: A Transport Protocol for Real-Time Applications", STD 64, RFC 3550, July 2003.

[RFC4288] Freed, N. and J. Klensin, "Media Type Specifications and Registration Procedures", BCP 13, RFC 4288, December 2005.

[RFC4733] Schulzrinne, H. and T. Taylor, "RTP Payload for DTMF Digits, Telephony Tones, and Telephony Signals", RFC 4733, December 2006.

Author’s Address

Harald T. Alvestrand Google Kungsbron 2 Stockholm, 11122 Sweden

Email: [email protected]


AVT A. BegenInternet-Draft CiscoIntended status: Standards Track C. PerkinsExpires: September 11, 2012 University of Glasgow March 10, 2012

Duplicating RTP Streams draft-begen-avtcore-rtp-duplication-01

Abstract

Packet loss is undesirable for real-time multimedia sessions, but can occur due to congestion, or other unplanned network outages. This is especially true for IP multicast networks, where packet loss patterns can vary greatly between receivers. One technique that can be used to recover from packet loss without incurring unbounded delay for all the receivers is to duplicate the packets and send them in separate redundant streams. This document explains how Real-time Transport Protocol (RTP) streams can be duplicated without breaking RTP media streams, or RTP Control Protocol (RTCP) rules.

Status of this Memo




This Internet-Draft will expire on September 11, 2012.

Copyright Notice


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

Begen & Perkins Expires September 11, 2012 [Page 1]

Internet-Draft RTP Duplication March 2012

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.

Table of Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 2. Terminology and Requirements Notation . . . . . . . . . . . . 3 3. Dual Streaming Use Cases . . . . . . . . . . . . . . . . . . . 3 3.1. Temporal Redundancy . . . . . . . . . . . . . . . . . . . 4 3.2. Spatial Redundancy . . . . . . . . . . . . . . . . . . . . 4 3.3. Dual Streaming over a Single Path or Multiple Paths . . . 5 4. Use of RTP and RTCP with Temporal Redundancy . . . . . . . . . 6 4.1. RTCP Considerations . . . . . . . . . . . . . . . . . . . 6 4.2. Signaling Considerations . . . . . . . . . . . . . . . . . 6 5. Use of RTP and RTCP with Spatial Redundancy . . . . . . . . . 7 5.1. RTCP Considerations . . . . . . . . . . . . . . . . . . . 8 5.2. Signaling Considerations . . . . . . . . . . . . . . . . . 8 6. Use of RTP and RTCP with Temporal and Spatial Redundancy . . . 9 7. Security Considerations . . . . . . . . . . . . . . . . . . . 9 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 9 9. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 10 10. References . . . . . . . . . . . . . . . . . . . . . . . . . . 10 10.1. Normative References . . . . . . . . . . . . . . . . . . . 10 10.2. Informative References . . . . . . . . . . . . . . . . . . 10 Authors’ Addresses . . . . . . . . . . . . . . . . . . . . . . . . 11



1. Introduction

The Real-time Transport Protocol (RTP) [RFC3550] is widely used today for delivering IPTV traffic, and other real-time multimedia sessions. Many of these applications support very large numbers of receivers, and rely on intra-domain UDP/IP multicast for efficient distribution of traffic within the network.

While this combination has proved successful, there does exist a weakness. As [RFC2354] noted, packet loss is not avoidable, even in a carefully managed network. This loss might be due to congestion, it might also be a result of an unplanned outage caused by a flapping link, link or interface failure, a software bug, or a maintenance person accidentally cutting the wrong fiber. Since UDP/IP flows do not provide any means for detecting loss and retransmitting packets, it leaves up to the RTP layer and the applications to detect, and recover from, packet loss.

One technique to recover from packet loss without incurring unbounded delay for all the receivers is to duplicate the packets and send them in separate redundant streams. Variations on this idea have been implemented and deployed today [IC2011]. However, duplication of RTP streams without breaking the RTP and RTCP functionality has not been documented properly. This document explains how duplication can be achieved for RTP streams.

Stream duplication offers a simple way to protect media flows from packet loss. It has a comparatively high bandwidth overhead, since everything is sent twice, but with a low processor overhead. It is also very predictable in its overheads. Alternative approaches may be suitable in some cases, for example retransmission-based recovery [RFC4588] or forward error correction [RFC5109].

2. Terminology and Requirements Notation

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119].

3. Dual Streaming Use Cases

Dual streaming refers to a technique that involves transmitting two redundant RTP streams of the same content, with each stream capable of supporting the playback when there is no packet loss. Therefore, adding an additional RTP stream provides a protection against packet



loss. The level of protection depends on how the packets are sent and transmitted inside the network.

It is important to note that dual streaming can easily be extended to support cases when more than two streams are desired. However, using three or more streams is rare in practise, due to the high overhead that it incurs.

3.1. Temporal Redundancy

From a routing perspective, two streams are considered identical if the following two IP header fields are the same, since they will be both routed over the same path:

o IP Source Address

o IP Destination Address

Two routing-plane identical RTP streams might carry the same payload, but can use different Synchronization Sources (SSRC) to differentiate the RTP packets belonging to each stream. In the context of dual RTP streaming, we assume that the source duplicates the RTP packets and sends them in separate RTP streams, each with a unique SSRC. All the redundant streams are transmitted in the same RTP session.

For example, one main and one redundant RTP stream can be sent to the same IP destination address and UDP destination port with a certain delay between them [I-D.begen-mmusic-temporal-interleaving]. The streams carry the same payload in their respective RTP packets with identical sequence numbers. This allows receivers (or other nodes responsible for gap filling and duplicate suppression) to identify and suppress the duplicate packets, and subsequently produce a hopefully loss-free and duplication-free output stream. This process is called stream merging.

3.2. Spatial Redundancy

An RTP source might be associated with multiple network interfaces, allowing it to send two redundant streams from two separate source addresses. Such streams can be routed over diverse or identical paths depending on the routing algorithm used inside the network. At the receiving end, the node responsible for duplicate suppression can look into various RTP header fields, for example SSRC and sequence number, to identify and suppress the duplicate packets.

If source-specific multicast (SSM) transport is used to carry such redundant streams, there will be a separate SSM session for each redundant stream since the streams are sourced from different



interfaces (i.e., IP addresses). Thus, the receiving host has to join each SSM session separately.

Alternatively, an RTP source might send the redundant streams to separate IP destination addresses.

3.3. Dual Streaming over a Single Path or Multiple Paths

Having described the characteristics of the streams, one can reach the following conclusions:

1. When two routing-plane identical streams are used, the two streams will have identical IP headers. This makes it impractical to forward the packets onto different paths. In order to minimize packet loss, the packets belonging to one stream are often interleaved with packets belonging to the other, and with a delay, so that if there is a packet loss, such a delay would allow the same packet from the other stream to reach the receiver because the chances that the same packet is lost in transit again is often small. This is what is also known as Time-shifted Redundancy, Temporal Redundancy or simply Delayed Duplication [I-D.begen-mmusic-temporal-interleaving] [IC2011]. This approach can be used with both types of dual streaming, described in Section 3.1 and Section 3.2.

2. If the two streams have different IP headers, an additional opportunity arises in that one is able to build a network, with physically diverse paths, to deliver the two streams concurrently to the intended receivers. This reduces the delay when packet loss occurs and needs to be recovered. Additionally, it also further reduces chances for packet loss. An unrecoverable loss happens only when two network failures happen in such a way that the same packet is affected on both paths. This is referred to as Spatial Diversity or Spatial Redundancy [IC2011]. The techniques used to build diverse paths are beyond the scope of this document.

Note that spatial redundancy often offers less delay in recovering from packet loss provided that the forwarding delay of the network paths are more or less the same. For both temporal and spatial redundancy approaches, packet misordering might still happen and needs to be handled using the sequence numbers of some sort (e.g., RTP sequence numbers).

To summarize, dual streaming allows an application and a network to work together to provide a near zero-loss transport with a bounded or minimum delay. The additional advantage includes a predictable bandwidth overhead that is proportional to the minimum bandwidth



needed for the multimedia session, but independent of the number of receivers experiencing a packet loss and requesting a retransmission. For a survey and comparison of similar approaches, refer to [IC2011].

4. Use of RTP and RTCP with Temporal Redundancy

To achieve temporal redundancy, the main and redundant RTP streams MUST be sent using the same 5-tuple of transport protocol, source and destination IP addresses, and source and destination transport ports. This is perhaps overly restrictive, but with the possible presence of network address and port translation (NAPT) devices, using anything other than an identical 5-tuple can also cause spatial redundancy.

Since main and redundant RTP streams follow an identical path, they are part of the same RTP session. Accordingly, the sender MUST choose a different SSRC for the redundant RTP stream than it chose for the main RTP stream, following the rules in [RFC3550] Section 8.

4.1. RTCP Considerations

If RTCP is being sent for the main RTP stream, then the sender MUST also generate RTCP for the redundant RTP stream. The RTCP for the redundant RTP stream is generated exactly as-if the redundant RTP stream were a regular media stream. The sender MUST NOT duplicate the RTCP packets sent for the main RTP stream when sending the duplicate stream, instead it MUST generate new RTCP reports for the duplicate stream. The sender MUST use the same RTCP CNAME in the RTCP reports it sends for the main and redundant streams, so that the receiver can synchronize them.

Both the main and redundant RTP streams, and their corresponding RTCP reports, will be received. If RTCP is used, receivers MUST generate RTCP reports for both main and redundant streams in the usual way, treating them as entirely separate media streams.

4.2. Signaling Considerations

Signaling is needed to allow the receiver to determine that an RTP stream is a redundant copy of another, rather than a separate stream that needs to be rendered in parallel. There are two parts to this: an SDP extension is needed in the offer/answer exchange to negotiate support for temporal redundancy; and signalling is needed to indicate which stream is the duplicate (the latter can be done in-band using an RTCP extension, or out-of-band by signalling the SSRCs used by the duplicate streams in SDP).

We require out-of-band signalling for both features. The required



SDP attribute to signal duplication in the SDP offer/answer exchange (’duplication-delay’) is defined in [I-D.begen-mmusic-temporal-interleaving]. The required SDP grouping semantics are defined in [I-D.begen-mmusic-redundancy-grouping].

In the following SDP example, a video stream is duplicated, and the main and redundant streams are transmitted in two separate SSRCs (1000 and 1010):

v=0 o=ali 1122334455 1122334466 IN IP4 dup.example.com s=Delayed Duplication t=0 0 m=video 30000 RTP/AVP 100 c=IN IP4 233.252.0.1/127 a=source-filter:incl IN IP4 233.252.0.1 198.51.100.1 a=rtpmap:100 MP2T/90000 a=ssrc:1000 cname:[email protected] a=ssrc:1010 cname:[email protected] a=ssrc-group:DUP 1000 1010 a=duplication-delay:100 a=mid:Group1

It is RECOMMENDED that the SSRC listed first in the "a=ssrc-group:" line is sent first, with the other RTP SSRC being the time-delayed duplicate. This is not critical, however, and receivers should size their playout buffers based on the "a=duplication-delay:" attribute, and play the stream that arrives first in preference, with the other stream acting as a repair stream, irrespective of the order in which they are signalled.

5. Use of RTP and RTCP with Spatial Redundancy

When using spatial redundancy, the redundant RTP stream is sent on using a different source and/or destination address/port pair. This will be a separate RTP session to the session conveying the main RTP stream.

The SSRCs used for the main and redundant streams MUST be chosen randomly, following the rules in Section 8 of [RFC3550]. Accordingly, they will almost certainly not match each other. The sender MUST, however, use the same RTCP CNAME for both the main and redundant streams, and MUST include an "a=ssrc:... srcname:..." attribute to correlate the flows. An "a=group:DUP" attribute is used to indicate duplication.



5.1. RTCP Considerations

If RTCP is being sent for the main RTP stream, then the sender MUST also generate RTCP for the redundant RTP stream. The RTCP for the redundant RTP stream is generated exactly as-if the redundant RTP stream were a regular media stream; the sender MUST NOT duplicate the RTCP packets sent for the main RTP stream. The sender MUST use the same RTCP CNAME in the RTCP reports it sends for the main and redundant streams, so that the receiver can synchronize them.

The main and redundant streams are conceptually synchronised using the standard RTCP SR-based mechanism, deriving a mapping between their timelines. The RTP timestamps and sequence numbers SHOULD be identical in the main and redundant streams, however, making the mapping trivial in most cases.

Both main and redundant streams, and their corresponding RTCP, will be received. If RTCP is used, receivers MUST generate RTCP reports for both main and redundant streams in the usual way, treating them as entirely separate media streams.

5.2. Signaling Considerations

The required SDP grouping semantics have been defined in [I-D.begen-mmusic-redundancy-grouping]. In the following example, the redundant streams have different IP destination addresses. The example shows the same UDP port number and IP source addresses, but either or both could have been different for the two streams.



v=0 o=ali 1122334455 1122334466 IN IP4 dup.example.com s=DUP Grouping Semantics t=0 0 a=group:DUP S1a S1b m=video 30000 RTP/AVP 100 c=IN IP4 233.252.0.1/127 a=source-filter:incl IN IP4 233.252.0.1 198.51.100.1 a=rtpmap:100 MP2T/90000 a=ssrc:1000 cname:[email protected] a=ssrc:1000 srcname:45:a8:f4:19:b4:c3 a=mid:S1a m=video 30000 RTP/AVP 101 c=IN IP4 233.252.0.2/127 a=source-filter:incl IN IP4 233.252.0.2 198.51.100.1 a=rtpmap:101 MP2T/90000 a=ssrc:1010 cname:[email protected] a=ssrc:1010 srcname:45:a8:f4:19:b4:c3 a=mid:S1b

6. Use of RTP and RTCP with Temporal and Spatial Redundancy

This uses the same RTP/RTCP mechanisms, plus a combination of both sets of signaling.


The security considerations of [RFC3550], [I-D.begen-mmusic-temporal-interleaving], and [I-D.begen-mmusic-redundancy-grouping] apply.

If stream de-duplication is done by an in-network middlebox, rather than by an end system, that middlebox can work if Secure RTP (SRTP) encryption is used [RFC3711], since the RTP headers are in the clear. Doing so would break the authentication when the SSRC is rewritten, unless the de-duplication middlebox were trusted to re-authenticate the packets. This would require additional signalling which is not specified here, since de-duplication in the receiver end system is expected to be the more common use case.


No IANA actions are required.



9. Acknowledgments

Thanks to Magnus Westerlund for his suggestions.

10. References




[I-D.begen-mmusic-temporal-interleaving] Begen, A., Cai, Y., and H. Ou, "Delayed Duplication Attribute in the Session Description Protocol", draft-begen-mmusic-temporal-interleaving-03 (work in progress), October 2011.

[I-D.begen-mmusic-redundancy-grouping] Begen, A., Cai, Y., and H. Ou, "Duplication Grouping Semantics in the Session Description Protocol", draft-begen-mmusic-redundancy-grouping-02 (work in progress), October 2011.

[RFC3711] Baugher, M., McGrew, D., Naslund, M., Carrara, E., and K. Norrman, "The Secure Real-time Transport Protocol (SRTP)", RFC 3711, March 2004.


[RFC2354] Perkins, C. and O. Hodson, "Options for Repair of Streaming Media", RFC 2354, June 1998.

[RFC4588] Rey, J., Leon, D., Miyazaki, A., Varsa, V., and R. Hakenberg, "RTP Retransmission Payload Format", RFC 4588, July 2006.

[RFC5109] Li, A., "RTP Payload Format for Generic Forward Error Correction", RFC 5109, December 2007.

[IC2011] Evans, J., Begen, A., Greengrass, J., and C. Filsfils, "Toward Lossless Video Transport (to appear in IEEE Internet Computing)", November 2011.



Authors’ Addresses

Ali Begen Cisco 181 Bay Street Toronto, ON M5J 2T3 CANADA


Colin Perkins University of Glasgow School of Computing Science Glasgow, G12 8QQ UK



AVTCore R. van BrandenburgInternet-Draft H. StokkingIntended status: Standards Track O. van DeventerExpires: September 10, 2012 TNO F. Boronat M. Montagud Universitat Politecnica de Valencia K. Gross AVA Networks March 9, 2012

RTCP for inter-destination media synchronization draft-ietf-avtcore-idms-03

Abstract

This document gives information on an RTCP Packet Type and RTCP XR Block Type including associated SDP parameters for Inter-Destination Media Synchronization (IDMS). The RTCP XR Block Type, registered with IANA based on an ETSI specification, is used to collect media playout information from participants in a group playing-out (watching, listening, etc.) a specific RTP media stream. The RTCP packet type specified by this document is used to distribute a common target playout point to which all the distributed receivers, sharing a media experience, can synchronize.

Typical use cases in which IDMS is usefull are social TV, shared service control (i.e. applications where two or more geographically separated users are watching a media stream together), distance learning, networked video walls, networked loudspeakers, etc.

Status of this Memo




van Brandenburg, et al.Expires September 10, 2012 [Page 1]

Internet-Draft RTCP for IDMS March 9, 2012


Copyright Notice


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.



Table of Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.1. Inter-destination Media Synchronization . . . . . . . . . 4 1.2. Applicability of RTCP to IDMS . . . . . . . . . . . . . . 4 1.3. Applicability of SDP to IDMS . . . . . . . . . . . . . . . 5 1.4. This document and ETSI TISPAN . . . . . . . . . . . . . . 5 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 5 3. Overview of IDMS operation . . . . . . . . . . . . . . . . . . 5 4. Inter-destination media synchronization use cases . . . . . . 7 5. Architecture for inter-destination media synchronization . . . 8 5.1. Media Synchronization Application Server (MSAS) . . . . . 8 5.2. Synchronization Client (SC) . . . . . . . . . . . . . . . 9 5.3. Communication between MSAS and SCs . . . . . . . . . . . . 9 6. RTCP XR Block Type for IDMS . . . . . . . . . . . . . . . . . 9 7. RTCP Packet Type for IDMS (IDMS report) . . . . . . . . . . . 11 8. Timing and NTP Considerations . . . . . . . . . . . . . . . . 13 8.1. Leap Seconds . . . . . . . . . . . . . . . . . . . . . . . 14 9. SDP Parameter for RTCP XR IDMS Block Type . . . . . . . . . . 14 10. SDP Parameter for RTCP IDMS Packet Type . . . . . . . . . . . 15 11. Compatibility with ETSI TISPAN . . . . . . . . . . . . . . . . 16 12. On the use of presentation timestamps . . . . . . . . . . . . 16 13. Security Considerations . . . . . . . . . . . . . . . . . . . 17 14. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 17 15. Contributors . . . . . . . . . . . . . . . . . . . . . . . . . 18 16. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . 18 17. References . . . . . . . . . . . . . . . . . . . . . . . . . . 19 17.1. Normative References . . . . . . . . . . . . . . . . . . . 19 17.2. Informative References . . . . . . . . . . . . . . . . . . 19 Authors’ Addresses . . . . . . . . . . . . . . . . . . . . . . . . 20



1. Introduction

1.1. Inter-Destination Media Synchronization

Inter-Destination Media Synchronization (IDMS) refers to the playout of media streams at two or more geographically distributed locations in a time synchronized manner. It can be applied to both unicast and multicast media streams and can be applied to any type and/or combination of streaming media, such as audio, video and text (subtitles). [Ishibashi2006] and [Boronat2009] provide an overview of technologies and algorithms for IDMS.

IDMS requires the exchange of information on media receipt and playout times among participants in an IDMS session. It may also require signaling for the initiation and maintenance of IDMS sessions and groups of receivers.

The presented RTCP specification for IDMS is independent of the used synchronization algorithm, which is out-of-scope of this document.

1.2. Applicability of RTCP to IDMS

Currently, most multimedia applications make use of RTP and RTCP [RFC3550]. RTP provides end-to-end network transport functions suitable for applications requiring real-time data transport, such as audio, video or data, over multicast or unicast network services. The timestamps, sequence numbers, and payload (content) type identification mechanisms provided by RTP packets are very useful for reconstructing the original media timing, and for reordering and detecting packet loss at the client side.

The data transport is augmented by a control protocol (RTCP) to allow monitoring of the data delivery in a manner that is scalable to large multicast networks, and to provide minimal control and identification functionality.

RTP receivers and senders provide reception quality feedback by sending out RTCP Receiver Report (RR) and Sender Report (SR) packets [RFC3550], respectively, which may be augmented by eXtended Reports (XR) [RFC3611]. Thus, the feedback reporting features provided by RTCP make QoS monitoring possible and can be used for troubleshooting and fault tolerance management in multicast distribution services such as IPTV.

These protocols are intended to be tailored through modifications and additions in order to include profile-specific information required by particular applications, and the guidelines on doing so are specified in [RFC5868].



IDMS involves the collection, summarizing and distribution of RTP packet arrival and playout times. As information on RTP packet arrival times and playout times can be considered reception quality feedback information, RTCP is well suited for carrying out IDMS, which may facilitate the implementation and deployment in typical multimedia applications.

1.3. Applicability of SDP to IDMS

RTCP XR [RFC3611] defines the Extended Report (XR) packet type for the RTP Control Protocol (RTCP), and defines how the use of XR packets can be signaled by an application using the Session Description Protocol (SDP) [RFC4566].

SDP signaling is used to set up and maintain a synchronization group between Synchronization Clients (SCs). This document describes two SDP parameters for doing this, one for the RTCP XR block type and one for the new RTCP packet type.

1.4. This document and ETSI TISPAN

ETSI TISPAN [TS183063] has specified architecture and protocol for IDMS using RTCP XR exchange and SDP signaling. For more information on how this document relates to [TS183063], see Section 11.

2. Terminology

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119 [RFC2119] and indicate requirement levels for compliant implementations.

3. Overview of IDMS operation

This section provides a brief example of how the RTCP functionality is used for achieving IDMS. The section is tutorial in nature and does not contain any normative statements.



Alice’s . . . . . . .tv:abc.com . . . . . . . . . Bob’s TV (Sync Client) (Sync Server) Laptop (Sync Client) | | | | Media Session | | |<=====================>| | | Invite(URL,Sync-group ID) | |------------------------------------------------->| | | Media Session Set-up | | |<========================>| | | | | Call set-up | |<================================================>| | | | | RTP Packet | RTP Packet | |<----------------------|------------------------->| | RR + IDMS XR | | |---------------------->| RR + IDMS XR | | |<-------------------------| | RTCP IDMS packet | RTCP IDMS packet | |<----------------------|------------------------->| | | |

Alice is watching TV in her living room. At some point she sees that a football game of Bob’s favorite team is on. She sends him an invite to watch the program together. Embedded in the invitation is the link to the media server and a unique sync-group identifier.

Bob, who is also at home, receives the invite on his laptop. He accepts Alice’s invitation and the RTP client on his laptop sets up a session to the media server. A VoIP connection to Alice’s TV is also set up, so that Alice and Bob can talk while watching the game together.

As is common with RTP, both the RTP client in Alice’s TV as well as the one in Bob’s laptop send periodic RTCP Receiver Reports (RR) to the media server. However, in order to make sure Alice and Bob see the events in the football game at (approximately) the same time, their clients also periodically send an IDMS XR block to the sync server function of the media server. Included in the XR blocks are timestamps on when both Alice and Bob received (or played out) a particular RTP packet.

The sync server function in the media server calculates a reference client from the received IDMS XR blocks (e.g. by selecting whichever client received the packet the latest as the reference client). It then sends an RTCP IDMS packet containing the playout information of this reference client to the sync clients of both Alice and Bob.



In this case Bob has the slowest connection and the reference client therefore includes a delay similar to the one experienced by Bob. Upon reception of this information, Alice’s RTP client can choose what to do with this information. In this case it decreases its playout rate temporarily until it matches with the reference client playout (and thus matches Bob’s playout). Another option for Alice’s TV would be to simply pause playback until it catches up. The exact implementation of the synchronization algorithm is up to the client.

Upon reception of the reference client RTCP IDMS packet, Bob’s client does not have to do anything since it is already synchronized to the reference client (since it is based on Bob’s delay). Note that other synchronization algorithms may introduce even more delay than the one experienced by the most delayed client, e.g. to account for delay variations, for new clients joining an existing synchronization group, etc.

4. Inter-Destination Media Synchronization use cases

There are a large number of use cases imaginable in which IDMS might be useful. This section will highlight some of them. It should be noted that this section is in no way meant to be exhaustive.

A first usage scenario for IDMS is Social TV. Social TV is the combination of media content consumption by two or more users at different devices and locations and real-time communication between those users. An example of Social TV, is when two or more users are watching the same television broadcast at different devices and locations, while communicating with each other using text, audio and/or video. A skew in their media playout processes can have adverse effects on their experience. A well-known use case here is one friend experiencing a goal in a football match well before or after other friend(s).

Another potential use case for IDMS is a networked video wall. A video wall consists of multiple computer monitors, video projectors, or television sets tiled together contiguously or overlapped in order to form one large screen. Each of the screens reproduces a portion of the larger picture. In some implementations, each screen may be individually connected to the network and receive its portion of the overall image from a network-connected video server or video scaler. Screens are refreshed at 60 hertz (every 16-2/3 milliseconds) or potentially faster. If the refresh is not synchronized, the effect of multiple screens acting as one is broken.

A third usage scenario is that of the networked loudspeakers, in which two or more speakers are connected to the network individually. Such situations can for example be found in large conference rooms,



legislative chambers, classrooms (especially those supporting distance learning) and other large-scale environments such as stadiums. Since humans are more susceptible to differences in audio delay, this use case needs even more accuracy than the video wall use case. Depending on the exact application, the need for accuracy can then be in the range of microseconds.

5. Architecture for Inter-Destination Media Synchronization

The architecture for IDMS, which is based on a sync-maestro architecture [Boronat2009], is sketched below. The Synchronization Client (SC) and Media Synchronization Application Server (MSAS) entities are shown as additional functionality for the RTP receiver and sender respectively.

It should be noted that a master/slave type of architecture is also supported by having one of the SC devices also act as an MSAS. In this case the MSAS functionality is thus embedded in an RTP receiver instead of an RTP sender.

+-----------------------+ +-----------------------+ | | SR + | | | RTP Receiver | RTCP | RTP Sender | | | IDMS | | | +-----------------+ | <----- | +-----------------+ | | | | | | | | | | | Synchronization | | | | Media | | | | Client | | | | Synchronization | | | | (SC) | | | | Application | | | | | | | | Server | | | | | | RR+XR | | (MSAS) | | | | | | -----> | | | | | +-----------------+ | | +-----------------+ | | | | | +-----------------------+ +-----------------------+

5.1. Media Synchronization Application Server (MSAS)

An MSAS collects RTP packet arrival times and playout times from one or more SC(s) in a synchronization group. The MSAS summarizes and distributes this information to the SCs in the synchronization group as synchronization settings, e.g. by determining the SC with the most lagged playout and using its reported RTP packet arrival time and playout time as a summary.



5.2. Synchronization Client (SC)

An SC reports on RTP packet arrival times and playout times of a media stream. It can receive summaries of such information, and use that to adjust its playout buffer.

5.3. Communication between MSAS and SCs

Two different message types are used for the communication between MSAS and SCs. For the SC->MSAS message containing the playout information of a particular client, an RTCP XR Block Type is used (see Section 6). For the MSAS->SC message containing the synchronization settings instructions, a new RTCP Packet Type is defined (see Section 7).

6. RTCP XR Block Type for IDMS

This section describes the RTCP XR Block Type for reporting IDMS information on an RTP media stream. Its definition is based on [RFC3611]. The RTCP XR is used to provide feedback information on receipt times and presentation times of RTP packets to e.g. a Sender [RFC3611], a Feedback Target [RFC5576] or a Third Party Monitor [RFC3611].

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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |V=2|P| Resrv | PT=XR=207 | length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | SSRC of packet sender | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ | BT=12 | SPST |Resrv|P| block length=7 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | PT | Resrv | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Media Stream Correlation Identifier | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | SSRC of media source | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Packet Received NTP timestamp, most significant word | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Packet Received NTP timestamp, least significant word | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Packet Received RTP timestamp | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Packet Presented NTP timestamp | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+



The first 64 bits form the header of the RTCP XR, as defined in [RFC3611]. The SSRC of packet sender identifies the sender of the specific RTCP packet.

The IDMS report block consists of 8 32-bit words, with the following fields:

Block Type (BT): 8 bits. It identifies the block format. Its value SHALL be set to 12.

Synchronization Packet Sender Type (SPST): 4 bits. This field identifies the role of the packet sender for this specific eXtended Report. It can have the following values:

SPST=0 Reserved For future use.

SPST=1 The packet sender is an SC. It uses this XR to report synchronization status information. Timestamps relate to the SC input.

SPST=2 This setting is reserved in order to preserve compatibility with ETSI TISPAN [TS183063]. See Section 11 for more information.

SPST=3-15 Reserved For future use.

Reserved bits (Resrv): 3 bits. These bits are reserved for future definition. In the absence of such a definition, the bits in this field MUST be set to zero and MUST be ignored by the receiver.

Packet Presented NTP timestamp flag (P): 1 bit. Bit set to 1 if the Packet Presented NTP timestamp field contains a value, 0 if it is empty. If this flag is set to zero, then the Packet Presented NTP timestamp SHALL be ignored.

Block Length: 16 bits. This field indicates the length of the block in 32 bit words minus one and SHALL be set to 7, as this RTCP Block Type has a fixed length.

Payload Type (PT): 7 bits. This field identifies the format of the media payload, according to [RFC3551]. The media payload is associated with an RTP timestamp clock rate. This clock rate provides the time base for the RTP timestamp counter.

Reserved bits (Resrv): 25 bits. These bits are reserved for future use and SHALL be set to 0.

Media Stream Correlation Identifier: 32 bits. This identifier is used to correlate synchronized media streams. The value 0 (all bits



are set "0") indicates that this field is empty. The value 2^32-1 (all bits are set "1") is reserved for future use. If the RTCP Packet Sender is an SC (SPST=1) or an MSAS (SPST=2), then the Media Stream Correlation Identifier maps on the Synchronization Group Identifier (SyncGroupId) to which the report applies.

SSRC: 32 bits. The SSRC of the media source SHALL be set to the value of the SSRC identifier carried in the RTP header [RFC3550] of the RTP packet to which the XR relates.

Packet Received NTP timestamp: 64 bits. This timestamp reflects the wall clock time at the moment of arrival of the first octet of the RTP packet to which the XR relates. It is formatted based on the NTP timestamp format as specified in [RFC5905]. See Section 8 for more information on how this field is used.

Packet Received RTP timestamp: 32 bits. This timestamp has the value of the RTP timestamp carried in the RTP header [RFC3550] of the RTP packet to which the XR relates. Several consecutive RTP packets will have equal timestamps if they are (logically) generated at once, e.g., belong to the same video frame. It may well be the case that one receiver reports on the first RTP packet having a certain RTP timestamp and a second receiver reports on the last RTP packet having that same RTP timestamp. This would lead to an error in the synchronization algorithm due to the faulty interpretation of considering both reports to be on the same RTP packet. To solve this, an SC SHOULD report on RTP packets in which a certain RTP timestamp shows up for the first time.

Packet Presented NTP timestamp: 32 bits. This timestamp reflects the wall clock time at the moment the data contained in the first octet of the associated RTP packet is presented to the user. It is based on the time format used by NTP and consists of the least significant 16 bits of the NTP seconds part and the most significant 16 bits of the NTP fractional second part. If this field is empty, then it SHALL be set to 0 and the Packet Presented NTP timestamp flag (P) SHALL be set to 0. Presented here means the moment the data is played out to the user of the system, i.e. sound played out through speakers, video images being displayed on some display, etc. The accuracy resulting from the synchronization algorithm will only be as good as the accuracy with which the receivers can determine the delay between receiving packets and presenting them to the end-user.

7. RTCP Packet Type for IDMS (IDMS report)

This section specifies the RTCP Packet Type for indicating synchronization settings instructions to the receivers of the RTP media stream. Its definition is based on [RFC3550].



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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |V=2|P| Resrv | PT=TBD | length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | SSRC of packet sender | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ | SSRC of media source | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Media Stream Correlation Identifier | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Packet Received NTP timestamp, most significant word | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Packet Received NTP timestamp, least significant word | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Packet Received RTP timestamp | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Packet Presented NTP timestamp, most significant word | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Packet Presented NTP timestamp, least significant word | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

The first 64 bits form the header of the RTCP Packet Type, as defined in [RFC3550]. The SSRC of packet sender identifies the sender of the specific RTCP packet.

The RTCP IDMS packet consists of 7 32-bit words, with the following fields:

SSRC: 32 bits. The SSRC of the media source SHALL be set to the value of the SSRC identifier carried in the RTP header [RFC3550] of the RTP packet to which the RTCP IDMS packet relates.

Media Stream Correlation Identifier: 32 bits. This identifier is used to correlate synchronized media streams. The value 0 (all bits are set "0") indicates that this field is empty. The value 2^32-1 (all bits are set "1") is reserved for future use. The Media Stream Correlation Identifier maps on the SyncGroupId of the group to which this packet is sent.

Packet Received NTP timestamp: 64 bits. This timestamp reflects the wall clock time at the reference client at the moment it received the first octet of the RTP packet to which this packet relates. It can be used by the synchronization algorithm on the receiving SC to adjust its playout timing in order to achieve synchronization, e.g. to set the required playout delay. The timestamp is formatted based on the NTP timestamp format as specified in [RFC5905]. See Section 8 for more information on how this field is used.



Packet Received RTP timestamp: 32 bits. This timestamp has the value of the RTP timestamp carried in the RTP header [RFC3550] of the RTP packet to which the XR relates. This SHOULD relate to the first arriving RTP packet containing this particular RTP timestamp, in case multiple RTP packets contain the same RTP timestamp.

Packet Presented NTP timestamp: 64 bits. This timestamp reflects the wall clock time at the reference client at the moment it presented the data contained in the first octet of the associated RTP packet to the user. The timestamp is formatted based on the NTP timestamp format as specified in [RFC5905]. If this field is empty, then it SHALL be set to 0. This field MAY be left empty if none or only one of the receivers reported on presentated timestamps. Presented here means the moment the data is played out to the user of the system.

In some use cases (e.g. phased array transducers), the level of control an MSAS might need to have over the exact moment of playout is so precise that a 32bit Presented Timestamp will not suffice. For this reason, this RTCP Packet Type for IDMS includes a 64bit Presented Timestamp field. Since an MSAS will in practice always add some extra delay to the delay reported by the most lagged receiver (to account for packet jitter), it suffices for the IDMS XR Block Type with which the SCs report on their playout to have a 32bit Presented Timestamp field.

8. Timing and NTP Considerations

To achieve IDMS, the different receivers involved need synchronized clocks as a common timeline for synchronization. Depending on the synchronization accuracy required, different clock synchronization methods can be used. For social TV, synchronization accuracy should be achieved on the order of hundreds of milliseconds. In that case, correct use of NTP on receivers will in most situations achieve the required accuracy. As a guideline, to deal with clock drift of receivers, receivers should synchronize their clocks at the beginning of a synchronized session. In case of high required accuracy, the synchronized clocks of different receivers should not drift beyond the accuracy required for the synchronization mechanism. In practice, this can mean that receivers need to synchronize their clocks repeatedly during a synchronization session.

Because of the stringent synchronization requirements for achieving good audio in some use cases, a high accuracy will be needed. In this case, use of the global NTP system may not be sufficient. For improved accuracy, a local NTP server could be set up, or some other more accurate clock synchronization mechanism can be used, such as GPS time or the Precision Time Protocol [IEEE-1588].



[I-D.draft-williams-avtcore-clksrc] defines a set of SDP parameters for signaling the clock synchronization source or sources available to and used by the individual receivers. Using these paramenters, an SC can indicate which synchronization source is being used at the moment, the last time the SC synchronized with this source and the synchronization frequency. An SC can also indicate any other synchronization sources available to it. This allows multiple SCs in an IDMS session to use the same or a similar clock source for their session.

Applications performing IDMS may or may not be able to choose a synchronization method for the system clock, because this may be a system-wide setting which the application cannot change. How applications deal with this is up to the implementation. The application might control the system clock, or it might use a separate application clock or even a separate IDMS session clock. It might also report on the system clock and the synchronization method used, without being able to change it.

[I-D.draft-williams-avtcore-clksrc] also presents some general guidelines on dealing with leap seconds within an RTP context. When relying on NTP for clock synchronization, IDMS is particularly sensitive to leap second induced timing discrepancies. 9. SDP Parameter for RTCP XR IDMS Block Type

The SDP parameter sync-group is used to signal the use of the RTCP XR block for IDMS. It is also used to carry an identifier of the synchronization group to which clients belong or will belong. This SDP parameter extends rtcp-xr-attrib as follows, using Augmented Backus-Naur Form [RFC5234].

rtcp-xr-attrib = "a=" "rtcp-xr" ":" [xr-format *(SP xr-format)] CRLF ; Original definition from [RFC3611], section 5.1

xr-format =/ grp-sync ; Extending xr-format for Inter-Destination Media Synchronization

grp-sync = "grp-sync" [",sync-group=" SyncGroupId]

SyncGroupId = 1*DIGIT ; Numerical value from 0 till 4294967295

DIGIT = %x30-39

SyncGroupId is a 32-bit unsigned integer represented in decimal. SyncGroupId identifies a group of SCs for IDMS. It maps on the Media Stream Correlation Identifier as described in sections 6 and 7. The value SyncGroupId=0 represents an empty SyncGroupId. The value 4294967295 (2^32-1) is reserved for future use.



The following is an example of the SDP attribute for IDMS

a=rtcp-xr:grp-sync,sync-group=42

10. SDP Parameter for RTCP IDMS Packet Type

The SDP parameter rtcp-idms is used to signal the use of the RTCP IDMS Packet Type for IDMS. It is also used to carry an identifier for the synchronization group to which clients belong or will belong. The SDP parameter is used as a media-level attribute during session setup. This SDP parameter is defined as follows, using Augmented Backus-Naur Form [RFC5234].

rtcp-idms = "a=" "rtcp-idms" ":" [sync-grp] CRLF

sync-grp = "sync-group=" SyncGroupId

SyncGroupId = 1*DIGIT ; Numerical value from 0 till 4294967295

DIGIT = %x30-39

SyncGroupId is a 32-bit unsigned integer and represented in decimal. SyncGroupId identifies a group of SCs for IDMS. The value SyncGroupId=0 represents an empty SyncGroupId. The value 4294967295 (2^32-1) is reserved for future use.

The following is an example of the SDP attribute for IDMS.

a=rtcp-idms:sync-group=42

11. Compatibility with ETSI TISPAN

As described in Section 1.4, ETSI TISPAN has also described a mechanism for IDMS in [TS183063]. One of the main differences between the TISPAN document and this document is the fact that the TISPAN solution uses an RTPC XR block for both the SC->MSAS message and the MSAS->SC message (by selecting different SPST-types), while this document specifies a new RTCP Packet Type for the MSAS->SC message. The message from MSAS to SC is not in any way a report on how a receiver sees a session, and therefore a separate RTCP packet type is more appropriate than the XR block solution chosen in ETSI TISPAN.

In order to maintain backward-compatibility, the RTCP XR block used for SC->MSAS signaling specified in this document is fully compatible with the TISPAN defined XR block.



For the MSAS->SC signaling, it is recommended to use the RTCP IDMS Packet Type defined in this document. The TISPAN XR block with SPST=2 MAY be used for purposes of compatibility with the TISPAN solution, but MUST NOT be used if all nodes involved support the new RTCP IDMS Packet Type.

The above means that the IANA registry contains two SDP parameters for the MSAS->SC signaling; one for the ETSI TISPAN solution and one for the IETF solution. This also means that if all elements in the SDP negotiation support the IETF solution they SHOULD use the new RTCP IDMS Packet Type.

12. On the use of presentation timestamps

A receiver can report on different timing events, i.e. on packet arrival times and on playout times. A receiver SHALL report on arrival times and a receiver MAY report on playout times. RTP packet arrival times are relatively easy to report on. Normally, the processing and playout of the same media stream by different receivers will take roughly the same amount of time. It can suffice for many applications, such as social TV, to synchronize on packet arrival times. Also, if the receivers are in some way controlled, e.g. having the same buffer settings and decoding times, high accuracy can be achieved. However, if all receivers in a synchronization session have the ability to report on, and thus synchronize on packet presented times, this may be more accurate. It is up to applications and implementations of this RTCP extension whether to implement and use this.


The specified RTCP XR Block Type in this document is used to collect, summarize and distribute information on packet reception- and playout-times of streaming media. The information may be used to orchestrate the media playout at multiple devices.

Errors in the information, either accidental or malicious, may lead to undesired behavior. For example, if one device erroneously reports a two-hour delayed playout, then another device in the same synchronization group could decide to delay its playout by two hours as well, in order to keep its playout synchronized. A user would likely interpret this two hour delay as a malfunctioning service.

Therefore, the application logic of both Synchronization Clients and Media Synchronization Application Servers should check for inconsistent information. Differences in playout time exceeding



configured limits (e.g. more than ten seconds) could be an indication of such inconsistent information.

No new mechanisms are introduced in this document to ensure confidentiality. Encryption procedures, such as those being suggested for a Secure RTP (SRTP) at the time that this document was written, can be used when confidentiality is a concern to end hosts.


New RTCP Packet Types and RTCP XR Block Types are subject to IANA registration. For general guidelines on IANA considerations for RTCP XR, refer to [RFC3611].

[TS 183 063] assigns the block type value 12 in the RTCP XR Block Type Registry to "Inter-Destination Media Synchronization Block". [TS183063] also registers the SDP [RFC4566] parameter "grp-sync" for the "rtcp-xr" attribute in the RTCP XR SDP Parameters Registry.

Further, this document defines a new RTCP packet type called IDMS report. This new packet type is registered with the IANA registry of RTP parameters, based on the specification in Section 10.

Further, this document defines a new SDP parameter "rtcp-idms" within the existing IANA registry of SDP Parameters.

The SDP attribute "rtcp-idms" defined by this document is registered with the IANA registry of SDP Parameters as follows:

SDP Attribute ("att-field"):

Attribute name: rtcp-idms

Long form: RTCP report block for IDMS

Type of name: att-field

Type of attribute: media level

Subject to charset: no

Purpose: see sections 7 and 10 of this document

Reference: this document



Values: see this document

15. Contributors

The following people have participated as co-authors or provided substantial contributions to this document: Omar Niamut, Fabian Walraven, Ishan Vaishnavi, Rufael Mekuria.

16. Conclusions

This document describes the RTCP XR block type for IDMS, the RTCP IDMS report and the associated SDP parameters for Inter-Destination Media Synchronization.

17. References




[I-D.draft-williams-avtcore-clksrc] Williams, A., van Brandenburg, R., Stokking, H., and K. Gross, "RTP Clock Source Signalling, draft-williams-avtcore-clksrc-00", March 2012.

[RFC2119] Bradner, S., "Key Words for use in RFCs to Indicate Requirement Levels, RFC 2119", March 1997.

[RFC3550] Schulzrinne, H., Casner, S., Frederick, R., and V. Jacobson, "RTP: A Transport Protocol for Real-Time Applications, RFC3550", July 2003.

[RFC3551] Schulzrinne, H. and S. Casner, "RTP Profile for Audio and Video conferences with Minimal Control, RFC3551", July 2003.

[RFC3611] Friedman, T., Ed., Caceres, R., Ed., and A. Clark, Ed., "RTP Control Protocol Extended Reports (RTCP XR), RFC3611", November 2003.

[RFC4566] Handley, M., Jacobson, V., and C. Perkins, "SDP: Session Description Protocol, RFC4566", July 2006.

[RFC5234] Crocker, D., Ed. and P. Overell, "Augmented BNF for Syntax Specifications, RFC5234", January 2008.

[RFC5576] Lenox, J., Ott, J., and T. Schierl, "Source-Specific Media Atrributes in the Session Description Protocol, RFC5576", June 2009.

[RFC5760] Ott, J., Chesterfield, J., and E. Schooler, "RTP Control Protocol (RTCP) Extensions for Single-Source Multicast Sessions with Unicast Feedback, RFC5760", February 2010.

[RFC5905] Mills, D., Martin, J., Ed., Burbank, J., and W. Kasch, "Network Time Protocol Version 4: Protocol and Algorithms Specifications, RFC5905", February 2010.

[TS183063] "IMS-based IPTV stage 3 specification, TS 183 063 v3.4.1", June 2010.


[Boronat2009] Boronat, F., Lloret, J., and M. Garcia, "Multimedia group



and inter-stream synchronization techniques: a comparative study", Elsevier Information Systems 34 (2009), pp. 108- 131".

[IEEE-1588] "1588-2008 - IEEE Standard for a Precision Clock Synchronization Protocol for Networked Measurement and Control Systems", 2008.

[Ishibashi2006] Ishibashi, Y., Nagasaka, M., and N. Fujiyoshi, "Subjective Assessment of Fairness among users in multipoint communications", Proceedings of the 2006 ACM SIGCHI internation conference on Advances in computer entertainment technology, 2006".

[RFC5868] Ott, J. and C. Perkins, "Guidelines for Extending the RTP Control Protocol (RTCP), RFC5968", September 2010.


Ray van Brandenburg TNO Brassersplein 2 Delft 2612CT the Netherlands

Phone: +31-88-866-7000 Email: [email protected]

Hans Stokking TNO Brassersplein 2 Delft 2612CT the Netherlands


M. Oskar van Deventer TNO Brassersplein 2 Delft 2612CT the Netherlands




Fernando Boronat Universidad Politecnica de Valencia IGIC Institute, Universitat Politecnica de Valencia-Campus de Gandia (UPV), C/ Paraninfo, 1, Grao de Gandia Valencia 46730 Spain

Phone: +34 962 849 341 Email: [email protected]

Mario Montagud Universidad Politecnica de Valencia IGIC Institute, Universitat Politecnica de Valencia-Campus de Gandia (UPV), C/ Paraninfo, 1, Grao de Gandia Valencia 46730 Spain

Phone: +34 962 849 341 Email: [email protected]

Kevin Gross AVA Networks



AVTCORE J. LennoxInternet-Draft VidyoIntended status: Informational March 5, 2012Expires: September 6, 2012

Real-Time Transport Protocol (RTP) Topology Considerations for Offer/ Answer-Initiated Sessions. draft-lennox-avtcore-rtp-topo-offer-answer-00

Abstract

This document discusses a number of considerations related to the topologies of Real-Time Transport Protocol (RTP) sessions initated using the Session Description (SDP) unicast Offer/Answer Model, especially as applied to source-multiplexed sessions.

The primary observation is that certain topologies cannot be created by unicast SDP Offer/Answer. Notably, the it is not possible to negotiate the topology that RFC 5117 calls Topo-Transport-Translator (or "relay").

As a consequence of this limitation, certain topological assumptions can safely be made for RTP sessions initiated using unicast SDP Offer/Answer; and therefore, certain optimizations to RTP are possible in such sessions. This document also describes the optimizations that these assumptions make possible.

Status of this Memo





Copyright Notice

Copyright (c) 2012 IETF Trust and the persons identified as the

Lennox Expires September 6, 2012 [Page 1]

Internet-Draft RTP Offer/Answer Topology Considerations March 2012

document authors. All rights reserved.


Table of Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 3 3. RTP Topologies . . . . . . . . . . . . . . . . . . . . . . . . 3 4. RTP Topologies and SDP Offer/Answer . . . . . . . . . . . . . 5 5. Advantages for Assuming RTCP rewriting . . . . . . . . . . . . 6 5.1. Independent RTCP bandwidth . . . . . . . . . . . . . . . . 6 5.2. Optimization of receiver reports . . . . . . . . . . . . . 7 6. Normative recommendations . . . . . . . . . . . . . . . . . . 8 7. Limitations of media and RTCP modifying middleboxes . . . . . 9 8. Security Considerations . . . . . . . . . . . . . . . . . . . 10 9. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 10 10. References . . . . . . . . . . . . . . . . . . . . . . . . . . 10 10.1. Normative References . . . . . . . . . . . . . . . . . . . 10 10.2. Informative References . . . . . . . . . . . . . . . . . . 11 Author’s Address . . . . . . . . . . . . . . . . . . . . . . . . . 11



1. Introduction

For interactive conferencing, by the far most common method of negotiating Real-Time Tranport Protocol (RTP) [RFC3550] sessions is to use the Session Description Protocol [RFC4566] Offer/Answer Model [RFC3264]. In particular, conferences are typically negotated using offer/answer unicast streams.

As discussed in [RFC5117], however, RTP sessions can be arranged in a fairly large number of topologies, more complexly than the simple dichotomy of "unicast" or "multicast" used by the Offer/Answer model. Most of the unicast-based topologies in RFC 5117 can be negotiated as SDP Offer/Answer unicast streams. However, for reasons that are explained in Section 3, one topology in particular cannot be negotiated using SDP Offer/Answer: Topo-Transport-Translator.

While this might initially seem to be a limitation of SDP Offer/ Answer, it actually turns out that if an endpoint can assume that its RTP topologies are limited to those that can be negotated using offer/answer, a number of RTP optimizations become possible. These are discussed in Section 5, with specific recommendations in Section 6.

2. Terminology

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119 [RFC2119] and indicate requirement levels for compliant implementations.

3. RTP Topologies

[RFC5117] discusses multi-endpoint topologies of RTP sessions in detail. For the purposes of this document, a few topologies in particular are of interest, and will be described in detail.

+---+ +---+ | A |<------->| B | +---+ +---+

Figure 1: Point to Point

The simplest topology, shown in Figure 1, is Point to Point (Topo-



Point-to-Point). A sends to B, and only B, while B sends to A, and only A. An endpoint can still use multiple synchronization sources (SSRCs) in a session.

This is topology is straightforwardly negotated by SDP Offer-Answer unicast streams.

+-----+ +---+ / \ +---+ | A |----/ \---| B | +---+ / Multi- \ +---+ + Cast + +---+ \ Network / +---+ | C |----\ /---| D | +---+ \ / +---+ +-----+

Figure 2: Point to Multipoint Using Multicast

Secondly, in the Topo-Multicast topology, shown in Figure 2, traffic from each endpoint in an RTP session is received by all other session participants, transported by the network level by sending to a special IP address. These sessions can negotiated by the Offer/ Answer model as multicast streams.

Multicast sessions are often supported within individual domains, but are not typically supported accross the public Internet.

+---+ +------------+ +---+ | A |<---->| |<---->| B | +---+ | | +---+ | Translator | +---+ | | +---+ | C |<---->| |<---->| D | +---+ +------------+ +---+

Figure 3: RTP Translator (Relay) with Only Unicast Paths

Finally, for the purposes of this discussion, one other topology described in [RFC5117] is specifically relevant; the others can all be generalized. The specific topology is the topology Topo- Transport-Translator, illustrated in Figure 3, which simply forwards unicast traffic (both media and Real-Time Transport Control Protocol (RTCP) traffic) among all the unicast connections to a central



translator.

+---+ +------------+ +---+ | A |<---->| Media & |<---->| B | +---+ | RTCP | +---+ | Modifier | +---+ | or | +---+ | C |<---->| Terminator|<---->| D | +---+ +------------+ +---+

Figure 4: RTCP-Modifying Central Box

By contrast, the topologies Topo-Media-Translator, Topo-Mixer, Topo- Video-Switch-Mixer, and Topo-RTCP-Terminating-MCU can all be summarized by the diagram shown in Figure 4. These topologies all have the common property that they have a central box (a media translator, mixer, or MCU) which terminates media and RTCP traffic, and forwards it, modified to a greater or lesser extent, to the topology’s other endpoints.

4. RTP Topologies and SDP Offer/Answer

The SDP Offer/Answer Model [RFC3264] specifies different behavior depending on whether the address of the transport connection being negotiated is a unicast and multicast address. Effectively, offer/ answer assumes that the stream being negotated corresponds either to Topo-Point-to-Point or Topo-Multicast.

Most of the RFC 5117 unicast topologies described in Section 3 can be negotiated by the centralized point negotating with each endpoint using the Offer/Answer unicast mode. However, the Topo-Transport- Translator cannot be.

The difficulty is that SDP Offer/Answer unicast exchange is designed to negotate each end of the excahnge’s separate view of the session, and each endpoint has a fair bit of control over what its view of the session should look like. However, because the translator at the center of the Topo-Transport-Translator topology forwards media and RTCP unmodified, it is necessary that all participants have a common view of all non-transport aspects of the session.

Thus, the freedom that the Offer/Answer model gives each endpoint to control its view of the session prevents the central box from enforcing a single, uniform view of it. Among the session aspects that can be different among the session participants are:



Media Types: Unicast Offer/Answer participants have the freedom to remove media types from the session description (in an answer or an updated offer). They also have the freedom to change some fmtp media type parameters. Moreover, though RFC 3264 indicates that it is NOT RECOMMENDED, they have the freedom to change the mapping between media typees and RTP payload type numbers. Bandwidth: An answerer can specify a bandwidth attribute (SDP b= value) for any media stream, indicating the bandwidth that the answerer would like the offerer to use when sending media. This does bear any relationship to bandwidth values in use in the other direction. (This is somewhat problematic as SDP bandwidth parameters are used to calculate RTCP bandwidth, and thus RTP session membership timeout intervals.) Packetization Time: An SDP offer or answer can specify the ptime attribute (packetization time interval) with which it wants to receive media. This value is independent in offers and answers.

In addition to these mismatches for the attributes specified by the core SDP Offer/Answer specification, there are of course many extensions to SDP which specify Offer/Answer behavior. These are not discussed here, but many of them would have similar issues with the Topo-Transport-Translator topology.

Any of the media and RTCP terminating topologies described in Section 3 as modifying media and RTCP will be able to repair these mismatches, or else reject an endpoint that asks for a configuration beyond its capacity to repair. The mismatch difficulties arise only for the Topo-Transport-Translator.

5. Advantages for Assuming RTCP rewriting

If we assume that we always have a central box that can rewrite, or generates its own, media and RTCP, a number of optimizations and protocol clarifications become possible.

5.1. Independent RTCP bandwidth

SDP Unicast Offer/answer allows RTP session bandwidth to be specified independently in each direction of the offer/answer exchange. The assumption is that bandwidth in each direction is (over the relevant bottleneck links) non-rival, and that the available bitrates can in some circumstances be dramatically asymmetric.

It has always been somewhat unclear how offer/answer assymetric bandwidths interact with the RTCP bandwidth fraction (5%, or the SDP bandwidth modifiers).



If we assume that RTCP is never passively relayed, but rather will always either be consumed locally or will actively be rewritten before being forwarded, this problem largely goes away. Each side of the unicast RTP session domain gets the appropriate fraction of its (sending) RTP bandwidth to send RTCP. It can divide this fraction among its sources as it wishes, subject ot the constraint that a regular report is sent for each source with appropriate frequency to prevent timeouts. Group size estimation is only needed for timeout calculation. It can be done independently for sending and receiving media.

Since RTCP bandwidth can be shared among all the sources, a sender can then also send feedback from multiple of its sources in a single compound RTCP packet, up to transport MTU issues, reducing transport overhead.

5.2. Optimization of receiver reports

For the benefit of Topo-Multicast and Topo-Transport-Translator, in an RTCP session, all session participants send RTCP reception reports (in SR or RR RTCP packets) for every active RTP source from which they have received packets in the previous reporting interval.

This means that the number of reception reports is quadratic in the number of sources in a conference. (Specifically, the number equals the number of conference participants, times the number of active senders whos sent during a report interval. However, because the report interval itself scales with the number of sources, this will in a many-to-many conference converge to being quadratic in the number of sources.)

In cases where there is an media-and-RTCP-modifying middlebox, this quadratic behavior is useless. The relevant reception report information is that between and endpoint and the middlebox, since the middlebox can often perform reliability and repair mechanisms on its own. These excess reception reports then increase the size of RTCP packets, which by the formulas for calculating RTCP packet transmission schedules reduces the RTCP timing interval. Thus, these excess reception reports consume bandwidth which could instead be used for timely RTCP feedback of relevant data.

These quadratic reception reports are particularly useless in scenarios where a given session participant is sending multiple sources of its own (rather than forwarding multiple remote sources) in the same RTP session. Examples of such use cases are the CLUE Telepresence model [I-D.lennox-clue-rtp-usage], bundling of multiple media types onto a single RTP session [I-D.ietf-mmusic-sdp-bundle-negotiation], and single-session RTP



retransmission [RFC4588]; in general, this will apply to most scenarios in which SDP source descriptions [RFC5576] are used.

The most useless data is reception reports by one local source about another, since these will always (by definition) be "received" perfectly (with zero loss and jitter) by their sender.

Nearly as useless redundant feedback from multiple co-located sources about the same remote source. Since RTP traffic is in fact received by an endpoint, not a source, this information will either be identical (if an endpoint choses to synchronize its RTCP feedback messages) or multiple, non-commensurate transmissions of the same information (if it does not).

Also often useless is feedback by one remote source about another one -- while there are some conceivable use cases where this could be relevant information (for instance, a monitoring application), in most conferencing models, this is uninteresting and unimportant.

6. Normative recommendations

Based on the analysis in Section 5, this section makes some normative recommendations for the behavior of RTP endpoints in sessions negotiated using unicast SDP Offer/Answer.

(Open issue: it is possible that these recommendations might need to be a normative update to [RFC3550]; alternatively, they may just be implementation guidance.)

When an RTP [RFC3550] session is negotiated using unicast SDP offer/ answer [RFC3264], RTCP bandwidth, and thus RTCP packet intervals and RTP group membership timeout rules, MUST be calculated separately for the receiving and sending direction, using the rules specified in [RFC3550] as modified by any SDP attributes or the RTP profile in use. An endpoint MAY send RTCP up to its available bandwidth, independent of the bandwidth consumed in the reverse direction, again subject to the SDP modifiers and profiles in use.

An endpoint MAY choose to send multiple sources’ RTCP messages in a single compound RTCP packet (though such compound packets SHOULD NOT exceed the path MTU, if avoidable and if it is known). This will reduce the average compound RTCP packet size, and thus increase the frequency with which RTCP messages can be sent. Regular (non- feedback) RTCP compound packets MUST still begin with an SR or RR packet, but otherwise MAY contain RTCP packets in any order. Receivers MUST be prepared to receive such compound packets.



An endpoint SHOULD NOT send reception reports from one of its own sources about another one ("cross-reports"). Endpoints receiving reception reports MUST be prepared that their peers might not be sending reception reports about their own sources.

Similarly, an endpoint sending multiple sources SHOULD NOT send reception reports about a remote source from more than one of its local sources. Instead, it SHOULD create or pick one local source as the "reporting" source for each remote source, which sends full report blocks; all its other sources SHOULD be treated as if they were disconnected, and never saw that remote source. This reporting source MAY be one of the sending sources in the session, or MAY be a receive-only source created simply for the purpose of sending feedback. An endpoint MAY choose different local sources as the reporting source for different remote sources (for example, if it is using bundle [I-D.ietf-mmusic-sdp-bundle-negotiation], it could choose to send reports about remote audio sources from its local audio source, and reports about remote video sources from its local video source), or it MAY choose a single local source for all its reports. If the reporting source leaves the session (sends BYE), another reporting source MUST be chosen. If the AVPF [RFC4585] RTP profile, or one if its secure equivalents, is in use, this "reporting" source SHOULD also be the source for any AVPF feedback messages about its remote sources, as well. Endpoints interpreting reception reports MUST be prepared to receive RTCP SR or RR messages where only one remote source is reporting about its sources.

7. Limitations of media and RTCP modifying middleboxes

There are a few limitations of media and RTCP modifying middleboxes, compared to what can be done by the Topo-Transport-Translator topology.

A media and RTCP modifying middlebox will, necessarily, be more complex (and thus be more expensive, or have lower capacity), than a pure transport forwarder.

It is not possible to deploy new RTCP extensions across an unmofidified RTCP-modifying central box, as that box will not know how to re-write these extensions so they are correctly forwarded.

If SRTP is in use, these central middleboxes must be trusted with the SRTP keying material. (Since SRTP keying material is usually negotiated hop-by-hop, they may be doing a complete SRTP decryption and re-encryption, with unrelated keys, and possibly even translating between different ciphers or cipher strengths.)



It is possible, if the recommendations of Section 6 are in use, that a naive RTCP monitor might think that an RTP flow should actually be interpreted as Topo-Transport-Translator. In this case, it might think that there is a network disconnection between the non-reporting sources and the sources on which they are not reporting. However, architecturally it is very unclear if such monitors actually exist, for conferencing applications, or would care about a disconnection of this sort.


See the security considerations of [RFC5117]. Notably, as discussed in Section 7, a centralized media and RTCP modifying box will need to terminate SRTP and SRTCP, and so must be a trusted entity.


This document makes no requests of IANA.

Note to the RFC Editor: please remove this section before publication.

10. References



[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.

[RFC4585] Ott, J., Wenger, S., Sato, N., Burmeister, C., and J. Rey, "Extended RTP Profile for Real-time Transport Control Protocol (RTCP)-Based Feedback (RTP/AVPF)", RFC 4585, July 2006.




[I-D.ietf-mmusic-sdp-bundle-negotiation] Holmberg, C. and H. Alvestrand, "Multiplexing Negotiation Using Session Description Protocol (SDP) Port Numbers", draft-ietf-mmusic-sdp-bundle-negotiation-00 (work in progress), February 2012.

[I-D.lennox-clue-rtp-usage] Romanow, A., Lennox, J., and P. Witty, "Real-Time Transport Protocol (RTP) Usage for Telepresence Sessions", draft-lennox-clue-rtp-usage-02 (work in progress), February 2012.


[RFC5117] Westerlund, M. and S. Wenger, "RTP Topologies", RFC 5117, January 2008.

[RFC5576] Lennox, J., Ott, J., and T. Schierl, "Source-Specific Media Attributes in the Session Description Protocol (SDP)", RFC 5576, June 2009.

Author’s Address

Jonathan Lennox Vidyo, Inc. 433 Hackensack Avenue Seventh Floor Hackensack, NJ 07601 US



Network Working Group C. PerkinsInternet-Draft University of GlasgowIntended status: Standards Track V. SinghExpires: September 5, 2012 Aalto University March 4, 2012

RTP Congestion Control: Circuit Breakers for Unicast Sessions draft-perkins-avtcore-rtp-circuit-breakers-00

Abstract

The Real-time Transport Protocol (RTP) is widely used for telephony, video conferencing, and telepresence applications. These applications are often used over best-effort UDP/IP networks. If congestion control is not implemented then network congestion will deteriorate the user’s multimedia experience. This document does not propose a congestion control algorithm. Instead, it specifies a minimal set of "circuit-breakers". Circuit-breakers are conditions under which an RTP flow should cease to transmit media to protect the network from excessive congestion. It is expected that all RTP applications running on best-effort networks will be able to run without triggering these circuit breakers in normal operation.

Status of this Memo





Copyright Notice


This document is subject to BCP 78 and the IETF Trust’s Legal Provisions Relating to IETF Documents

Perkins & Singh Expires September 5, 2012 [Page 1]

Internet-Draft RTP Circuit Breakers March 2012

(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.

Table of Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 3 3. Background . . . . . . . . . . . . . . . . . . . . . . . . . . 3 4. RTP Circuit Breakers for Systems Using the RTP/AVP Profile . . 6 4.1. RTP/AVP Circuit Breaker #1: Timeout . . . . . . . . . . . 7 4.2. RTP/AVP Circuit Breaker #2: Congestion . . . . . . . . . . 8 5. RTP Circuit Breakers for Systems Using the RTP/AVPF Profile . 10 6. Impact of RTCP XR . . . . . . . . . . . . . . . . . . . . . . 10 7. Impact of Explicit Congestion Notification (ECN) . . . . . . . 11 8. Session Timeout . . . . . . . . . . . . . . . . . . . . . . . 11 9. References . . . . . . . . . . . . . . . . . . . . . . . . . . 11 9.1. Normative References . . . . . . . . . . . . . . . . . . . 11 9.2. Informative References . . . . . . . . . . . . . . . . . . 12 Authors’ Addresses . . . . . . . . . . . . . . . . . . . . . . . . 13



1. Introduction

The Real-time Transport Protocol (RTP) [RFC3550] is widely used in voice-over-IP, video teleconferencing, and telepresence systems. Many of these systems run over best-effort IP networks, and can suffer from packet loss and increased latency due to network congestion. Designing effective RTP congestion control algorithms, to adapt the transmission of RTP-based media to match the available network capacity, while also maintaining the user experience, is a difficult but important problem. Many such congestion control and media adaptation algorithms have been proposed, but to date there is no consensus on the correct approach, or even that a single standard algorithm is desirable.

This memo does not attempt to propose a new RTP congestion control algorithm. Rather, it proposes a minimal set of "circuit breakers"; conditions under which there should be general agreement that an RTP flow is causing serious congestion, and should cease transmission. It is expected that any future standards-track congestion control algorithms for RTP will operate within the envelope defined by this memo.

2. Terminology


3. Background

We consider congestion control for unicast RTP traffic flows. This is the problem of adapting the transmission of an audio/visual data flow, encapsulated within an RTP transport session, from one sender to one receiver so that it matches the available network bandwidth. Such adaptation must be done in a way that limits the disruption to the user experience caused by both packet loss and excessive rate changes.

Congestion control for unicast RTP traffic can be implemented in one of two places in the protocol stack. One approach is to run the RTP traffic over a congestion controlled transport protocol, for example over TCP, and to adapt the media encoding to match the dictates of the transport-layer congestion control algorithm. This is safe for the network, but may be suboptimal for the media quality unless the transport protocol is designed to support real-time media flows. We do not consider this class of applications further in this memo, as



their network safety is guaranteed by the underlying transport.

Alternatively, RTP flows can be run over a non-congestion controlled transport protocol, for example UDP, performing rate adaptation at the application layer based on RTP Control Protocol (RTCP) feedback. With a well-designed, network-aware, application, this allows highly effective media quality adaptation, but there is potential to disrupt the network operation if the application does not adapt its sending rate in a timely and effective manner. This memo focusses on this class of application.

Congestion control relies on monitoring the delivery of a media flow, and responding to adapt the transmission of that flow when there are signs that the network path is congested. Network congestion may be detected in one of three ways: 1) a receiver may infer the onset of congestion by observing an increase in one-way delay caused by queue build-up within the network; 2) if Explicit Congestion Notification (ECN) [RFC3168] is supported, the network may signal the presence of congestion by marking packets with ECN Congestion Experienced (CE) marks; or 3) in the extreme case, congestion will cause packet loss, which can be detected by observing a gap in the received RTP sequence numbers. Once the onset of congestion is observed, the receiver must send feedback to the sender to indicate that the transmission rate should be reduced. How the sender reduces the transmission rate is highly dependent on the media codec being used, and is outside the scope of this memo.

There are several ways in which a receiver may send feedback to a media sender within the RTP framework:

o The base RTP specification [RFC3550] defines RTCP Reception Report (RR) packets to convey reception quality feedback information, and Sender Report (SR) packets to convey information about the media transmission. RTCP SR packets contain data that can be used to reconstruct media timing at a receiver, along with a count of the total number of octets and packets sent. RTCP RR packets report on the fraction of packets lost in the last reporting interval, the cumulative number of packets lost, the highest sequence number received, and the inter-arrival jitter. The RTCP RR packets also contain timing information that allows the sender to estimate the network round trip time (RTT) to the receivers. RTCP reports are sent periodically, with the reporting interval being determined by the number of participants in the session and a configured session bandwidth estimate. The interval between reports sent from each receiver tends to be on the order of a few seconds on average, and it is randomised to avoid synchronisation of reports from multiple receivers. If a receiver detects problems, the base RTP specification contains no provisions for sending the feedback



report early and must wait until the next scheduled reporting interval.

o The RTCP Extended Reports (XR) [RFC3611] allow reporting of more complex and sophisticated reception quality metrics, but do not change the RTCP timing rules. RTCP extended reports of potential interest for congestion control purposes are 1) the extended packet loss, discard or burst/gap metrics for reacting based on loss patterns; and 2) the end-system delay metrics for delay-based congestion control.

o Rapid feedback about the occurrence of congestion events can be achieved using the Extended RTP Profile for RTCP-Based Feedback (RTP/AVPF) [RFC4585]. This modifies the RTCP timing rules to allow RTCP reports to be sent early, in some cases immediately, provided the average RTCP reporting interval remains unchanged. It also defines new transport-layer feedback messages, including negative acknowledgements (NACKs), that can be used to report on specific congestion events. The use of the RTP/AVPF profile is dependent on signalling, but is otherwise generally backwards compatible, as it keeps the same average RTCP reporting interval as the base RTP specification. The Codec control messages [RFC5104] extend the RTP/AVPF profile with additional feedback message that can be used to influence that way in which rate adaptation occurs. The dynamics of how rapidly feedback can be sent are unchanged.

o Finally, the RTP and RTCP extensions for Explicit Congestion Notification (ECN) [I-D.ietf-avtcore-ecn-for-rtp] can be used to provide feedback on the number of packets that received an ECN Congestion Experienced (CE) mark. This extension builds on the RTP/AVPF profile to allow rapid congestion feedback.

In addition to these mechanisms for providing feedback, the sender can include an RTP header extension in each packet to record packet transmission times. There are two methods: [RFC5450] represents the transmission time in terms of a time-offset from the RTP timestamp of the packet, while [RFC6051] includes an explicit NTP-format sending timestamp (potentially more accurate, but a higher header overhead). Accurate sending timestamps can be helpful for estimating queuing delays, to get an early indication of the onset of congestion.

Taken together, these various mechanisms allow receivers to provide feedback on the senders when congestion events occur, with varying degrees of timeliness and accuracy. The key distinction is between systems that use only the basic RTCP mechanisms, without RTP/AVPF rapid feedback, and those that use the RTP/AVPF extensions, and can respond to congestion more rapidly.



4. RTP Circuit Breakers for Systems Using the RTP/AVP Profile

The feedback mechanisms defined in [RFC3550] are the minimum that can be required for a baseline circuit breaker mechanism suitable for all unicast applications of RTP. Accordingly, for an RTP circuit breaker to be useful, it should be able to detect that an RTP flow is causing excessive congestion using only basic RTCP features, without needing RTCP XR feedback or the RTP/AVPF profile for rapid RTCP reports.

Three potential congestion signals are available from the basic RTCP SR/RR packets, and are reported for each SSRC in the RTP session:

1. The sender can estimate the network round-trip time once per RTCP reporting interval, based on the contents and timing of RTCP SR and RR packets.

2. Receivers report estimated jitter (the statistical variance of the RTP data packet inter-arrival time) calculated over the RTCP reporting interval. Due to the nature of the jitter calculation ([RFC3550], section 6.4.4), the jitter is only meaningful for RTP flows that send a single data packet for each RTP timestamp value (i.e., audio flows, or video flows where each frame comprises one RTP packet).

3. Receivers report the fraction of packets lost during the RTCP reporting interval, and the cumulative number of packets lost over the entire RTP session.

These congestion signals limit the possible circuit breakers, since they give only limited visibility into the behaviour of the network.

RTT estimates are widely used in congestion control algorithms, as a proxy for queuing delay measures in delay-based congestion control, or to determine connection timeouts. RTT estimates derived from RTCP SR and RR packets send according to the RTP/AVP timing rules are too infrequent to be useful though, and don’t give enough information to distinguish a delay change due to routing updates from queuing delay caused by congestion. Accordingly, we do not use the RTT estimate alone as an RTP circuit breaker.

Increased jitter can be a signal of transient network congestion, but in the highly aggregated form reported in RTCP RR packets, it offers insufficient information to estimate the extent or persistence of congestion. Jitter reports are a useful early warning of potential network congestion, but provide an insufficiently strong signal to be used as a circuit breaker.

The remaining congestion signals are the packet loss fraction and the



cumulative number of packets lost. These are robust indicators of congestion in networks where packet loss is primarily due to queue overflows, although less accurate in networks where losses can be caused by non-congestive packet corruption. TCP also uses packet loss as a congestion signal.

Two packet loss regimes can be observed: 1) RTCP RR packets show a non-zero packet loss fraction, while the extended highest sequence number received continues to increment; and 2) RR packets show a loss fraction of zero, but the extended highest sequence number received does not increment even though the sender has been transmitting RTP data packets. The former corresponds to the TCP congestion avoidance state, and indicates a congested path that is still delivering data; the latter corresponds to a TCP timeout, and is most likely due to a path failure. We derive circuit breaker conditions for these two loss regimes.

4.1. RTP/AVP Circuit Breaker #1: Timeout

If RTP data packets are being sent while the corresponding RTCP RR packets report a non-increasing extended highest sequence number received, this is an indication that those RTP data packets are not reaching the receiver. This could be a short-term issue affecting only a few packets, perhaps caused by a slow-to-open firewall or a transient connectivity problem, but if the issue persists, it is a sign of a more ongoing and significant problem. Accordingly, if a sender of RTP data packets receives two or more consecutive RTCP RR packets from the same receiver that correspond to its transmission and have a non-increasing extended highest sequence number received field, then that sender SHOULD cease transmission.

Systems that usually send at a high data rate, but which can reduce their data rate significantly (i.e., by an order of magnitude), MAY first reduce their sending rate to this lower value, but MUST then cease transmission if the problem does not resolve itself within a further two RTCP reporting intervals. An example of this might be a video conferencing system that backs off to sending audio only, before completely dropping the call.

The choice of two RTCP reporting intervals is to give enough time for transient problems to resolve themselves, but to stop problem flows quickly enough to avoid causing serious problems. A single RTCP report showing no reception could be caused by numerous transient faults, and so should not stop transmission. More than two RTCP reports could avoid false positives, but would lead to problematic flows running for a long time before being cut off.



4.2. RTP/AVP Circuit Breaker #2: Congestion

If RTP data packets are being sent, and the corresponding RTCP RR packets show non-zero packet loss fraction and increasing extended highest sequence number received, then the RTP data packets are arriving at the receiver, but some degree of congestion is occurring. The RTP/AVP profile [RFC3551] states that:

If best-effort service is being used, RTP receivers SHOULD monitor packet loss to ensure that the packet loss rate is within acceptable parameters. Packet loss is considered acceptable if a TCP flow across the same network path and experiencing the same network conditions would achieve an average throughput, measured on a reasonable timescale, that is not less than the RTP flow is achieving. This condition can be satisfied by implementing congestion control mechanisms to adapt the transmission rate (or the number of layers subscribed for a layered multicast session), or by arranging for a receiver to leave the session if the loss rate is unacceptably high.

The comparison to TCP cannot be specified exactly, but is intended as an "order-of-magnitude" comparison in timescale and throughput. The timescale on which TCP throughput is measured is the roundtrip time of the connection. In essence, this requirement states that it is not acceptable to deploy an application (using RTP or any other transport protocol) on the best-effort Internet which consumes bandwidth arbitrarily and does not compete fairly with TCP within an order of magnitude.

The throughput of a long-lived TCP connection can be estimated using the TCP throughput equation:

s X = -------------------------------------------------------------- R*sqrt(2*b*p/3) + (t_RTO * (3*sqrt(3*b*p/8) * p * (1+32*p^2)))

Where:

X is the transmit rate in bytes/second.

s is the packet size in bytes. If the RTP data packets vary in size, then the average size should be used.

R is the round trip time in seconds.



p is the loss event rate, between 0 and 1.0, of the number of loss events as a fraction of the number of packets transmitted.

t_RTO is the TCP retransmission timeout value in seconds, approximated by setting t_RTO = 4*R.

b is the number of packets acknowledged by a single TCP acknowledgement ([RFC3448] recommends the use of b=1 since many TCP implementations do not use delayed acknowledgements).

This is the same approach to estimated TCP throughput that is used in [RFC3448]. Two parameters must be estimated in order to calculate the throughput: the round trip time, R, and the loss event rate, p. The round trip time can be estimated from RTCP. This is done too infrequently for accurate statistics, but is the best that can be done with the standard RTCP mechanisms.

RTCP RR packets contain the packet loss fraction, rather than the loss event rate, so p cannot be reported (TCP typically treats the loss of multiple packets within a single RTT as one loss event, but RTCP RR packets report the overall fraction of packets lost, not caring about when the losses occurred). Using the loss fraction in place of the loss event rate can overestimate the loss. We believe that this overestimate will not be significant, given that we are only interested in order of magnitude comparison (Floyd et al, "Equation-Based Congestion Control for Unicast Applications", Proc. SIGCOMM 2000, section 3.2.1, show that the difference is small for steady-state conditions and random loss, but using the loss fraction is more conservative in the case of bursty loss).

The congestion circuit breaker is therefore: when RTCP RR packets are received, estimate the TCP throughput using the above equation and the measured R, p (approximated by the loss fraction), and s. Compare this with the actual sending rate. If the actual sending rate has been more than an order of magnitude greater than the throughput equation estimate for two or more RTCP reporting intervals, stop transmitting.

Again, we use two reporting intervals to avoid triggering the circuit breaker on transient failures. This circuit breaker is a worst-case condition, and congestion control should be performed to keep well within this bound. It is expected that the circuit breaker will only be triggered if the usual congestion control fails for some reason.

(tbd -- we need to base the circuit breaker condition on something, so TCP seems a logical choice. Following TCP limits too closely is inappropriate for many applications of RTP, though, since they have different dynamics. Is the above lax enough to not disrupt valid



applications, but tight enough to provide meaningful protection for the network?)

5. RTP Circuit Breakers for Systems Using the RTP/AVPF Profile

More rapid feedback allows more responsiveness. The receiver SHOULD provide feedback more often during, or at onset of, congestion, and provide feedback less often when there is no congestion.

(tbd -- mechanisms may probably need to be designed in conjunction with the different classes of congestion control that can leverage RTP/AVPF; e.g., we might need to specify limits for TFRC-like or delay-based algorithms using RTP/AVPF feedback.)

(tbd -- a high-level question to be answered is whether we need to specify anything different for the circuit breaker for AVPF, or if we leave that unchanged, and focus solely on the dynamics, to ensure the circuit breaker is never triggered.)

6. Impact of RTCP XR

(tbd)

This improves the information, but doesn’t change the dynamics of the congestion control loop. Suspect the impact will actually be quite small.

Packets discarded [I-D.ietf-xrblock-rtcp-xr-discard] or bytes discarded [I-D.ietf-xrblock-rtcp-xr-discard-rle-metrics] due to late arrival by the receiver may indicate congestion. Congestion control should consider the discarded packets as if they were lost packets.

The RTCP RR reports the loss fraction over an RTCP interval which is insufficient to distinguish between solitary or bursty losses. To provide rough sense of duration of losses or discards, an endpoint may use burst/gap reporting for loss [I-D.ietf-xrblock-rtcp-xr-burst-gap-loss] and discard [I-D.ietf-xrblock-rtcp-xr-burst-gap-discard]. For more accurate reporting the receiver may use Run-length encoded (RLE) lost [RFC3611] or discarded [I-D.ietf-xrblock-rtcp-xr-discard-rle-metrics] packets.

For precise measurement of network roundtrip delay the receiver can signal its end-system delay [I-D.ietf-xrblock-rtcp-xr-delay] [RFC3611].



A receiver may also indicate onset or end of congestion by reporting the distribution of the inter-packet delay variation [I-D.ietf-xrblock-rtcp-xr-pdv] [RFC3611].

7. Impact of Explicit Congestion Notification (ECN)

ECN-CE marked packets SHOULD be treated as if it were lost for the purposes of congestion control, when determining the optimal rate at which to send. However, it seems unwise to treat the receipt of multiple ECN-CE marked packets as a circuit breaker, since it is likely that ECN-capable and non-ECN-capable paths will exist for a long time to come. Rather, consider packet loss as the circuit breaker condition as for non-ECN flows.

8. Session Timeout

From a usability perspective, if there is no audio or video response from the other peer, it is likely that the user may terminate the session.

According to RFC 3550 [RFC3550], any participant that has not sent an RTP packet within the last two RTCP interval is removed from the sender list. To avoid timing out the specific flow, the endpoint MUST send corresponding RTCP reports. Interactive Connectivity Establishment (ICE) [RFC5245] recommends that the timeout MUST NOT be less than 15 seconds.

If no RTCP RR arrives for two complete SR intervals, the sender SHOULD cease transmission. However, if the endpoint can reduce the media rate then it MAY first reduce the rate to the lower value, but terminate the transmission if still no RTCP RR is received in the next two SR intervals.

9. References



[RFC3448] Handley, M., Floyd, S., Padhye, J., and J. Widmer, "TCP Friendly Rate Control (TFRC): Protocol Specification", RFC 3448, January 2003.

[RFC3550] Schulzrinne, H., Casner, S., Frederick, R., and V.



Jacobson, "RTP: A Transport Protocol for Real-Time Applications", STD 64, RFC 3550, July 2003.

[RFC3551] Schulzrinne, H. and S. Casner, "RTP Profile for Audio and Video Conferences with Minimal Control", STD 65, RFC 3551, July 2003.

[RFC3611] Friedman, T., Caceres, R., and A. Clark, "RTP Control Protocol Extended Reports (RTCP XR)", RFC 3611, November 2003.

[RFC4585] Ott, J., Wenger, S., Sato, N., Burmeister, C., and J. Rey, "Extended RTP Profile for Real-time Transport Control Protocol (RTCP)-Based Feedback (RTP/AVPF)", RFC 4585, July 2006.


[I-D.ietf-avtcore-ecn-for-rtp] Westerlund, M., Johansson, I., Perkins, C., and K. Carlberg, "Explicit Congestion Notification (ECN) for RTP over UDP", draft-ietf-avtcore-ecn-for-rtp-06 (work in progress), February 2012.

[I-D.ietf-xrblock-rtcp-xr-burst-gap-discard] Clark, A., Hunt, G., Wu, W., and R. Huang, "RTCP XR Report Block for Burst/Gap Discard metric Reporting", draft-ietf-xrblock-rtcp-xr-burst-gap-discard-02 (work in progress), January 2012.

[I-D.ietf-xrblock-rtcp-xr-burst-gap-loss] Clark, A., Hunt, G., Zhao, J., Wu, W., and S. Zhang, "RTCP XR Report Block for Burst/Gap Loss metric Reporting", draft-ietf-xrblock-rtcp-xr-burst-gap-loss-01 (work in progress), January 2012.

[I-D.ietf-xrblock-rtcp-xr-delay] Hunt, G., Gross, K., and A. Clark, "RTCP XR Report Block for Delay metric Reporting", draft-ietf-xrblock-rtcp-xr-delay-01 (work in progress), December 2011.

[I-D.ietf-xrblock-rtcp-xr-discard] Hunt, G., Clark, A., Zorn, G., and W. Wu, "RTCP XR Report Block for Discard metric Reporting", draft-ietf-xrblock-rtcp-xr-discard-01 (work in progress), December 2011.



[I-D.ietf-xrblock-rtcp-xr-discard-rle-metrics] Ott, J., Singh, V., and I. Curcio, "Real-time Transport Control Protocol (RTCP) Extension Report (XR) for Run Length Encoding of Discarded Packets", draft-ietf-xrblock-rtcp-xr-discard-rle-metrics-03 (work in progress), February 2012.

[I-D.ietf-xrblock-rtcp-xr-pdv] Hunt, G. and A. Clark, "RTCP XR Report Block for Packet Delay Variation Metric Reporting", draft-ietf-xrblock-rtcp-xr-pdv-02 (work in progress), December 2011.

[RFC3168] Ramakrishnan, K., Floyd, S., and D. Black, "The Addition of Explicit Congestion Notification (ECN) to IP", RFC 3168, September 2001.

[RFC5104] Wenger, S., Chandra, U., Westerlund, M., and B. Burman, "Codec Control Messages in the RTP Audio-Visual Profile with Feedback (AVPF)", RFC 5104, February 2008.

[RFC5245] Rosenberg, J., "Interactive Connectivity Establishment (ICE): A Protocol for Network Address Translator (NAT) Traversal for Offer/Answer Protocols", RFC 5245, April 2010.

[RFC5450] Singer, D. and H. Desineni, "Transmission Time Offsets in RTP Streams", RFC 5450, March 2009.

[RFC6051] Perkins, C. and T. Schierl, "Rapid Synchronisation of RTP Flows", RFC 6051, November 2010.


Colin Perkins University of Glasgow School of Computing Science Glasgow G12 8QQ United Kingdom




Varun Singh Aalto University School of Science and Technology Otakaari 5 A Espoo, FIN 02150 Finland

Email: [email protected] URI: http://www.netlab.tkk.fi/˜varun/


AVT Core Working Group V. SinghInternet-Draft T. KarkkainenIntended status: Experimental J. OttExpires: August 30, 2012 S. Ahsan Aalto University L. Eggert NetApp February 27, 2012

Multipath RTP (MPRTP) draft-singh-avtcore-mprtp-04

Abstract

The Real-time Transport Protocol (RTP) is used to deliver real-time content and, along with the RTP Control Protocol (RTCP), forms the control channel between the sender and receiver. However, RTP and RTCP assume a single delivery path between the sender and receiver and make decisions based on the measured characteristics of this single path. Increasingly, endpoints are becoming multi-homed, which means that they are connected via multiple Internet paths. Network utilization can be improved when endpoints use multiple parallel paths for communication. The resulting increase in reliability and throughput can also enhance the user experience. This document extends the Real-time Transport Protocol (RTP) so that a single session can take advantage of the availability of multiple paths between two endpoints.

Status of this Memo




This Internet-Draft will expire on August 30, 2012.

Copyright Notice

Singh, et al. Expires August 30, 2012 [Page 1]

Internet-Draft Multipath RTP February 2012



Table of Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.1. Requirements Language . . . . . . . . . . . . . . . . . . 5 1.2. Terminology . . . . . . . . . . . . . . . . . . . . . . . 5 1.3. Use-cases . . . . . . . . . . . . . . . . . . . . . . . . 6 2. Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2.1. Functional goals . . . . . . . . . . . . . . . . . . . . . 6 2.2. Compatibility goals . . . . . . . . . . . . . . . . . . . 7 3. RTP Topologies . . . . . . . . . . . . . . . . . . . . . . . . 7 4. MPRTP Architecture . . . . . . . . . . . . . . . . . . . . . . 7 5. Example Media Flow Diagrams . . . . . . . . . . . . . . . . . 9 5.1. Streaming use-case . . . . . . . . . . . . . . . . . . . . 9 5.2. Conversational use-case . . . . . . . . . . . . . . . . . 10 6. MPRTP Functional Blocks . . . . . . . . . . . . . . . . . . . 11 7. Available Mechanisms within the Functional Blocks . . . . . . 12 7.1. Session Setup . . . . . . . . . . . . . . . . . . . . . . 12 7.1.1. Connectivity Checks . . . . . . . . . . . . . . . . . 12 7.1.2. Adding New or Updating Interfaces . . . . . . . . . . 12 7.1.3. In-band vs. Out-of-band Signaling . . . . . . . . . . 12 7.2. Expanding RTP . . . . . . . . . . . . . . . . . . . . . . 14 7.3. Expanding RTCP . . . . . . . . . . . . . . . . . . . . . . 14 7.4. Failure Handling and Teardown . . . . . . . . . . . . . . 14 8. MPRTP Protocol . . . . . . . . . . . . . . . . . . . . . . . . 15 8.1. Overview . . . . . . . . . . . . . . . . . . . . . . . . . 15 8.1.1. Gather or Discovering Candidates . . . . . . . . . . . 16 8.1.2. NAT Traversal . . . . . . . . . . . . . . . . . . . . 16 8.1.3. Choosing between In-band (in RTCP) and Out-of-band (in SDP) Interface Advertisement . . . . . . . . . . . 16 8.1.4. In-band Interface Advertisement . . . . . . . . . . . 17 8.1.5. Subflow ID Assignment . . . . . . . . . . . . . . . . 17 8.1.6. Active and Passive Subflows . . . . . . . . . . . . . 17 8.2. RTP Transmission . . . . . . . . . . . . . . . . . . . . . 18 8.3. Playout Considerations at the Receiver . . . . . . . . . . 18



8.4. Subflow-specific RTCP Statistics and RTCP Aggregation . . 18 8.5. RTCP Transmission . . . . . . . . . . . . . . . . . . . . 19 9. Packet Formats . . . . . . . . . . . . . . . . . . . . . . . . 19 9.1. RTP Header Extension for MPRTP . . . . . . . . . . . . . . 19 9.1.1. MPRTP RTP Extension for a Subflow . . . . . . . . . . 21 9.2. RTCP Extension for MPRTP (MPRTCP) . . . . . . . . . . . . 21 9.2.1. MPRTCP Extension for Subflow Reporting . . . . . . . . 23 9.2.1.1. MPRTCP for Subflow-specific SR, RR and XR . . . . 24 9.3. MPRTCP Extension for Interface advertisement . . . . . . . 26 10. RTCP Timing reconsiderations for MPRTCP . . . . . . . . . . . 27 11. SDP Considerations . . . . . . . . . . . . . . . . . . . . . . 27 11.1. Signaling MPRTP Header Extension in SDP . . . . . . . . . 28 11.2. Signaling MPRTP capability in SDP . . . . . . . . . . . . 28 11.3. MPRTP Interface Advertisement in SDP (out-of-band signaling) . . . . . . . . . . . . . . . . . . . . . . . . 29 11.3.1. "interface" attribute . . . . . . . . . . . . . . . . 29 11.3.2. Example . . . . . . . . . . . . . . . . . . . . . . . 30 11.4. MPRTP with ICE . . . . . . . . . . . . . . . . . . . . . . 30 11.5. Increased Throughput . . . . . . . . . . . . . . . . . . . 31 11.6. Offer/Answer . . . . . . . . . . . . . . . . . . . . . . . 31 11.6.1. In-band Signaling Example . . . . . . . . . . . . . . 32 11.6.2. Out-of-band Signaling Example . . . . . . . . . . . . 32 11.6.2.1. Without ICE . . . . . . . . . . . . . . . . . . . 32 11.6.2.2. With ICE . . . . . . . . . . . . . . . . . . . . . 33 12. MPRTP in RTSP . . . . . . . . . . . . . . . . . . . . . . . . 35 12.1. Solution Overview without ICE . . . . . . . . . . . . . . 35 12.2. Solution Overview with ICE . . . . . . . . . . . . . . . . 37 12.3. RTSP Extensions . . . . . . . . . . . . . . . . . . . . . 39 12.3.1. MPRTP Interface Transport Header Parameter . . . . . . 39 12.3.2. MPRTP Feature Tag . . . . . . . . . . . . . . . . . . 40 12.3.3. Status Codes . . . . . . . . . . . . . . . . . . . . . 40 12.3.4. New Reason for PLAY_NOTIFY . . . . . . . . . . . . . . 40 12.3.5. re-SETUP . . . . . . . . . . . . . . . . . . . . . . . 41 13. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 41 13.1. MPRTP Header Extension . . . . . . . . . . . . . . . . . . 42 13.2. MPRTCP Packet Type . . . . . . . . . . . . . . . . . . . . 42 13.3. SDP Attributes . . . . . . . . . . . . . . . . . . . . . . 43 13.3.1. "mprtp" attribute . . . . . . . . . . . . . . . . . . 43 13.4. RTSP . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 13.4.1. RTSP Feature Tag . . . . . . . . . . . . . . . . . . . 44 13.4.2. RTSP Transport Parameters . . . . . . . . . . . . . . 44 13.4.3. Notify-Reason value . . . . . . . . . . . . . . . . . 44 14. Security Considerations . . . . . . . . . . . . . . . . . . . 44 15. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 44 16. References . . . . . . . . . . . . . . . . . . . . . . . . . . 45 16.1. Normative References . . . . . . . . . . . . . . . . . . . 45 16.2. Informative References . . . . . . . . . . . . . . . . . . 46 Appendix A. Interoperating with Legacy Applications . . . . . . . 46



Appendix B. Change Log . . . . . . . . . . . . . . . . . . . . . 47 B.1. changes in draft-singh-avtcore-mprtp-04 . . . . . . . . . 47 B.2. changes in draft-singh-avtcore-mprtp-03 . . . . . . . . . 47 B.3. changes in draft-singh-avtcore-mprtp-02 . . . . . . . . . 48 B.4. changes in draft-singh-avtcore-mprtp-01 . . . . . . . . . 48 Authors’ Addresses . . . . . . . . . . . . . . . . . . . . . . . . 48



1. Introduction

Multi-homed endpoints are becoming common in today’s Internet, e.g., devices that support multiple wireless access technologies such as 3G and Wireless LAN. This means that there is often more than one network path available between two endpoints. Transport protocols, such as RTP, have not been designed to take advantage of the availability of multiple concurrent paths and therefore cannot benefit from the increased capacity and reliability that can be achieved by pooling their respective capacities.

Multipath RTP (MPRTP) is an OPTIONAL extension to RTP [1] that allows splitting a single RTP stream into multiple subflows that are transmitted over different paths. In effect, this pools the resource capacity of multiple paths. Multipath RTCP (MPRTCP) is an extension to RTCP, it is used along with MPRTP to report per-path sender and receiver characteristics.

Other IETF transport protocols that are capable of using multiple paths include SCTP [11], MPTCP [12] and SHIM6 [13]. However, these protocols are not suitable for realtime communications.

1.1. Requirements Language

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [2].

1.2. Terminology

o Endpoint: host either initiating or terminating an RTP flow.

o Interface: logical or physical component that is capable of acquiring a unique IP address.

o Path: sequence of links between a sender and a receiver. Typically, defined by a set of source and destination addresses.

o Subflow: flow of RTP packets along a specific path, i.e., a subset of the packets belonging to an RTP stream. The combination of all RTP subflows forms the complete RTP stream. Typically, a subflow would map to a unique path, i.e., each combination of IP addresses and port pairs (5-tuple, including protocol) is a unique subflow.



1.3. Use-cases

The primary use-case for MPRTP is transporting high bit-rate streaming multimedia content between endpoints, where at least one is multi-homed. Such endpoints could be residential IPTV devices that connect to the Internet through two different Internet service providers (ISPs), or mobile devices that connect to the Internet through 3G and WLAN interfaces. By allowing RTP to use multiple paths for transmission, the following gains can be achieved:

o Higher quality: Pooling the resource capacity of multiple Internet paths allows higher bit-rate and higher quality codecs to be used. From the application perspective, the available bandwidth between the two endpoints increases.

o Load balancing: Transmitting an RTP stream over multiple paths reduces the bandwidth usage on a single path, which in turn reduces the impact of the media stream on other traffic on that path.

o Fault tolerance: When multiple paths are used in conjunction with redundancy mechanisms (FEC, re-transmissions, etc.), outages on one path have less impact on the overall perceived quality of the stream.

A secondary use-case for MPRTP is transporting Voice over IP (VoIP) calls to a device with multiple interfaces. Again, such an endpoint could be a mobile device with multiple wireless interfaces. In this case, little is to be gained from resource pooling, i.e., higher capacity or load balancing, because a single path should be easily capable of handling the required load. However, using multiple concurrent subflows can improve fault tolerance, because traffic can shift between the subflows when path outages occur. This results in very fast transport-layer handovers that do not require support from signaling.

2. Goals

This section outlines the basic goals that multipath RTP aims to meet. These are broadly classified as Functional goals and Compatibility goals.

2.1. Functional goals

Allow unicast RTP session to be split into multiple subflows in order to be carried over multiple paths. This may prove beneficial in case of video streaming.



o Increased Throughput: Cumulative capacity of the two paths may meet the requirements of the multimedia session. Therefore, MPRTP MUST support concurrent use of the multiple paths.

o Improved Reliability: MPRTP SHOULD be able to send redundant packets or re-transmit packets along any available path to increase reliability.

The protocol SHOULD be able to open new subflows for an existing session when new paths appear and MUST be able to close subflows when paths disappear.

2.2. Compatibility goals

MPRTP MUST be backwards compatible; an MPRTP stream needs to fall back to be compatible with legacy RTP stacks if MPRTP support is not successfully negotiated.

o Application Compatibility: MPRTP service model MUST be backwards compatible with existing RTP applications, i.e., an MPRTP stack MUST be able to work with legacy RTP applications and not require changes to them. Therefore, the basic RTP APIs MUST remain unchanged, but an MPRTP stack MAY provide extended APIs so that the application can configure any additional features provided by the MPRTP stack.

o Network Compatibility: individual RTP subflows MUST themselves be well-formed RTP flows, so that they are able to traverse NATs and firewalls. This MUST be the case even when interfaces appear after session initiation. Interactive Connectivity Establishment (ICE) [3] MAY be used for discovering new interfaces or performing connectivity checks.

3. RTP Topologies

RFC 5117 [14] describes a number of scenarios using mixers and translators in single-party (point-to-point), and multi-party (point- to-multipoint) scenarios. RFC 3550 [1] (Section 2.3 and 7.x) discuss in detail the impact of mixers and translators on RTP and RTCP packets. MPRTP assumes that if a mixer or translator exists in the network, then either all of the multiple paths or none of the multiple paths go via this component.

4. MPRTP Architecture

In a typical scenario, an RTP session uses a single path. In an



MPRTP scenario, an RTP session uses multiple subflows that each use a different path. Figure 1 shows the difference.

+--------------+ Signaling +--------------+ | |------------------------------------>| | | Client |<------------------------------------| Server | | | Single RTP flow | | +--------------+ +--------------+

+--------------+ Signaling +--------------+ | |------------------------------------>| | | Client |<------------------------------------| Server | | |<------------------------------------| | +--------------+ MPRTP subflows +--------------+

Figure 1: Comparison between traditional RTP streaming and MPRTP

+-----------------------+ +-------------------------------+ | Application | | Application | +-----------------------+ +-------------------------------+ | | | MPRTP | + RTP + +- - - - - - - -+- - - - - - - -+ | | | RTP subflow | RTP subflow | +-----------------------+ +---------------+---------------+ | UDP | | UDP | UDP | +-----------------------+ +---------------+---------------+ | IP | | IP | IP | +-----------------------+ +---------------+---------------+

Figure 2: MPRTP Architecture

Figure 2 illustrates the differences between the standard RTP stack and the MPRTP stack. MPRTP receives a normal RTP session from the application and splits it into multiple RTP subflows. Each subflow is then sent along a different path to the receiver. To the network, each subflow appears as an independent, well-formed RTP flow. At the receiver, the subflows are combined to recreate the original RTP session. The MPRTP layer performs the following functions:

o Path Management: The layer is aware of alternate paths to the other host, which may, for example, be the peer’s multiple interfaces. This enables the endpoint to transmit differently marked packets along separate paths. MPRTP also selects interfaces to send and receive data. Furthermore, it manages the port and IP address pair bindings for each subflow.



o Packet Scheduling: the layer splits a single RTP flow into multiple subflows and sends them across multiple interfaces (paths). The splitting MAY BE done using different path characteristics.

o Subflow recombination: the layer creates the original stream by recombining the independent subflows. Therefore, the multipath subflows appear as a single RTP stream to applications.

5. Example Media Flow Diagrams

There may be many complex technical scenarios for MPRTP, however, this memo only considers the following two scenarios: 1) a unidirectional media flow that represents the streaming use-case, and 2) a bidirectional media flow that represents a conversational use- case.

5.1. Streaming use-case

In the unidirectional scenario, the receiver (client) initiates a multimedia session with the sender (server). The receiver or the sender may have multiple interfaces and both endpoints are MPRTP- capable. Figure 3 shows this scenario. In this case, host A has multiple interfaces. Host B performs connectivity checks on host A’s multiple interfaces. If the interfaces are reachable, then host B streams multimedia data along multiple paths to host A. Moreover, host B also sends RTCP Sender Reports (SR) for each subflow and host A responds with a normal RTCP Receiver Report (RR) for the overall session as well as the receiver statistics for each subflow. Host B distributes the packets across the subflows based on the individually measured path characteristics.

Alternatively, to reduce media startup time, host B may start streaming multimedia data to host A’s initiating interface and then perform connectivity checks for the other interfaces. This method of updating a single path session to a multipath session is called "multipath session upgrade".



Host A Host B ----------------------- ---------- Interface A1 Interface A2 Interface B1 ----------------------- ---------- | | |------------------------------------->| Session setup with |<-------------------------------------| multiple interfaces | | | | | | | (RTP data B1->A1, B1->A2) | |<=====================================| | |<========================| | | | Note: there may be more scenarios.

Figure 3: Unidirectional media flow

5.2. Conversational use-case

In the bidirectional scenario, multimedia data flows in both directions. The two hosts exchange their lists of interfaces with each other and perform connectivity checks. Communication begins after each host finds suitable address, port pairs. Interfaces that receive data send back RTCP receiver statistics for that path (based on the 5-tuple). The hosts balance their multimedia stream across multiple paths based on the per path reception statistics and its own volume of traffic. Figure 4 describes an example of a bidirectional flow.

Host A Host B --------------------------- --------------------------- Interface A1 Interface A2 Interface B1 Interface B2 --------------------------- --------------------------- | | | | | | | |Session setup |----------------------------------->| |with multiple |<-----------------------------------| |interfaces | | | | | | | | | (RTP data B1<->A1, B2<->A2) | | |<===================================| | | |<===================================| |===================================>| | | |===================================>| | | | | Note: the address pairs may have other permutations, and they may be symmetric or asymmetric combinations.



Figure 4: Bidirectional flow

6. MPRTP Functional Blocks

This section describes some of the functional blocks needed for MPRTP. We then investigate each block and consider available mechanisms in the next section.

1. Session Setup: Interfaces may appear or disappear at anytime during the session. To preserve backward compatibility with legacy applications, a multipath session MUST look like a bundle of individual RTP sessions. A multipath session may be upgraded from a typical single path session, as and when new interfaces appear or disappear. However, it is also possible to setup a multipath session from the beginning, if the interfaces are available at the start of the multimedia session.

2. Expanding RTP: For a multipath session, each subflow MUST look like an independent RTP flow, so that individual RTCP messages can be generated per subflow. Furthermore, MPRTP splits the single multimedia stream into multiple subflows based on path characteristics (e.g. RTT, loss-rate, receiver rate, bandwidth- delay product etc.) and dynamically adjusts the load on each link.

3. Adding Interfaces: Interfaces on the host need to be regularly discovered and advertised. This can be done at session setup and/or during the session. Discovering interfaces is outside the scope of this document.

4. Expanding RTCP: MPRTP MUST provide per path RTCP reports for monitoring the quality of the path, for load balancing, or for congestion control, etc. To maintain backward compatibility with legacy applications, the receiver MUST also send aggregate RTCP reports along with the per-path reports.

5. Maintenance and Failure Handling: In a multi-homed endpoint interfaces may appear and disappear. If this occurs at the sender, it has to re-adjust the load on the available links. On the other hand, if this occurs at the receiver, then the multimedia data transmitted by the sender to those interfaces is lost. This data may be re-transmitted along a different path i.e., to a different interface on the receiver. Furthermore, the endpoint has to either explicitly signal the disappearance of an interface, or the other endpoint has to detect it (by lack of media packets, RTCP feedback, or keep-alive packets).



6. Teardown: The MPRTP layer releases the occupied ports on the interfaces.

7. Available Mechanisms within the Functional Blocks

This section discusses some of the possible alternatives for each functional block mentioned in the previous section.

7.1. Session Setup

MPRTP session can be set up in many possible ways e.g., during handshake, or upgraded mid-session. The capability exchange may be done using out-of-band signaling (e.g., Session Description Protocol (SDP) [15] in Session Initiation Protocol (SIP) [16], Real-Time Streaming Protocol (RTSP) [17]) or in-band signaling (e.g., RTP/RTCP header extension, Session Traversal Utilities for NAT (STUN) messages).

7.1.1. Connectivity Checks

The endpoint SHOULD be capable of performing connectivity checks (e.g., using ICE [3]). If the IP addresses of the endpoints are reachable (e.g., globally addressable, same network etc) then connectivity checks are not needed.

7.1.2. Adding New or Updating Interfaces

Interfaces can appear and disappear during a session, the endpoint MUST be capable of advertising the changes in its set of usable interfaces. Additionally, the application or OS may define a policy on when and/or what interfaces are usable. However, MPRTP requires a method to advertise or notify the other endpoint about the updated set of usable interfaces.

7.1.3. In-band vs. Out-of-band Signaling

MTRTP nodes will generally use a signaling protocol to establish their MPRTP session. With the existence of such a signaling relationship, two alternatives become available to exchange information about the available interfaces on each side for extending RTP sessions to MPRTP and for modifying MPRTP sessions: in-band and out-of-band signaling.

In-band signaling refers to using mechanisms of RTP/RTCP itself to communicate interface addresses, e.g., a dedicated RTCP extensions along the lines of the one defined to communicate information about the feedback target for RTP over SSM [4]. In-band signaling does not



rely on the availability of a separate signaling connection and the information flows along the same path as the media streams, thus minimizing dependencies. Moreover, if the media channel is secured (e.g., by means of SRTP), the signaling is implicitly protected as well if SRTCP encryption and authentication are chosen. In-band signaling is also expected to take a direct path to the peer, avoiding any signaling overlay-induced indirections and associated processing overheads in signaling elements -- avoiding such may be especially worthwhile if frequent updates may occur as in the case of mobile users. Finally, RTCP is usually sent sufficiently frequently (in point-to-point settings) to provide enough opportunities for transmission and (in case of loss) retransmission of the corresponding RTCP packets.

Examples for in-band signaling include RTCP extensions as noted above or suitable extensions to STUN.

Out-of-band signaling refers to using a separate signaling connection (via SIP, RTSP, or HTTP) to exchange interface information, e.g., expressed in SDP. Clear benefits are that signaling occurs at setup time anyway and that experience and SDP syntax (and procedures) are available that can be re-used or easily adapted to provide the necessary capabilities. In contrast to RTCP, SDP offers a reliable communication channel so that no separate retransmissions logic is necessary. In SDP, especially when combined with ICE, connectivity check mechanisms including sophisticated rules are readily available. While SDP is not inherently protected, suitable security may need to be applied anyway to the basic session setup.

Examples for out-of-band signaling are dedicated extensions to SDP; those may be combined with ICE.

Both types of mechanisms have their pros and cons for middleboxes. With in-band signaling, control packets take the same path as the media packets and they can be directly inspected by middleboxes so that the elements operating on the signaling channel do not need to understand new SDP. With out-of-band signaling, only the middleboxes processing the signaling need to be modified and those on the data forwarding path can remain untouched.

Overall, it may appear sensible to provide a combination of both mechanisms: out-of-band signaling for session setup and initial interface negotiation and in-band signaling to deal with frequent changes in interface state (and for the potential case, albeit rather theoretical case of MPRTP communication without a signaling channel).

In its present version, this document explores both options to provide a broad understanding of how the corresponding mechanisms



would look like.

[[Comment.1: Some have suggested STUN may be suitable for doing in- band interface advertisement. This is still under consideration, but depends on implementation challenges as many legacy systems don’t implement STUN and many RTP systems ignore STUN messages. --Editor]]

7.2. Expanding RTP

RTCP [1] is generated per media session. However, with MPRTP, the media sender spreads the RTP load across several interfaces. The media sender SHOULD make the path selection, load balancing and fault tolerance decisions based on the characteristics of each path. Therefore, apart from normal RTP sequence numbers defined in [1], the MPRTP sender MUST add subflow-specific sequence numbers to help calculate fractional losses, jitter, RTT, playout time, etc., for each path, and a subflow identifier to associate the characteristics with a path. The RTP header extension for MPRTP is shown in Section 9.1.

7.3. Expanding RTCP

To provide accurate per path information an MPRTP endpoint MUST send (SR/RR) report for each unique subflow along with the overall session RTCP report. Therefore, the additional subflow reporting affects the RTCP bandwidth and the RTCP reporting interval. RTCP report scheduling for each subflow may cause a problem for RTCP recombination and reconstruction in cases when 1) RTCP for a subflow is lost, and 2) RTCP for a subflow arrives later than the other subflows. (There may be other cases as well.)

The sender distributes the media across different paths using the per path RTCP reports. However, this document doesn’t cover algorithms for congestion control or load balancing.

7.4. Failure Handling and Teardown

An MPRTP endpoint MUST keep alive subflows that have been negotiated but no media is sent on them. Moreover, using the information in the subflow reports, a sender can monitor an active subflow for failure (errors, unreachability, congestion) and decide not to use (make the active subflow passive), or teardown the subflow.

If an interface disappears, the endpoint MUST send an updated interface advertisement without the interface and release the the associated subflows.



8. MPRTP Protocol

Host A Host B ----------------------- ----------------------- Interface A1 Interface A2 Interface B1 Interface B2 ----------------------- ----------------------- | | | | | | (1) | | |--------------------------------------->| | |<---------------------------------------| | | | (2) | | |<=======================================| | |<=======================================| (3) | | | (4) | | |<- - - - - - - - - - - - - - - - - - - -| | |<- - - - - - - - - - - - - - - - - - - -| | |<- - - - - - - - - - - - - - - - - - - -| | | | (5) | | |- - - - - - - - - - - - - - - - - - - ->| | |<=======================================| (6) | |<=======================================| | | |<========================================| |<=======================================| | | |<========================================|

Key: | Interface ---> Signaling Protocol <=== RTP Packets - -> RTCP Packet

Figure 5: MPRTP New Interface

8.1. Overview

The bullet points explain the different steps shown in Figure 5 for upgrading a single path multimedia session to multipath session.

(1) The first two interactions between the hosts represents the establishment of a normal RTP session. This may performed e.g. using SIP or RTSP.

(2) When the RTP session has been established, host B streams media using its interface B1 to host A at interface A1.

(3) Host B supports sending media using MPRTP and becomes aware of an additional interface B2.



(4) Host B advertises the multiple interface addresses.

(5) Host A supports receiving media using MPRTP and becomes aware of an additional interface A2.

Side note, even if an MPRTP-capable host has only one interface, it MUST respond to the advertisement with its single interface.

(6) Each host receives information about the additional interfaces and the appropriate endpoints starts to stream the multimedia content using the additional paths.

If needed, each endpoint will need to independently perform connectivity checks (not shown in figure) and ascertain reachability before using the paths.

8.1.1. Gather or Discovering Candidates

The endpoint periodically polls the operating system or is notified when an additional interface appears. If the endpoint wants to use the additional interface for MPRTP it MUST advertise it to the other peers. The endpoint may also use ICE [3] to gather additional candidates.

8.1.2. NAT Traversal

After gathering their interface candidates, the endpoints decide internally if they wish to perform connectivity checks.

[[Comment.2: Legacy applications do not require ICE for session establishment, therefore, MPRTP should not require it as well. --Editor]]

If the endpoint chooses to perform connectivity checks then it MUST first advertise the gathered candidates as ICE candidates in SDP during session setup and let ICE perform the connectivity checks. As soon as a sufficient number of connectivity checks succeed, the endpoint can use the successful candidates to advertise its MPRTP interface candidates.

8.1.3. Choosing between In-band (in RTCP) and Out-of-band (in SDP) Interface Advertisement

If there is no media flowing at the moment and the application wants to use the interfaces from the start of the session, it should advertise them in SDP (See Section 11.3). Alternatively, the endpoint can setup the session as a single path media session and upgrade the session to multipath by advertising the session in-band



(See Section 8.1.4). Moreover, if an interface appears and disappears, the endpoint SHOULD prefer to advertise it in-band because the endpoint would not have to wait for a response from the other endpoint before starting to use the interface. However, if there is a conflict between an in-band and out-of-band advertisement, i.e., the endpoint receives an in-band advertisement while it has a pending out-of-band advertisement, or vice versa then the session is setup using out-of-band mechanisms.

8.1.4. In-band Interface Advertisement

To advertise the multiple interfaces in RTCP, an MPRTP-capable endpoint MUST add the MPRTP Interface Advertisement defined in Figure 13 with the RTCP Sender Report (SR). Each unique address is encapsulated in an Interface Advertisement block and contains the IP address, RTP and RTCP port addresses. The Interface Advertisement blocks are ordered based on a decreasing priority level. On receiving the MPRTP Interface Advertisement, an MPRTP-capable receiver MUST respond with the set of interfaces (subset or all available) it wants to use.

If the sender and receiver have only one interface, then the endpoints MUST indicate the negotiated single path IP, RTP port and RTCP port addresses.

8.1.5. Subflow ID Assignment

After interface advertisements have been exchanged, the endpoint MUST associate a Subflow ID to each unique subflow. Each combination of sender and receiver IP addresses and port pairs (5-tuple) is a unique subflow. If the connectivity checks have been performed then the endpoint MUST only use the subflows for which the connectivity checks have succeeded.

8.1.6. Active and Passive Subflows

To send and receive data an endpoint MAY use any number of subflows from the set of available subflows. The subflows that carry media data are called active subflows, while those subflows that don’t send any media packets (fallback paths) are called passive subflows.

An endpoint MUST multiplex the subflow specific RTP and RTCP packets on the same port to keep the NAT bindings alive. This is inline with the recommendation made in RFC6263[18]. Moreover, if an endpoint uses ICE, multiplexing RTP and RTCP reduces the number of components from 2 to 1 per media stream. If no MPRTP or MPRTCP packets are received on a particular subflow at a receiver, the receiver SHOULD remove the subflow from active and passive lists and not send any



MPRTCP reports for that subflow. It may keep the bindings in a dead- pool, so that if the 5-tuple or subflow reappears, it can quickly reallocate it based on past history.

8.2. RTP Transmission

If both endpoints are MPRTP-capable and if they want to use their multiple interfaces for sending the media stream then they MUST use the MPRTP header extensions. They MAY use normal RTP with legacy endpoints (see Appendix A).

An MPRTP endpoint sends RTP packets with an MPRTP extension that maps the media packet to a specific subflow (see Figure 8). The MPRTP layer SHOULD associate an RTP packet with a subflow based on a scheduling strategy. The scheduling strategy may either choose to augment the paths to create higher throughput or use the alternate paths for enhancing resilience or error-repair. Due to the changes in path characteristics, the endpoint should be able change its scheduling strategy during an ongoing session. The MPRTP sender MUST also populate the subflow specific fields described in the MPRTP extension header (see Section 9.1.1).

8.3. Playout Considerations at the Receiver

A media receiver, irrespective of MPRTP support or not, should be able to playback the media stream because the received RTP packets are compliant to [1], i.e., a non-MPRTP receiver will ignore the MPRTP header and still be able to playback the RTP packets. However, the variation of jitter and loss per path may affect proper playout. The receiver can compensate for the jitter by modifying the playout delay (i.e., by calculating skew across all paths) of the received RTP packets.

8.4. Subflow-specific RTCP Statistics and RTCP Aggregation

Aggregate RTCP provides the overall media statistics and follows the normal RTCP defined in RFC3550 [1]. However, subflow specific RTCP provides the per path media statistics because the aggregate RTCP report may not provide sufficient per path information to an MPRTP scheduler. Specifically, the scheduler should be aware of each path’s RTT and loss-rate, which an aggregate RTCP cannot provide. The sender/receiver MUST use non-compound RTCP reports defined in RFC5506 [5] to transmit the aggregate and subflow-specific RTCP reports. Also, each subflow and the aggregate RTCP report MUST follow the timing rules defined in [6].

The RTCP reporting interval is locally implemented and the scheduling of the RTCP reports may depend on the the behavior of each path. For



instance, the RTCP interval may be different for a passive path than an active path to keep port bindings alive. Additionally, an endpoint may decide to share the RTCP reporting bit rate equally across all its paths or schedule based on the receiver rate on each path.

8.5. RTCP Transmission

The sender sends an RTCP SR on each active path. For each SR the receiver gets, it echoes one back to the same IP address-port pair that sent the SR. The receiver tries to choose the symmetric path and if the routing is symmetric then the per-path RTT calculations will work out correctly. However, even if the paths are not symmetric, the sender would at maximum, under-estimate the RTT of the path by a factor of half of the actual path RTT.

9. Packet Formats

In this section we define the protocol structures described in the previous sections.

9.1. RTP Header Extension for MPRTP

The MPRTP header extension is used to distribute a single RTP stream over multiple subflows.

The header conforms to the one-byte RTP header extension defined in [7]. The header extension contains a 16-bit length field that counts the number of 32-bit words in the extension, excluding the four-octet extension header (therefore zero is a valid length, see Section 5.3.1 of [1] for details).

The RTP header for each subflow is defined below:



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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |V=2|P|1| CC |M| PT | sequence number | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | timestamp | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | synchronization source (SSRC) identifier | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ | 0xBE | 0xDE | length=N-1 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | ID | LEN | MPID |LENGTH | MPRTP header data | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + | .... | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ | RTP payload | | .... | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

Figure 6: Generic MPRTP header extension

MPID:

The MPID field corresponds to the type of MPRTP packet. Namely:

+---------------+--------------------------------------------------+ | MPID ID | Use | | Value | | +---------------+--------------------------------------------------+ | 0x0 | Subflow RTP Header. For this case the Length is | | | set to 6 | +---------------+--------------------------------------------------+

Figure 7: RTP header extension values for MPRTP (H-Ext ID)

length

The 4-bit length field is the length of extension data in bytes not including the H-Ext ID and length fields. The value zero indicates there is no data following.

MPRTP header data

Carries the MPID specific data as described in the following sub-sections.



9.1.1. MPRTP RTP Extension for a Subflow

The RTP header for each subflow is defined below:

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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |V=2|P|1| CC |M| PT | sequence number | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | timestamp | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | synchronization source (SSRC) identifier | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ | 0xBE | 0xDE | length=3 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | ID | LEN=4 | 0x0 | LEN=4 | Subflow ID | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Subflow-specific Seq Number | Pad (0) | Pad (0) | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ | RTP payload | | .... | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

Figure 8: MPRTP header for subflow

MP ID = 0x0

Indicates that the MPRTP header extension carries subflow specific information.

length = 4

Subflow ID: Identifier of the subflow. Every RTP packet belonging to the same subflow carries the same unique subflow identifier.

Flow-Specific Sequence Number (FSSN): Sequence of the packet in the subflow. Each subflow has its own strictly monotonically increasing sequence number space.

9.2. RTCP Extension for MPRTP (MPRTCP)

The MPRTP RTCP header extension is used to 1) provide RTCP feedback per subflow to determine the characteristics of each path, and 2) advertise each interface.



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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |V=2|P|reserved | PT=MPRTCP=211 | length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | SSRC of packet sender | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ | SSRC_1 (SSRC of first source) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | MPRTCP_Type | block length | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ MPRTCP Reports | | ... | | ... | | ... | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+

Figure 9: Generic RTCP Extension for MPRTP (MPRTCP) [appended to normal SR/RR]

MPRTCP: 8 bits

Contains the constant 211 to identify this as an Multipath RTCP packet.

length: 16 bits

As described for the RTCP packet (see Section 6.4.1 of the RTP specification [1]), the length of this is in 32-bit words minus one, including the header and any padding.

MPRTCP_Type: 8-bits

The MPRTCP_Type field corresponds to the type of MPRTP RTCP packet. Namely:

+---------------+--------------------------------------------------+ | MPRTCP_Type | Use | | Value | | +---------------+--------------------------------------------------+ | 0 | Subflow Specific Report | | 1 | Interface Advertisement (IPv4 Address) | | 2 | Interface Advertisement (IPv4 Address) | | 3 | Interface Advertisement (DNS Address) | +---------------+--------------------------------------------------+

Figure 10: RTP header extension values for MPRTP (MPRTCP_Type)



block length: 8-bits

The 8-bit length field is the length of the encapsulated MPRTCP reports in 32-bit word length not including the MPRTCP_Type and length fields. The value zero indicates there is no data following.

MPRTCP Reports: variable size

Defined later in 9.2.1 and 9.3.

9.2.1. MPRTCP Extension for Subflow Reporting

When sending a report for a specific subflow the sender or receiver MUST add only the reports associated with that 5-tuple. Each subflow is reported independently using the following MPRTCP Feedback header.

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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |V=2|P|reserved | PT=MPRTCP=211 | length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | SSRC of packet sender | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ | SSRC_1 (SSRC of first source) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | MPRTCP_Type=0 | block length | Subflow ID #1 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Subflow-specific reports | | .... | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | MPRTCP_Type=0 | block length | Subflow ID #2 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Subflow-specific reports | | .... | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

Figure 11: MPRTCP Subflow Reporting Header

MPRTCP_Type: 0

The value indicates that the encapsulated block is a subflow report.

block length: 8-bits

The 8-bit length field is the length of the encapsulated subflow- specific reports in 32-bit word length not including the



MPRTCP_Type and length fields.

Subflow ID: 16 bits

Subflow identifier is the value associated with the subflow the endpoint is reporting about. If it is a sender it MUST use the Subflow ID associated with the 5-tuple. If it is a receiver it MUST use the Subflow ID received in the Subflow-specific Sender Report.

Subflow-specific reports: variable

Subflow-specific report contains all the reports associated with the Subflow ID. For a sender, it MUST include the Subflow- specific Sender Report (SSR). For a receiver, it MUST include Subflow-specific Receiver Report (SRR). Additionally, if the receiver supports subflow-specific extension reports then it MUST append them to the SRR.

9.2.1.1. MPRTCP for Subflow-specific SR, RR and XR

[[Comment.3: inside the context of subflow specific reports can we reuse the payload type code for Sender Report (PT=200), Receiver Report (PT=201), Extension Report (PT=207).Transport and Payload specific RTCP messages are session specific and SHOULD be used as before. --Editor]]

Example:



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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |V=2|P|reserved | PT=MPRTCP=211 | length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | SSRC of packet sender | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ | SSRC_1 (SSRC of first source) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | MPRTCP_Type=0 | block length | Subflow ID #1 | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ |V=2|P| RC | PT=SR=200 | length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | SSRC of sender | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ | NTP timestamp, most significant word | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | NTP timestamp, least significant word | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | RTP timestamp | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | subflow’s packet count | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | subflow’s octet count | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | MPRTCP_Type=0 | block length | Subflow ID #2 | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ |V=2|P| RC | PT=RR=201 | length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | SSRC of packet sender | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ | fraction lost | cumulative number of packets lost | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | extended highest sequence number received | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | inter-arrival jitter | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | last SR (LSR) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | delay since last SR (DLSR) | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ | Subflow specific extension reports | | .... | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

Figure 12: Example of reusing RTCP SR and RR inside an MPRTCP header (Bi-directional use-case, in case of uni-directional flow the subflow will only send an SR or RR).



9.3. MPRTCP Extension for Interface advertisement

This sub-section defines the RTCP header extension for in-band interface advertisement by the receiver. The interface advertisement block describes a method to represent IPv4, IPv6 and generic DNS-type addresses in a block format. It is based on the sub-reporting block in [4]. The interface advertisement SHOULD immediately follow the Receiver Report. If the Receiver Report is not present, then it MUST be appended to the Sender Report.

The endpoint MUST advertise the interfaces it wants to use whenever an interface appears or disappears and also when it receives an Interface Advertisement.

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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |V=2|P|reserved | PT=MPRTCP=211 | length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | SSRC of packet sender | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ | SSRC_1 (SSRC of first source) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | MPRTCP_Type | block length | RTP Port | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Interface Address #1 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | MPRTCP_Type | block length | RTP Port | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Interface Address #2 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | MPRTCP_Type | block length | RTP Port | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Interface Address #.. | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | MPRTCP_Type | block length | RTP Port | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Interface Address #m | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+

Figure 13: MPRTP Interface Advertisement. (appended to SR/RR)

MPRTCP_Type: 8 bits

The MPRTCP_Type corresponds to the type of interface address. Namely:



1: IPv4 address

2: IPv6 address

3: DNS name

block length: 8 bits

The length of the Interface Advertisement block in bytes.

For an IPv4 address, this should be 9 (i.e., 5 octets for the header and 4 octets for IPv4 address).

For an IPv6 address, this should be 21.

For a DNS name, the length field indicates the number of octets making up the string plus the 5 byte header.

RTP Port: 2 octets

The port number to which the sender sends RTP data. A port number of 0 is invalid and MUST NOT be used.

Interface Address: 4 octets (IPv4), 16 octets (IPv6), or n octets (DNS name)

The address to which receivers send feedback reports. For IPv4 and IPv6, fixed-length address fields are used. A DNS name is an arbitrary-length string. The string MAY contain Internationalizing Domain Names in Applications (IDNA) domain names and MUST be UTF-8 [8] encoded.

10. RTCP Timing reconsiderations for MPRTCP

MPRTP endpoints MUST conform to the timing rule imposed in [6], i.e., the total RTCP rate between the participants MUST NOT exceed 5% of the media rate. For each endpoint, a subflow MUST send the aggregate and subflow-specific report. The endpoint SHOULD schedule the RTCP reports for the active subflows based on the share of the transmitted or received bit rate to the average media bit rate, this method ensures fair sharing of the RTCP bandwidth. Alternatively, the endpoint MAY schedule the reports in round-robin.

11. SDP Considerations



11.1. Signaling MPRTP Header Extension in SDP

To indicate the use of the MPRTP header extensions (see Section 9) in SDP, the sender MUST use the following URI in extmap: urn:ietf:params:rtp-hdrext:mprtp. This is a media level parameter. Legacy RTP (non-MPRTP) clients will ignore this header extension, but can continue to parse and decode the packet (see Appendix A).

Example:

v=0 o=alice 2890844526 2890844527 IN IP4 192.0.2.1 s= c=IN IP4 192.0.2.1 t=0 0 m=video 49170 RTP/AVP 98 a=rtpmap:98 H264/90000 a=fmtp:98 profile-level-id=42A01E; a=extmap:1 urn:ietf:params:rtp-hdrext:mprtp

11.2. Signaling MPRTP capability in SDP

A participant of a media session MUST use SDP to indicate that it supports MPRTP. Not providing this information will make the other endpoint ignore the RTCP extensions.

mprtp-attrib = "a=" "mprtp" [ SP mprtp-optional-parameter] CRLF ; flag to enable MPRTP

The endpoint MUST use ’a=mprtp’, if it is able to send and receive MPRTP packets. Generally, senders and receivers MUST indicate this capability if they support MPRTP and would like to use it in the specific media session being signaled. To exchange the additional interfaces, the endpoint SHOULD use the in-band signaling (Section 9.3). Alternatively, advertise in SDP (Section 11.3).

MPRTP endpoint multiplexes RTP and RTCP on a single port, sender MUST indicate support by adding "a=rtcp-mux" in SDP. If an endpoint receives an SDP without "a=rtcp-mux" but contains "a=mprtp", then the endpoint MUST infer support for multiplexing.

[[Comment.4: If a=mprtp is indicated, does the endpoint need to indicate a=rtcp-mux? because MPRTP mandates RTP RTCP multiplexing. --Editor]]



11.3. MPRTP Interface Advertisement in SDP (out-of-band signaling)

If the endpoint is aware of its multiple interfaces and wants to use them for MPRTP then it MAY use SDP to advertise these interfaces. Alternatively, it MAY use in-band signaling to advertise its interfaces, as defined in Section 9.3. The receiving endpoint MUST use the same mechanism to respond to an interface advertisement. In particular, if an endpoint receives an SDP offer, then it MUST respond to the offer in SDP.

11.3.1. "interface" attribute

The interface attribute is an optional media-level attribute and is used to advertise an endpoint’s interface address.

The syntax of the interface attribute is defined using the following Augmented BNF, as defined in [9]. The definitions of unicast- address, port, token, SP, and CRLF are according to RFC4566 [19].

mprtp-optional-parameter = mprtp-interface ; other optional parameters may be added later

mprtp-interface = "interface" ":" counter SP unicast-address ":" rtp_port *(SP interface-description-extension)

counter = 1*DIGIT rtp_port = port ;port from RFC4566

<mprtp-interface>: specifies one unicast IP address, the RTP and RTCP port number of the endpoint. The unicast address with lowest counter value MUST match the connection address (’c=’ line). Similarly, the RTP and RTCP ports MUST match the RTP and RTCP ports in the associated ’m=’ line. The counter should start at 1 and increment with each additional interface. Multiple interface lines MUST be ordered in a decreasing priority level as is the case with the Interface Advertisement blocks in in-band signaling (See Figure 13).

<unicast-address>: is taken from RFC4566 [19]. It is one of the IP addresses of the endpoint and allows the use of IPv4 addresses, IPv6 addresses and Fully Qualified Domain Names (FQDN). An endpoint MUST only include the IP address for which the connectivity checks have succeeded.

<port>: is from RFC4566 [19]. It is the RTP port associated with the unicast address and note that the RTP and RTCP ports are multiplexed for MPRTP subflows.



<counter>: is a monotonically increasing positive integer starting at 1. The counter MUST reset for each media line. The counter value for an ’mprtp-interface’ should remain the same for the session.

The ’mprtp-interface’ can be extended using the ’interface- description-extension’ parameter. An endpoint MUST ignore any extensions it does not understand.

11.3.2. Example

The ABNF grammar is illustrated by means of an example:

v=0 o=alice 2890844526 2890844527 IN IP4 192.0.2.1 s= c=IN IP4 192.0.2.1 t=0 0 m=video 49170 RTP/AVP 98 a=rtpmap:98 H264/90000 a=fmtp:98 profile-level-id=42A01E; a=extmap:1 urn:ietf:params:rtp-hdrext:mprtp a=rtcp-mux a=mprtp interface:1 192.0.2.1:49170 ;primary interface a=mprtp interface:2 198.51.100.1:51372 ;additional interface

11.4. MPRTP with ICE

If the endpoints intend to use ICE [3] for discovering interfaces and running connectivity checks then the following two step procedure MUST be followed:

1. Advertise ICE candidates: in the initial OFFER the endpoints exchange candidates, as defined in ICE [3]. Thereafter the endpoints run connectivity checks.

2. Advertise MPRTP interfaces: When a sufficient number of connectivity checks succeed the endpoint MUST send an updated offer containing the interfaces that they want to use for MPRTP.

When an endpoint uses ICE’s regular nomination [3] procedure, it chooses the best ICE candidate as the default path. In the case of an MPRTP endpoint, if more than one ICE candidate succeeded the connectivity checks then an MPRTP endpoint MAY advertise (some of) these as "mprtp-interfaces" in an updated offer.

When an endpoint uses ICE’s aggressive nomination [3] procedure, the selected candidate may change as more ICE checks complete. Instead of sending updated offers as additional ICE candidates appear



(transience), the endpoint it MAY use in-band signaling to advertise its interfaces, as defined in Section 9.3. Additionally, it MAY send an updated offer when the transience stabilizes.

If the default interface disappears and the paths used for MPRTP are different from the one in the c= and m= lines then the mprtp- intefaces with the lowest counter value should be promoted to the c= and m= lines in the updated offer.

When a new interface appears then the application/endpoint should internally decide if it wishes to use it and sends an updated offer with ICE candidates of the new interface. The receiving endpoint responds to the offer with all its ICE candidates in the answer and starts connectivity checks between all its candidates and the offerer’s new ICE candidate. Similarly, the initiating endpoint starts connectivity checks between the new interface and all the received ICE candidates in the answer. If the connectivity checks succeed, the initiating endpoint MAY send an updated offer with the new interface as an additional mprtp-interface.

11.5. Increased Throughput

The MPRTP layer MAY choose to augment paths to increase throughput. If the desired media rate exceeds the current media rate, the endpoints MUST renegotiate the application specific ("b=AS:xxx") [19] bandwidth.

11.6. Offer/Answer

When the SDP [19] is used to negotiate MPRTP sessions following the offer/answer model [15], the "a=mprtp" attribute (see Section 11.2) indicates the desire to use multiple interfaces to send or receive media data. The initial SDP offer MUST include this attribute at the media level. If the answerer wishes to also use MPRTP, it MUST include a media-level "a=mprtp" attribute in the answer. If the answer does not contain an "a=mprtp" attribute, the offerer MUST NOT stream media over multiple paths and the offerer MUST NOT advertise additional MPRTP interfaces in-band or out-of-band.

When SDP is used in a declarative manner, the presence of an "a=mprtp" attribute signals that the sender can send or receive media data over multiple interfaces. The receiver SHOULD be capable to stream media to the multiple interfaces and be prepared to receive media from multiple interfaces.

The following sections shows examples of SDP offer and answer for in- band and out-of-band signaling.



11.6.1. In-band Signaling Example

The following offer/answer shows that both the endpoints are MPRTP capable and SHOULD use in-band signaling for interfaces advertisements.

Offer: v=0 o=alice 2890844526 2890844527 IN IP4 192.0.2.1 s= c=IN IP4 192.0.2.1 t=0 0 m=video 49170 RTP/AVP 98 a=rtpmap:98 H264/90000 a=fmtp:98 profile-level-id=42A01E; a=rtcp-mux a=mprtp

Answer: v=0 o=bob 2890844528 2890844529 IN IP4 192.0.2.2 s= c=IN IP4 192.0.2.2 t=0 0 m=video 4000 RTP/AVP 98 a=rtpmap:98 H264/90000 a=fmtp:98 profile-level-id=42A01E; a=rtcp-mux a=mprtp

The endpoint MAY now use in-band RTCP signaling to advertise its multiple interfaces. Alternatively, it MAY make another offer with the interfaces in SDP (out-of-band signaling).

11.6.2. Out-of-band Signaling Example

If the multiple interfaces are included in an SDP offer then the receiver MUST respond to the request with an SDP answer.

11.6.2.1. Without ICE

In this example, the offerer advertises two interfaces and the answerer responds with a single interface description. The endpoint MAY use one or both paths depending on the end-to-end characteristics of each path.



Offer: v=0 o=alice 2890844526 2890844527 IN IP4 192.0.2.1 s= c=IN IP4 192.0.2.1 t=0 0 m=video 49170 RTP/AVP 98 a=rtpmap:98 H264/90000 a=fmtp:98 profile-level-id=42A01E; a=rtcp-mux a=mprtp interface:1 192.0.2.1:49170 a=mprtp interface:2 198.51.100.1:51372

Answer: v=0 o=bob 2890844528 2890844529 IN IP4 192.0.2.2 s= c=IN IP4 192.0.2.2 t=0 0 m=video 4000 RTP/AVP 98 a=rtpmap:98 H264/90000 a=fmtp:98 profile-level-id=42A01E; a=rtcp-mux a=mprtp interface:1 192.0.2.2:4000

11.6.2.2. With ICE

In this example, the endpoint first sends its ICE candidates in the initial offer and the other endpoint answers with its ICE candidates.

Initial offer (with ICE candidates):

Offer: v=0 o=alice 2890844526 2890844527 IN IP4 192.0.2.1 s= c=IN IP4 192.0.2.1 t=0 0 a=ice-pwd:asd88fgpdd777uzjYhagZg a=ice-ufrag:8hhY a=mprtp m=video 49170 RTP/AVP 98 a=rtpmap:98 H264/90000 a=fmtp:98 profile-level-id=42A01E; a=rtcp-mux a=candidate:1 1 UDP 2130706431 192.0.2.1 49170 typ host a=candidate:2 1 UDP 1694498815 198.51.100.1 51372 typ host



Answer: v=0 o=bob 2890844528 2890844529 IN IP4 192.0.2.2 s= c=IN IP4 192.0.2.2 t=0 0 a=ice-pwd:YH75Fviy6338Vbrhrlp8Yh a=ice-ufrag:9uB6 a=mprtp m=video 4000 RTP/AVP 98 a=rtpmap:98 H264/90000 a=fmtp:98 profile-level-id=42A01E; a=rtcp-mux a=candidate:1 1 UDP 2130706431 192.0.2.2 4000 typ host

Thereafter, each endpoint conducts ICE connectivity checks and when sufficient number of connectivity checks succeed, the endpoint sends an updated offer. In the updated offer, the endpoint advertises its multiple interfaces for MPRTP.

Updated offer (with MPRTP interfaces):

Offer: v=0 o=alice 2890844526 2890844527 IN IP4 192.0.2.1 s= c=IN IP4 192.0.2.1 t=0 0 m=video 49170 RTP/AVP 98 a=rtpmap:98 H264/90000 a=fmtp:98 profile-level-id=42A01E; a=rtcp-mux a=mprtp interface:1 192.0.2.1:49170 a=mprtp interface:2 198.51.100.1:51372

Answer: v=0 o=bob 2890844528 2890844529 IN IP4 192.0.2.2 s= c=IN IP4 192.0.2.2 t=0 0 m=video 4000 RTP/AVP 98 a=rtpmap:98 H264/90000 a=fmtp:98 profile-level-id=42A01E; a=rtcp-mux a=mprtp interface:1 192.0.2.2:4000



12. MPRTP in RTSP

Endpoints MUST use RTSP 2.0 [17] for session setup. Endpoints MUST multiplex RTP and RTCP on a single port [10] and follow the recommendations made in Appendix C of [17].

12.1. Solution Overview without ICE

1. The RTSP Server should include all of its multiple interfaces via the SDP attribute ("a=mprtp interface") in the RTSP DESCRIBE message.

2. The RTSP Client should include all its multiple interface in the RTSP SETUP message using the new attribute ("dest_mprtp_addr=") in the Transport header. [[Comment.5: Do we need a new lower layer transport MPRTP?. --Editor]]

3. The RTSP Server responds to the RTSP SETUP message with a 200 OK containing its MPRTP interfaces (using the "src_mprtp_header=") in the Transport header. After this the RTSP Client can issue a PLAY request.

4. If a new interface appears or an old one disappear at the RTSP Client during playback, it should send a new RTSP SETUP message containing the updated interfaces ("dest_mprtp_addr") in the Transport header. [[Comment.6: Sending a re-SETUP to update the interfaces during PLAY state would require a change in behavior of the server. Similar to Section 5.12 of [draft-ietf-mmusic-rtsp-nat]. --Editor]]

5. If a new interface appear or an old one disappear at the RTSP Server during playback, the RTSP Server should send a PLAY_NOTIFY message with a new Notify-Reason: "src-mprtp-interface-update". The request must contain the updated interfaces ("dest_mprtp_addr") in the "MPRTP-Interfaces" header.

6. Alternatively, the RTSP Server or Client may use the RTCP (in- band) mechanism to advertise their interfaces.

[[Comment.7: Does it make sense to advertise out-of-band (in RTSP SETUP) when advertising in-band in RTCP is less complex? --Editor]]

The overview is illustrated by means of an example:



C->S: DESCRIBE rtsp://server.example.com/fizzle/foo RTSP/2.0 CSeq: 111 User-Agent: PhonyClient 1.3 Accept: application/sdp, application/example Supported: setup.mprtp, setup.rtp.rtcp.mux

S->C: RTSP/2.0 200 OK CSeq: 111 Date: 23 Jan 2011 15:35:06 GMT Server: PhonyServer 1.3 Content-Type: application/sdp Content-Length: 367 Supported: setup.mprtp, setup.rtp.rtcp.mux

v=0 o=mprtp-rtsp-server 2890844526 2890844527 IN IP4 192.0.2.1 s= c=IN IP4 192.0.2.1 t=0 0 m=video 49170 RTP/AVP 98 a=rtpmap:98 H264/90000 a=fmtp:98 profile-level-id=42A01E; a=extmap:1 urn:ietf:params:rtp-hdrext:mprtp a=rtcp-mux a=mprtp interface:1 192.0.2.1:49170 a=mprtp interface:2 198.51.100.1:51372

On receiving the response to the RTSP DESCRIBE message, the RTSP Client sends a RTSP SETUP message containing its MPRTP interfaces in the Transport header using the "dest_mprtp_addr=" attribute. The RTSP Server responds with a 200 OK containing both the RTSP client’s and the RTSP Server’s MPRTP interfaces.



C->S: SETUP rtsp://server.example.com/fizzle/foo/audio RTSP/2.0 CSeq: 112 Transport: RTP/AVPF/UDP; unicast; dest_mprtp_addr=" 1 192.0.2.2 4000"; RTCP-mux, RTP/AVP/UDP; unicast; dest_addr=":6970"/":6971", RTP/AVP/TCP;unicast;interleaved=0-1 Accept-Ranges: NPT, UTC User-Agent: PhonyClient 1.3 Supported: setup.mprtp, setup.rtp.rtcp.mux

S->C: RTSP/2.0 200 OK CSeq: 112 Session: 12345678 Transport: RTP/AVPF/UDP; unicast; dest_mprtp_addr=" 1 192.0.2.2 4000"; src_mprtp_addr="1 192.0.2.1 49170; 2 198.51.100.1 51372"; RTCP-mux Accept-Ranges: NPT Date: 23 Jan 2012 15:35:06 GMT Server: PhonyServer 1.3 Supported: setup.mprtp, setup.rtp.rtcp.mux

The RTSP Client can issue a PLAY request on receiving the 200 OK and media can start to stream once the RTSP server receives the PLAY request.

12.2. Solution Overview with ICE

This overview uses the ICE mechanisms [20] defined for RTSP 2.0 [17].

1. The RTSP Server should include the "a=rtsp-ice-d-m" attribute and also indicate that it supports MPRTP by including the "a=mprtp" attribute in the SDP of the RTSP DESCRIBE message.

2. The client sends a RTSP SETUP message containing the D-ICE in lower level transport and ICE candidates in the transport header. The RTSP Server and client then follow the procedures (Steps 2 to 8) described in [20].

3. When the connectivity checks conclude the RTSP Client can send an updated RTSP SETUP message with its MPRTP interfaces (ICE candidates that were successful) in the Transport header ("dest_mprtp_addr="). The RTSP Server responds to the RTSP SETUP message with a 200 OK containing its MPRTP interfaces (ICE candidates that were successful) in the Transport header ("src_mprtp_header="). After receiving the 200 OK, the RTSP client can issue the PLAY request.



4. Alternatively after the connectivity checks conclude, the RTSP Client can issue the PLAY request (Step 9 and 10 of [20]) and the endpoints can use the RTCP (in-band) mechanism to advertise their interfaces.

5. If a new interface appears or an old one disappears, the RTSP Client should issue an updated SETUP message with the new candidates (See Section 5.12 of [20]) or the RTSP Server should send a PLAY_NOTIFY message (See Section 5.13 of [20]). After connectivity checks succeed for the new interfaces, the RTSP Client can proceed with the instructions in Step 3 or 4.

The overview is illustrated by means of an example:

C->S: DESCRIBE rtsp://server.example.com/foo RTSP/2.0 CSeq: 312 User-Agent: PhonyClient 1.3 Accept: application/sdp, application/example Supported: setup.mprtp, setup.ice-d-m, setup.rtp.rtcp.mux

S->C: RTSP/2.0 200 OK CSeq: 312 Date: 23 Jan 2012 15:35:06 GMT Server: PhonyServer 1.3 Content-Type: application/sdp Content-Length: 367 Supported: setup.mprtp, setup.ice-d-m, setup.rtp.rtcp.mux

v=0 o=mprtp-rtsp-server 2890844526 2890842807 IN IP4 192.0.2.1 s=SDP Seminar i=A Seminar on the session description protocol u=http://www.example.com/lectures/sdp.ps [email protected] (Seminar Management) t=2873397496 2873404696 a=recvonly a=rtsp-ice-d-m a=control: * m=video 49170 RTP/AVP 98 a=rtpmap:98 H264/90000 a=fmtp:98 profile-level-id=42A01E; a=rtcp-mux a=mprtp a=control: /video



C->S: SETUP rtsp://server.example.com/foo/video RTSP/2.0 CSeq: 302 Transport: RTP/AVP/D-ICE; unicast; ICE-ufrag=9uB6; ICE-Password=YH75Fviy6338Vbrhrlp8Yh; candidates="1 1 UDP 2130706431 192.0.2.2 4000 typ host"; RTCP-mux, RTP/AVP/UDP; unicast; dest_addr=":6970"/":6971", RTP/AVP/TCP;unicast;interleaved=0-1 Accept-Ranges: NPT, UTC User-Agent: PhonyClient 1.3 Supported: setup.mprtp, setup.ice-d-m, setup.rtp.rtcp.mux

S->C: RTSP/2.0 200 OK CSeq: 302 Session: 12345678 Transport: RTP/AVP/D-ICE; unicast; RTCP-mux; ICE-ufrag=8hhY; ICE-Password= asd88fgpdd777uzjYhagZg; candidates=" 1 1 UDP 2130706431 192.0.2.1 49170 typ host; 2 1 UDP 1694498815 198.51.100.1 51372 typ host" Accept-Ranges: NPT Date: 23 Jan 2012 15:35:06 GMT Server: PhonyServer 1.3 Supported: setup.mprtp, setup.ice-d-m, setup.rtp.rtcp.mux

After the connectivity checks complete, the RTSP Client can send an updated RTSP SETUP message containing the MPRTP interfaces for which the connectivity checks were successful. These steps are the same as the ones in the previous example.

12.3. RTSP Extensions

12.3.1. MPRTP Interface Transport Header Parameter

This section defines a new RTSP transport parameter for carrying MPRTP interfaces. The transport parameters may only occur once in each transport specification. The parameter can contain one or more MPRTP interfaces. In the SETUP response if the RTSP server support MPRTP it MUST be include one or more MPRTP interfaces.



trns-parameter = <Defined in Section 20.2.3 of [I-D.ietf-mmusic-rfc2326bis]> trns-parameter =/ SEMI dest-mprtp-interface-par trns-parameter =/ SEMI src-mprtp-interface-par dest-mprtp-interface-par = "dest_mprtp_addr" EQUAL DQ SWS interface *(SEMI interface) SWS DQ src-mprtp-interface-par = "src_mprtp_addr" EQUAL DQ SWS interface *(SEMI interface) SWS DQ

interface = counter SP unicast-address SP rtp_port SP *(SP interface-description-extension)

counter = See section 11.3.1 unicast-address = See section 11.3.1 rtp_port = See section 11.3.1 interface-description-extension = See section 11.3.1

12.3.2. MPRTP Feature Tag

A feature tag is defined for indicating MPRTP support in the RTSP capabilities mechanism: "setup.mprtp". This feature tag indicates that the endpoint supports all the mandatory extensions defined in this specification and is applicable to all types of RTSP agents; clients, servers and proxies.

The MPRTP compliant RTSP Client MUST send the feature tag "setup.mprtp" in the "Supported" header of all DESCRIBE and SETUP requests.

12.3.3. Status Codes

TBD

12.3.4. New Reason for PLAY_NOTIFY

A new value used in the PLAY_NOTIFY methods Notify-Reason header is defined: "src-mprtp-interface-update". This reason indicates that the RTSP Server has updated set of MPRTP interfaces.

Notify-Reas-val =/ "src-mprtp-interface-update"

PLAY_NOTIFY requests with Notify-Reason header set to src-mprtp- interface-update MUST include a mprtp-interfaces header.



mprtp-interfaces = "mprtp-interfaces" HCOLON interface *(COMMA interface) interface = counter SP unicast-address SP rtp_port SP *(SP interface-description-extension)

counter = See section 11.3.1 unicast-address = See section 11.3.1 rtp_port = See section 11.3.1 interface-description-extension = See section 11.3.1

[[Comment.8: Do we need to add a new header attribute?. Alternatively, the RTSP Server could just send the PLAY_NOTIFY and let the RTSP client initiate a new RTSP SETUP message with its current interfaces and the RTSP Server can then respond with its updated set of interfaces. This will make it a 3-way exchange as opposed to a 1-way notification. Alternatively, using SET_PARAMETER reduces it to a 2-way exchange and can be initiated by both the RTSP Server and the RTSP Client. However, SET_PARAMETER can only be used when the endpoints are in SETUP state. --Editor]]

Example:

S->C: PLAY_NOTIFY rtsp://server.example.com/foo RTSP/2.0 CSeq: 305 Notify-Reason: src-mprtp-interface-update Session: 12345678 mprtp-interfaces: 2 192.0.2.10 48211, 3 198.51.100.11 38703 Server: PhonyServer 1.3

C->S: RTSP/2.0 200 OK CSeq: 305 User-Agent: PhonyClient 1.3

12.3.5. re-SETUP

The server SHALL support SETUP requests in PLAYING state if it is only updating the transport parameter (dest_mprtp_addr). If the session is established using ICE then the RTSP Server and Client MUST also follow the procedures described for re-SETUP in [20].


The following contact information shall be used for all registrations in this document:



Contact: Varun Singh mailto:[email protected] tel:+358-9-470-24785

Note to the RFC-Editor: When publishing this document as an RFC, please replace "RFC XXXX" with the actual RFC number of this document and delete this sentence.

13.1. MPRTP Header Extension

This document defines a new extension URI to the RTP Compact Header Extensions sub-registry of the Real-Time Transport Protocol (RTP) Parameters registry, according to the following data:

Extension URI: urn:ietf:params:rtp-hdrext:mprtp Description: Multipath RTP Reference: RFC XXXX

13.2. MPRTCP Packet Type

A new RTCP packet format has been registered with the RTCP Control Packet type (PT) Registry:

Name: MPRTCP Long name: Multipath RTCP Value: 211 Reference: RFC XXXX.

This document defines a substructure for MPRTCP packets. A new sub- registry has been set up for the sub-report block type (MPRTCP_Type) values for the MPRTCP packet, with the following registrations created initially:



Name: Subflow Specific Report Long name: Multipath RTP Subflow Specific Report Value: 0 Reference: RFC XXXX.

Name: IPv4 Address Long name: IPv4 Interface Address Value: 1 Reference: RFC XXXX.

Name: IPv6 Address Long name: IPv6 Interface Address Value: 2 Reference: RFC XXXX.

Name: DNS Name Long name: DNS Name indicating Interface Address Value: 3 Reference: RFC XXXX.

Further values may be registered on a first come, first served basis. For each new registration, it is mandatory that a permanent, stable, and publicly accessible document exists that specifies the semantics of the registered parameter as well as the syntax and semantics of the associated sub-report block. The general registration procedures of [19] apply.

13.3. SDP Attributes

This document defines a new SDP attribute, "mprtp", within the existing IANA registry of SDP Parameters.

13.3.1. "mprtp" attribute

o Attribute Name: MPRTP

o Long Form: Multipath RTP

o Type of Attribute: media-level

o Charset Considerations: The attribute is not subject to the charset attribute.

o Purpose: This attribute is used to indicate support for Multipath RTP. It can also provide one of many possible interfaces for communication. These interface addresses may have been validated using ICE procedures.



o Appropriate Values: See Section 11.2 and Section 11.3.1 of RFC XXXX.

13.4. RTSP

This document requests registration in a number of registries for RTSP.

13.4.1. RTSP Feature Tag

This document request that one RTSP 2.0 feature tag be registered in the "RTSP 2.0 feature tag" registry:

setup.mprtp See Section 12.3.2.

13.4.2. RTSP Transport Parameters

This document requests that 2 transport parameters be registered in RTSP 2.0’s "Transport Parameters":

"dest_mprtp_addr": See Section 12.3.1.

"src_mprtp_addr": See Section 12.3.1.

13.4.3. Notify-Reason value

This document requests that one assignment be done in the RTSP 2.0 Notify-Reason header value registry. The defined value is:

"src-mprtp-interface-update": See Section 12.3.4.


TBD

All drafts are required to have a security considerations section. See RFC 3552 [21] for a guide.


Varun Singh, Saba Ahsan, and Teemu Karkkainen are supported by Trilogy (http://www.trilogy-project.org), a research project (ICT- 216372) partially funded by the European Community under its Seventh Framework Program. The views expressed here are those of the author(s) only. The European Commission is not liable for any use that may be made of the information in this document.



The authors would also like acknowledge the contribution of Ralf Globisch and Thomas Schierl for providing the input and text for the MPRTP interface advertisement in SDP.

Thanks to Miguel A. Garcia, Ralf Globisch, Christer Holmberg, and Roni Even for providing valuable feedback on earlier versions of this draft

16. References


[1] Schulzrinne, H., Casner, S., Frederick, R., and V. Jacobson, "RTP: A Transport Protocol for Real-Time Applications", STD 64, RFC 3550, July 2003.

[2] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997.

[3] Rosenberg, J., "Interactive Connectivity Establishment (ICE): A Protocol for Network Address Translator (NAT) Traversal for Offer/Answer Protocols", RFC 5245, April 2010.

[4] Ott, J., Chesterfield, J., and E. Schooler, "RTP Control Protocol (RTCP) Extensions for Single-Source Multicast Sessions with Unicast Feedback", RFC 5760, February 2010.

[5] Johansson, I. and M. Westerlund, "Support for Reduced-Size Real-Time Transport Control Protocol (RTCP): Opportunities and Consequences", RFC 5506, April 2009.

[6] Ott, J., Wenger, S., Sato, N., Burmeister, C., and J. Rey, "Extended RTP Profile for Real-time Transport Control Protocol (RTCP)-Based Feedback (RTP/AVPF)", RFC 4585, July 2006.

[7] Singer, D. and H. Desineni, "A General Mechanism for RTP Header Extensions", RFC 5285, July 2008.

[8] Yergeau, F., "UTF-8, a transformation format of ISO 10646", STD 63, RFC 3629, November 2003.

[9] Crocker, D. and P. Overell, "Augmented BNF for Syntax Specifications: ABNF", STD 68, RFC 5234, January 2008.

[10] Perkins, C. and M. Westerlund, "Multiplexing RTP Data and Control Packets on a Single Port", RFC 5761, April 2010.




[11] Stewart, R., "Stream Control Transmission Protocol", RFC 4960, September 2007.

[12] Ford, A., Raiciu, C., Handley, M., Barre, S., and J. Iyengar, "Architectural Guidelines for Multipath TCP Development", RFC 6182, March 2011.

[13] Nordmark, E. and M. Bagnulo, "Shim6: Level 3 Multihoming Shim Protocol for IPv6", RFC 5533, June 2009.

[14] Westerlund, M. and S. Wenger, "RTP Topologies", RFC 5117, January 2008.

[15] Rosenberg, J. and H. Schulzrinne, "An Offer/Answer Model with Session Description Protocol (SDP)", RFC 3264, June 2002.

[16] Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston, A., Peterson, J., Sparks, R., Handley, M., and E. Schooler, "SIP: Session Initiation Protocol", RFC 3261, June 2002.

[17] Schulzrinne, H., Rao, A., Lanphier, R., Westerlund, M., and M. Stiemerling, "Real Time Streaming Protocol 2.0 (RTSP)", draft-ietf-mmusic-rfc2326bis-28 (work in progress), October 2011.

[18] Marjou, X. and A. Sollaud, "Application Mechanism for Keeping Alive the NAT Mappings Associated with RTP / RTP Control Protocol (RTCP) Flows", RFC 6263, June 2011.

[19] Handley, M., Jacobson, V., and C. Perkins, "SDP: Session Description Protocol", RFC 4566, July 2006.

[20] Goldberg, J., Westerlund, M., and T. Zeng, "A Network Address Translator (NAT) Traversal mechanism for media controlled by Real-Time Streaming Protocol (RTSP)", draft-ietf-mmusic-rtsp-nat-11 (work in progress), October 2011.

[21] Rescorla, E. and B. Korver, "Guidelines for Writing RFC Text on Security Considerations", BCP 72, RFC 3552, July 2003.

Appendix A. Interoperating with Legacy Applications

An MPRTP sender can use its multiple interfaces to send media to a legacy RTP client. The legacy receiver will ignore the subflow RTP header and the receiver’s de-jitter buffer will try to compensate for



the mismatch in per-path delay. However, the receiver can only send the overall or aggregate RTCP report which may be insufficient for an MPRTP sender to adequately schedule packets or detect if a path disappeared.

An MPRTP receiver can only use one of its interface when communicating with a legacy sender.

Appendix B. Change Log

Note to the RFC-Editor: please remove this section prior to publication as an RFC.

B.1. changes in draft-singh-avtcore-mprtp-04

o Fixed missing 0xBEDE header in MPRTP header format.

o Removed connectivity checks and keep-alives from in-band signaling.

o MPRTP and MPRTCP are multiplexed on a single port.

o MPRTCP packet headers optimized.

o Made ICE optional

o Updated Sections: 7.1.2, 8.1.x, 11.2, 11.4, 11.6.

o Added how to use MPRTP in RTSP (Section 12).

o Updated IANA Considerations section.


o Added this change log.

o Updated section 6, 7 and 8 based on comments from MMUSIC.

o Updated section 11 (SDP) based on comments of MMUSIC.

o Updated SDP examples with ICE and non-ICE in out-of-band signaling scenario.

o Added Appendix A on interop with legacy.

o Updated IANA Considerations section.




o MPRTCP protocol extensions use only one PT=210, instead of 210 and 211.

o RTP header uses 1-byte extension instead of 2-byte.

o Added section on RTCP Interval Calculations.

o Added "mprtp-interface" attribute in SDP considerations.


o Added MPRTP and MPRTCP protocol extensions and examples.

o WG changed from -avt to -avtcore.


Varun Singh Aalto University School of Science and Technology Otakaari 5 A Espoo, FIN 02150 Finland

Email: [email protected] URI: http://www.netlab.tkk.fi/˜varun/

Teemu Karkkainen Aalto University School of Science and Technology Otakaari 5 A Espoo, FIN 02150 Finland




Joerg Ott Aalto University School of Science and Technology Otakaari 5 A Espoo, FIN 02150 Finland


Saba Ahsan Aalto University School of Science and Technology Otakaari 5 A Espoo, FIN 02150 Finland


Lars Eggert NetApp Sonnenallee 1 Kirchheim 85551 Germany

Phone: +49 151 12055791 Email: [email protected] URI: http://eggert.org/


Network Working Group M. WesterlundInternet-Draft B. BurmanIntended status: Informational EricssonExpires: September 13, 2012 C. Perkins University of Glasgow March 12, 2012

RTP Multiplexing Architecture draft-westerlund-avtcore-multiplex-architecture-01

Abstract

Real-time Transport Protocol is a flexible protocol possible to use in a wide range of applications and network and system topologies. This flexibility and the implications of different choices should be understood by any application developer using RTP. To facilitate that understanding, this document contains an in-depth discussion of the usage of RTP’s multiplexing points; the RTP session, the Synchronization Source Identifier (SSRC), and the payload type. The focus is put on the first two, trying to give guidance and source material for an analysis on the most suitable choices for the application being designed.

Status of this Memo





Copyright Notice


This document is subject to BCP 78 and the IETF Trust’s Legal Provisions Relating to IETF Documents

Westerlund, et al. Expires September 13, 2012 [Page 1]

Internet-Draft RTP Multiplexing Architecture March 2012

(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.

Table of Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 2. Definitions . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.1. Requirements Language . . . . . . . . . . . . . . . . . . 5 2.2. Terminology . . . . . . . . . . . . . . . . . . . . . . . 5 3. RTP Multiplex Points . . . . . . . . . . . . . . . . . . . . . 6 3.1. Session . . . . . . . . . . . . . . . . . . . . . . . . . 6 3.2. SSRC . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 3.3. CSRC . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 3.4. Payload Type . . . . . . . . . . . . . . . . . . . . . . . 9 4. Multiple Streams Alternatives . . . . . . . . . . . . . . . . 10 5. RTP Topologies and Issues . . . . . . . . . . . . . . . . . . 11 5.1. Point to Point . . . . . . . . . . . . . . . . . . . . . . 12 5.1.1. RTCP Reporting . . . . . . . . . . . . . . . . . . . . 12 5.1.2. Compound RTCP Packets . . . . . . . . . . . . . . . . 13 5.2. Point to Multipoint Using Multicast . . . . . . . . . . . 13 5.3. Point to Multipoint Using an RTP Translator . . . . . . . 15 5.4. Point to Multipoint Using an RTP Mixer . . . . . . . . . . 16 5.5. Point to Multipoint using Multiple Unicast flows . . . . . 17 5.6. De-composite End-Point . . . . . . . . . . . . . . . . . . 18 6. Multiple Streams Discussion . . . . . . . . . . . . . . . . . 19 6.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . 19 6.2. RTP/RTCP Aspects . . . . . . . . . . . . . . . . . . . . . 19 6.2.1. The RTP Specification . . . . . . . . . . . . . . . . 19 6.2.2. Handling Varying sets of Senders . . . . . . . . . . . 22 6.2.3. Cross Session RTCP Requests . . . . . . . . . . . . . 22 6.2.4. Binding Related Sources . . . . . . . . . . . . . . . 22 6.2.5. Forward Error Correction . . . . . . . . . . . . . . . 24 6.2.6. Transport Translator Sessions . . . . . . . . . . . . 25 6.3. Interworking . . . . . . . . . . . . . . . . . . . . . . . 25 6.3.1. Interworking Applications . . . . . . . . . . . . . . 26 6.3.2. Multiple SSRC Legacy Considerations . . . . . . . . . 27 6.4. Signalling Aspects . . . . . . . . . . . . . . . . . . . . 28 6.4.1. Session Oriented Properties . . . . . . . . . . . . . 28 6.4.2. SDP Prevents Multiple Media Types . . . . . . . . . . 29 6.4.3. Media Stream Usage . . . . . . . . . . . . . . . . . . 29 6.5. Network Aspects . . . . . . . . . . . . . . . . . . . . . 30 6.5.1. Quality of Service . . . . . . . . . . . . . . . . . . 30



6.5.2. NAT and Firewall Traversal . . . . . . . . . . . . . . 31 6.5.3. Multicast . . . . . . . . . . . . . . . . . . . . . . 32 6.5.4. Multiplexing multiple RTP Session on a Single Transport . . . . . . . . . . . . . . . . . . . . . . 33 6.6. Security Aspects . . . . . . . . . . . . . . . . . . . . . 33 6.6.1. Security Context Scope . . . . . . . . . . . . . . . . 33 6.6.2. Key-Management for Multi-party session . . . . . . . . 34 6.6.3. Complexity Implications . . . . . . . . . . . . . . . 34 6.7. Multiple Media Types in one RTP session . . . . . . . . . 35 7. Arch-Types . . . . . . . . . . . . . . . . . . . . . . . . . . 37 7.1. Single SSRC per Session . . . . . . . . . . . . . . . . . 37 7.2. Multiple SSRCs of the Same Media Type . . . . . . . . . . 39 7.3. Multiple Sessions for one Media type . . . . . . . . . . . 40 7.4. Multiple Media Types in one Session . . . . . . . . . . . 41 7.5. Summary . . . . . . . . . . . . . . . . . . . . . . . . . 43 8. Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . 43 9. Proposal for Future Work . . . . . . . . . . . . . . . . . . . 44 10. RTP Specification Clarifications . . . . . . . . . . . . . . . 45 10.1. RTCP Reporting from all SSRCs . . . . . . . . . . . . . . 45 10.2. RTCP Self-reporting . . . . . . . . . . . . . . . . . . . 45 10.3. Combined RTCP Packets . . . . . . . . . . . . . . . . . . 45 11. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 46 12. Security Considerations . . . . . . . . . . . . . . . . . . . 46 13. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 46 14. References . . . . . . . . . . . . . . . . . . . . . . . . . . 46 14.1. Normative References . . . . . . . . . . . . . . . . . . . 46 14.2. Informative References . . . . . . . . . . . . . . . . . . 46 Appendix A. Dismissing Payload Type Multiplexing . . . . . . . . 49 Authors’ Addresses . . . . . . . . . . . . . . . . . . . . . . . . 51



1. Introduction

Real-time Transport Protocol (RTP) [RFC3550] is a commonly used protocol for real-time media transport. It is a protocol that provides great flexibility and can support a large set of different applications. RTP has several multiplexing points designed for different purposes. These enable support of multiple media streams and switching between different encoding or packetization of the media. By using multiple RTP sessions, sets of media streams can be structured for efficient processing or identification. Thus the question for any RTP application designer is how to best use the RTP session, the SSRC and the payload type to meet the application’s needs.

The purpose of this document is to provide clear information about the possibilities of RTP when it comes to multiplexing. The RTP application designer should understand the implications that come from a particular choice of RTP multiplexing points. The document will recommend against some usages as being unsuitable, in general or for particular purposes.

RTP was from the beginning designed for multiple participants in a communication session. This is not restricted to multicast, as some may believe, but also provides functionality over unicast, using either multiple transport flows below RTP or a network node that re- distributes the RTP packets. The re-distributing node can for example be a transport translator (relay) that forwards the packets unchanged, a translator performing media translation in addition to forwarding, or an RTP mixer that creates new conceptual sources from the received streams. In addition, multiple streams may occur when a single end-point have multiple media sources, like multiple cameras or microphones that need to be sent simultaneously.

This document has been written due to increased interest in more advanced usage of RTP, resulting in questions regarding the most appropriate RTP usage. The limitations in some implementations, RTP/ RTCP extensions, and signalling has also been exposed. It is expected that some limitations will be addressed by updates or new extensions resolving the shortcomings. The authors also hope that clarification on the usefulness of some functionalities in RTP will result in more complete implementations in the future.

The document starts with some definitions and then goes into the existing RTP functionalities around multiplexing. Both the desired behavior and the implications of a particular behavior depend on which topologies are used, which requires some consideration. This is followed by a discussion of some choices in multiplexing behavior and their impacts. Some arch-types of RTP usage are discussed.



Finally, some recommendations and examples are provided.

This document is currently an individual contribution, but it is the intention of the authors that this should become a WG document that objectively describes and provides suitable recommendations for which there is WG consensus. Currently this document only represents the views of the authors. The authors gladly accept any feedback on the document and will be happy to discuss suitable recommendations.

2. Definitions



2.2. Terminology

The following terms and abbreviations are used in this document:

End-point: A single entity sending or receiving RTP packets. It may be decomposed into several functional blocks, but as long as it behaves a single RTP stack entity it is classified as a single end-point.

Media Stream: A sequence of RTP packets using a single SSRC that together carries part or all of the content of a specific Media Type from a specific sender source within a given RTP session.

Media Source: The originator or source of a particular Media Stream. It can either be a single media capturing device such as a video camera, a microphone, or a specific output of a media production function, such as an audio mixer, or some video editing function.

Media Aggregate: All Media Streams related to a particular Source.

Media Type: Audio, video, text or data whose form and meaning are defined by a specific real-time application.

Multiplex: The operation of taking multiple entities as input, aggregating them onto some common resource while keeping the individual entities addressable such that they can later be fully and unambiguously separated (de-multiplexed) again.



RTP Session: As defined by [RFC3550], the end-points belonging to the same RTP Session are those that share a single SSRC space. That is, those end-points can see an SSRC identifier transmitted by any one of the other end-points. An end-point can receive an SSRC either as SSRC or as CSRC in RTP and RTCP packets. Thus, the RTP Session scope is decided by the end-points’ network interconnection topology, in combination with RTP and RTCP forwarding strategies deployed by end-points and any interconnecting middle nodes.

Source: See Media Source.

3. RTP Multiplex Points

This section describes the existing RTP tools that enable multiplexing of different media streams.

3.1. Session

The RTP Session is the highest semantic level in RTP and contains all of the RTP functionality.

Identifier: RTP in itself does not contain any Session identifier, but relies either on the underlying transport or on the used signalling protocol, depending on in which context the identifier is used (e.g. transport or signalling). Due to this, a single RTP Session may have multiple associated identifiers belonging to different contexts.

Position: Depending on underlying transport and signalling protocol. For example, when running RTP on top of UDP, an RTP endpoint can identify and delimit an RTP Session from other RTP Sessions through the UDP source and destination transport address, consisting of network address and port number(s). Commonly, RTP and RTCP use separate ports and the destination transport address is in fact an address pair, but in the case of RTP/RTCP multiplex [RFC5761] there is only a single port. Another example is SDP signalling [RFC4566], where the grouping framework [RFC5888] uses an identifier per "m="-line. If there is a one-to-one mapping between "m="-line and RTP Session, that grouping framework identifier can identify a single RTP Session.

Usage: Identify separate RTP Sessions.



Uniqueness: Globally unique within the general communication context for the specific end-point.

Inter-relation: Depending on the underlying transport and signalling protocol.

Special Restrictions: None.

A source that changes its source transport address during a session must also choose a new SSRC identifier to avoid being interpreted as a looped source.

The set of participants considered part of the same RTP Session is defined by[RFC3550] as those that share a single SSRC space. That is, those participants that can see an SSRC identifier transmitted by any one of the other participants. A participant can receive an SSRC either as SSRC or CSRC in RTP and RTCP packets. Thus, the RTP Session scope is decided by the participants’ network interconnection topology, in combination with RTP and RTCP forwarding strategies deployed by end-points and any interconnecting middle nodes.

3.2. SSRC

An RTP Session serves one or more Media Sources, each sending a Media Stream.

Identifier: Synchronization Source (SSRC), 32-bit unsigned number.

Position: In every RTP and RTCP packet header. May be present in RTCP payload. May be present in SDP signalling.

Usage: Identify individual Media Sources within an RTP Session. Refer to individual Media Sources in RTCP messages and SDP signalling.

Uniqueness: Randomly chosen, globally unique within an RTP Session and not dependent on network address.

Inter-relation: SSRC belonging to the same synchronization context (originating from the same end-point), within or between RTP Sessions, are indicated through use of identical SDES CNAME items in RTCP compound packets with those SSRC as originating source. SDP signalling can provide explicit SSRC grouping [RFC5576]. When CNAME is inappropriate or insufficient, there exist a few other methods to relate different SSRC. One such case is session-based RTP retransmission [RFC4588]. In some cases, the same SSRC Identifier value is used to relate streams in two different RTP



Sessions, such as in Multi-Session Transmission of scalable video [RFC6190].

Special Restrictions: All RTP implementations must be prepared to use procedures for SSRC collision handling, which results in an SSRC number change. A Media Source that changes its RTP Session identifier (e.g. source transport address) during a session must also choose a new SSRC identifier to avoid being interpreted as looped source. Note that RTP sequence number and RTP timestamp are scoped by SSRC and thus independent between different SSRCs.

A media source having an SSRC identifier can be of different types:

Real: Connected to a "physical" media source, for example a camera or microphone.

Conceptual: A source with some attributed property generated by some network node, for example a filtering function in an RTP mixer that provides the most active speaker based on some criteria, or a mix representing a set of other sources.

Virtual: A source that does not generate any RTP media stream in itself (e.g. an end-point only receiving in an RTP session), but anyway need a sender SSRC for use as source in RTCP reports.

Note that a "multimedia source" that generates more than one media type, e.g. a conference participant sending both audio and video, need not (and commonly should not) use the same SSRC value across RTP sessions. RTCP Compound packets containing the CNAME SDES item is the designated method to bind an SSRC to a CNAME, effectively cross- correlating SSRCs within and between RTP Sessions as coming from the same end-point. The main property attributed to SSRCs associated with the same CNAME is that they are from a particular synchronization context and may be synchronized at playback.

Note also that RTP sequence number and RTP timestamp are scoped by SSRC and thus independent between different SSRCs.

An RTP receiver receiving a previously unseen SSRC value must interpret it as a new source. It may in fact be a previously existing source that had to change SSRC number due to an SSRC conflict. However, the originator of the previous SSRC should have ended the conflicting source by sending an RTCP BYE for it prior to starting to send with the new SSRC, so the new SSRC is anyway effectively a new source.

Some RTP extension mechanisms already require the RTP stacks to handle additional SSRCs, like SSRC multiplexed RTP retransmission



[RFC4588]. However, that still only requires handling a single media decoding chain per pair of SSRCs.

3.3. CSRC

The Contributing Source (CSRC) can arguably be seen as a sub-part of a specific SSRC and thus a multiplexing point. It is optionally included in the RTP header, shares the SSRC number space and specifies which set of SSRCs that has contributed to the RTP payload. However, even though each RTP packet and SSRC can be tagged with the contained CSRCs, the media representation of an individual CSRC is in general not possible to extract from the RTP payload since it is typically the result of a media mixing (merge) operation (by an RTP mixer) on the individual media streams corresponding to the CSRC identifiers. Due to these restrictions, CSRC will not be considered a fully qualified multiplex point and will be disregarded in the rest of this document.

3.4. Payload Type

Each Media Stream can be represented in various encoding formats.

Identifier: Payload Type number.

Position: In every RTP header and in SDP signalling.

Usage: Identify a specific Media Stream encoding format. The format definition may be taken from [RFC3551] for statically allocated Payload Types, but should be explicitly defined in signalling, such as SDP, both for static and dynamic Payload Types. The term "format" here includes whatever can be described by out-of-band signaling means. In SDP, the term "format" includes media type, RTP timestamp sampling rate, codec, codec configuration, payload format configurations, and various robustness mechanisms such as redundant encodings [RFC2198].

Uniqueness: Scoped by sending end-point within an RTP Session. To avoid any potential for ambiguity, it is desirable that payload types are unique across all sending end-points within an RTP session, but this is often not true in practice. All SSRC in an RTP session sent from an single end-point share the same Payload Types definitions. The RTP Payload Type is designed such that only a single Payload Type is valid at any time instant in the SSRC’s RTP timestamp time line, effectively time-multiplexing different Payload Types if any change occurs. Used Payload Type may change on a per-packet basis for an SSRC, for example a speech codec making use of generic Comfort Noise



[RFC3389].

Inter-relation: There are some uses where Payload Type numbers need be unique across RTP Sessions. This is for example the case in Media Decoding Dependency [RFC5583] where Payload Types are used to describe media dependency across RTP Sessions. Another example is session-based RTP retransmission [RFC4588].

Special Restrictions: Using different RTP timestamp clock rates for the RTP Payload Types in use in the same RTP Session have issues such as loss of synchronization. Payload Type clock rate switching requires some special consideration that is described in the multiple clock rates specification [I-D.ietf-avtext-multiple-clock-rates].

If there is a true need to send multiple Payload Types for the same SSRC that are valid for the same RTP Timestamps, then redundant encodings [RFC2198] can be used. Several additional constraints than the ones mentioned above need to be met to enable this use, one of which is that the combined payload sizes of the different Payload Types must not exceed the transport MTU.

Other aspects of RTP payload format use are described in RTP Payload HowTo [I-D.ietf-payload-rtp-howto].

4. Multiple Streams Alternatives

This section reviews the alternatives to enable multi-stream handling. Let’s start with describing mechanisms that could enable multiple media streams, independent of the purpose for having multiple streams.

SSRC Multiplexing: Each additional Media Stream gets its own SSRC within a RTP Session.

Session Multiplexing: Using additional RTP Sessions to handle additional Media Streams

Payload Type Multiplexing: Using different RTP payload types for different additional streams.

Independent of the reason to use additional media streams, achieving it using payload type multiplexing is not a good choice as can be seen in the Appendix A. The RTP payload type alone is not suitable for cases where additional media streams are required. Streams need their own SSRCs, so that they get their own sequence number space. The SSRC itself is also important so that the media stream can be



referenced and reported on.

This leaves us with two main choices, either using SSRC multiplexing to have multiple SSRCs from one end-point in one RTP session, or create an additional RTP session to hold that additional SSRC. As the below discussion will show, in reality we cannot choose a single one of the two solutions. To utilize RTP well and as efficiently as possible, both are needed. The real issue is finding the right guidance on when to create RTP sessions and when additional SSRCs in an RTP session is the right choice.

In the below discussion, please keep in mind that the reasons for having multiple media streams vary and include but are not limited to the following:

o Multiple Media Sources

o Retransmission streams

o FEC stream

o Alternative Encodings

o Scalability layers

Thus the choice made due to one reason may not be the choice suitable for another reason. In the above list, the different items have different levels of maturity in the discussion on how to solve them. The clearest understanding is associated with multiple media sources of the same media type. However, all warrant discussion and clarification on how to deal with them.

5. RTP Topologies and Issues

The impact of how RTP Multiplex is performed will in general vary with how the RTP Session participants are interconnected; the RTP Topology [RFC5117]. This section describes the topologies and attempts to highlight the important behaviors concerning RTP multiplexing and multi-stream handling. It lists any identified issues regarding RTP and RTCP handling, and introduces additional topologies that are supported by RTP beyond those included in RTP Topologies [RFC5117]. The RTP Topologies that do not follow the RTP specification or do not attempt to utilize the facilities of RTP are ignored in this document.



5.1. Point to Point

This is the most basic use case with an RTP session containing two end-points. Each end-point has one or more SSRCs.

+---+ +---+ | A |<------->| B | +---+ +---+

Figure 1: Point to Point

5.1.1. RTCP Reporting

In cases when an end-point uses multiple SSRCs, we have found two closely related issues. The first is if every SSRC shall report on all other SSRC, even the ones originating from the same end-point. The reason for this would be to ensure that no monitoring function should suspect a breakage in the RTP session.

The second issue around RTCP reporting arise when an end-point receives one or more media streams, and when the receiving end-point itself sends multiple SSRC in the same RTP session. As transport statistics are gathered per end-point and shared between the nodes, all the end-point’s SSRC will report based on the same received data, the only difference will be which SSRCs sends the report. This could be considered unnecessary overhead, but for consistency it might be simplest to always have all sending SSRCs send RTCP reports on all media streams the end-point receives.

The current RTP text is silent about sending RTCP Receiver Reports for an endpoint’s own sources, but does not preclude either sending or omitting them. The uncertainty in the expected behavior in those cases has likely caused variations in the implementation strategy. This could cause an interoperability issue where it is not possible to determine if the lack of reports is a true transport issue, or simply a result of implementation.

Although this issue is valid already for the simple point to point case, it needs to be considered in all topologies. From the perspective of an end-point, any solution needs to take into account what a particular end-point can determine without explicit information of the topology. For example, a Transport Translator (Relay) topology will look quite similar to point to point on a transport level but is different on RTP level. Assume a first scenario with two SSRC being sent from an end-point to a Transport Translator, and a second scenario with two single SSRC remote end- points sending to the same Transport Translator. The main differences between those two scenarios are that in the second



scenario, the RTT may vary between the SSRCs (but it is not guaranteed), and the SSRCs may also have different CNAMEs.

5.1.2. Compound RTCP Packets

When an end-point has multiple SSRCs and it needs to send RTCP packets on behalf of these SSRCs, the question arises if and how RTCP packets with different source SSRCs can be sent in the same compound packet. If it is allowed, then some consideration of the transmission scheduling is needed.

5.2. Point to Multipoint Using Multicast

This section discusses the Point to Multi-point using Multicast to interconnect the session participants. This needs to consider both Any Source Multicast (ASM) and Source-Specific Multicast (SSM). There are large commercial deployments of multicast for applications like IPTV.

+-----+ +---+ / \ +---+ | A |----/ \---| B | +---+ / Multi- \ +---+ + Cast + +---+ \ Network / +---+ | C |----\ /---| D | +---+ \ / +---+ +-----+

Figure 2: Point to Multipoint Using Any Source Multicast

In Any Source Multicast, any of the participants can send to all the other participants, simply by sending a packet to the multicast group. That is not possible in Source Specific Multicast [RFC4607] where only a single source (Distribution Source) can send to the multicast group, creating a topology that looks like the one below:



+--------+ +-----+ |Media | | | Source-specific |Sender 1|<----->| D S | Multicast +--------+ | I O | +--+----------------> R(1) | S U | | | | +--------+ | T R | | +-----------> R(2) | |Media |<----->| R C |->+ | : | | |Sender 2| | I E | | +------> R(n-1) | | +--------+ | B | | | | | | : | U | +--+--> R(n) | | | : | T +-| | | | | : | I | |<---------+ | | | +--------+ | O |F|<---------------+ | | |Media | | N |T|<--------------------+ | |Sender M|<----->| | |<-------------------------+ +--------+ +-----+ RTCP Unicast

FT = Feedback Target Transport from the Feedback Target to the Distribution Source is via unicast or multicast RTCP if they are not co-located.

Figure 3: Point to Multipoint using Source Specific Multicast

In this topology a number of Media Senders (1 to M) are allowed to send media to the SSM group, sends media to the distribution source which then forwards the media streams to the multicast group. The media streams reach the Receivers (R(1) to R(n)). The Receiver’s RTCP cannot be sent to the multicast group. To support RTCP, an RTP extension for SSM [RFC5760] was defined to use unicast transmission to send RTCP from the receivers to one or more Feedback Targets (FT).

As multicast is a one to many distribution system, this must be taken into consideration. For example, the only practical method for adapting the bit-rate sent towards a given receiver for large groups is to use a set of multicast groups, where each multicast group represents a particular bit-rate. Otherwise the whole group gets media adapted to the participant with the worst conditions. The media encoding is either scalable, where multiple layers can be combined, or simulcast where a single version is selected. By either selecting or combing multicast groups, the receiver can control the bit-rate sent on the path to itself. It is also common that streams that improve transport robustness is sent in its own multicast group to allow for interworking with legacy or to support different levels of protection.

The result of this is three common behaviors for RTP multicast:



1. Use of multiple RTP sessions for the same media type.

2. The need for identifying RTP sessions that are related in one of several possible ways.

3. The need for binding related SSRCs in different RTP sessions together.

This indicates that Multicast is an important consideration when working with the RTP multiplexing and multi stream architecture questions. It is also important to note that so far there is no special mode for basic behavior between multicast and unicast usages of RTP. Yes, there are extensions targeted to deal with multicast specific cases, but the general applicability does need to be considered.

5.3. Point to Multipoint Using an RTP Translator

Transport Translators (Relays) are a very important consideration for this document as they result in an RTP session situation that is very similar to how an ASM group RTP session would behave.

+---+ +------------+ +---+ | A |<---->| |<---->| B | +---+ | | +---+ | Translator | +---+ | | +---+ | C |<---->| |<---->| D | +---+ +------------+ +---+

Figure 4: Transport Translator (Relay)

One of the most important aspects with the simple relay is that it is both easy to implement and require minimal amount of resources as only transport headers are rewritten, no RTP modifications nor media transcoding occur. Thus it is most likely the cheapest and most generally deployable method for multi-point sessions. The most obvious downside of this basic relaying is that the translator has no control over how many streams needs to be delivered to a receiver. Nor can it simply select to deliver only certain streams, as it creates session inconsistencies. If some middlebox temporarily stops a stream, this prevents some receivers from reporting on it. From the senders perspective it will look like a transport failure. Applications having needs to stop or switch streams in the central node should consider using an RTP mixer to avoid this issue.

The Transport Translator does not need to have an SSRC of itself, nor need it send any RTCP reports on the flows that pass it, but it may



choose to do that.

Use of a transport translator results in that any of the end-points will receive multiple SSRCs over a single unicast transport flow from the translator. That is independent of the other end-points having only a single or several SSRCs. End-points that have multiple SSRCs put further requirements on how SSRCs can be related or bound within and across RTP sessions and how they can be identified on an application level. The transport translator has a signalling requirement that also exist in any source multicast; all of the participants will need to have the same RTP and payload type configuration. Otherwise, A could for example be using payload type 97 as the video codec H.264 while B thinks it is MPEG-2. It should be noted that SDP offer/answer [RFC3264] has issues with ensuring this property.

A Media Translator can perform a large variety of media functions affecting the media stream passing the translator, coming from one source and destined to a particular end-point. The translator can transcode to a different bit-rate, transcode to use another encoder, change the packetization of the media stream, add FEC streams, or terminate RTP retransmissions. The latter behaviors require the translator to use SSRCs that only exist in a particular sub-domain of the RTP session, and it may also create additional sessions when the mechanism applied on one side so requires.

5.4. Point to Multipoint Using an RTP Mixer

The most commonly used topology in centralized conferencing is based on the RTP Mixer. The main reason for this is that it provides a very consistent view of the RTP session towards each participant. That is accomplished through the mixer having its own SSRCs and any media sent to the participants will be sent using those SSRCs. If the mixer wants to identify the underlying media sources for its conceptual streams, it can identify them using CSRC. The media stream the mixer provides can be an actual media mixing of multiple media sources. It might also be as simple as selecting one of the underlying sources based on some mixer policy or control signalling.

+---+ +------------+ +---+ | A |<---->| |<---->| B | +---+ | | +---+ | Mixer | +---+ | | +---+ | C |<---->| |<---->| D | +---+ +------------+ +---+

Figure 5: RTP Mixer



In the case where the mixer does stream selection, an application may in fact desire multiple simultaneous streams but only as many as the mixer can handle. As long as the mixer and an end-point can agree on the maximum number of streams and how the streams that are delivered are selected, this provides very good functionality. As these streams are forwarded using the mixer’s SSRCs, there are no inconsistencies within the session.

5.5. Point to Multipoint using Multiple Unicast flows

Based on the RTP session definition, it is clearly possible to have a joint RTP session over multiple transport flows like the below three end-point joint session. In this case, A needs to send its’ media streams and RTCP packets to both B and C over their respective transport flows. As long as all participants do the same, everyone will have a joint view of the RTP session.

+---+ +---+ | A |<---->| B | +---+ +---+ ^ ^ \ / \ / v v +---+ | C | +---+

Figure 6: Point to Multi-Point using Multiple Unicast Transports

This doesn’t create any additional requirements beyond the need to have multiple transport flows associated with a single RTP session. Note that an end-point may use a single local port to receive all these transport flows, or it might have separate local reception ports for each of the end-points.

There exists an alternative structure for establishing the above communication scenario (Figure 6) which uses independent RTP sessions between each pair of peers, i.e. three different RTP sessions. Unless independently adapted the same RTP media stream could be sent in both of the RTP sessions an end-point has. The difference exists in the behaviors around RTCP, for example common RTCP bandwidth for one joint session, rather than three independent pools, and the awareness based on RTCP reports between the peers of how that third leg is doing.



5.6. De-composite End-Point

There is some possibility that an RTP end-point implementation in fact reside on multiple devices, each with their own network address. A very basic use case for this would be to separate audio and video processing for a particular end-point, like a conference room, into one device handling the audio and another handling the video, being interconnected by some control functions allowing them to behave as a single end-point.

+---------------------+ | End-point A | | Local Area Network | | +------------+ | | +->| Audio |<+----\ | | +------------+ | \ +------+ | | +------------+ | +-->| | | +->| Video |<+--------->| B | | | +------------+ | +-->| | | | +------------+ | / +------+ | +->| Control |<+----/ | +------------+ | +---------------------+

Figure 7: De-composite End-Point

In the above usage, let us assume that the RTP sessions are different for audio and video. The audio and video parts will use a common CNAME and also have a common clock to ensure that synchronization and clock drift handling works despite the decomposition.

However, if the audio and video were in a single RTP session then this use case becomes problematic. This as all transport flow receivers will need to receive all the other media streams that are part of the session. Thus the audio component will receive also all the video media streams, while the video component will receive all the audio ones, doubling the site’s bandwidth requirements from all other session participants. With a joint RTP session it also becomes evident that a given end-point, as interpreted from a CNAME perspective, has two sets of transport flows for receiving the streams and the decomposition is not hidden.

The requirements that can derived from the above usage is that the transport flows for each RTP session might be under common control but still go to what looks like different end-points based on addresses and ports. A conclusion can also be reached that decomposition without using separate RTP sessions has downsides and potential for RTP/RTCP issues.



There exist another use case which might be considered as a de- composite end-point. However, as will be shown this should be considered a translator instead. An example of this is when an endpoint A sends a media flow to B. On the path there is a device C that on A’s behalf does something with the media streams, for example adds an RTP session with FEC information for A’s media streams. C will in this case need to bind the new FEC streams to A’s media stream by using the same CNAME as A.

+------+ +------+ +------+ | | | | | | | A |------->| C |-------->| B | | | | |---FEC-->| | +------+ +------+ +------+

Figure 8: When De-composition is a Translator

This type of functionality where C does something with the media stream on behalf of A is clearly covered under the media translator definition (Section 5.3).

6. Multiple Streams Discussion

6.1. Introduction

Using multiple media streams is a well supported feature of RTP. However, it can be unclear for most implementers or people writing RTP/RTCP applications or extensions attempting to apply multiple streams when it is most appropriate to add an additional SSRC in an existing RTP session and when it is better to use multiple RTP sessions. This section tries to discuss the various considerations needed. The next section then concludes with some guidelines.

6.2. RTP/RTCP Aspects

This section discusses RTP and RTCP aspects worth considering when selecting between SSRC multiplexing and Session multiplexing.

6.2.1. The RTP Specification

RFC 3550 contains some recommendations and a bullet list with 5 arguments for different aspects of RTP multiplexing. Let’s review Section 5.2 of [RFC3550], reproduced below:

"For efficient protocol processing, the number of multiplexing points should be minimized, as described in the integrated layer processing design principle [ALF]. In RTP, multiplexing is provided by the



destination transport address (network address and port number) which is different for each RTP session. For example, in a teleconference composed of audio and video media encoded separately, each medium SHOULD be carried in a separate RTP session with its own destination transport address.

Separate audio and video streams SHOULD NOT be carried in a single RTP session and demultiplexed based on the payload type or SSRC fields. Interleaving packets with different RTP media types but using the same SSRC would introduce several problems:

1. If, say, two audio streams shared the same RTP session and the same SSRC value, and one were to change encodings and thus acquire a different RTP payload type, there would be no general way of identifying which stream had changed encodings.

2. An SSRC is defined to identify a single timing and sequence number space. Interleaving multiple payload types would require different timing spaces if the media clock rates differ and would require different sequence number spaces to tell which payload type suffered packet loss.

3. The RTCP sender and receiver reports (see Section 6.4) can only describe one timing and sequence number space per SSRC and do not carry a payload type field.

4. An RTP mixer would not be able to combine interleaved streams of incompatible media into one stream.

5. Carrying multiple media in one RTP session precludes: the use of different network paths or network resource allocations if appropriate; reception of a subset of the media if desired, for example just audio if video would exceed the available bandwidth; and receiver implementations that use separate processes for the different media, whereas using separate RTP sessions permits either single- or multiple-process implementations.

Using a different SSRC for each medium but sending them in the same RTP session would avoid the first three problems but not the last two.

On the other hand, multiplexing multiple related sources of the same medium in one RTP session using different SSRC values is the norm for multicast sessions. The problems listed above don’t apply: an RTP mixer can combine multiple audio sources, for example, and the same treatment is applicable for all of them. It may also be appropriate to multiplex streams of the same medium using different SSRC values in other scenarios where the last two problems do not apply."



Let’s consider one argument at a time. The first is an argument for using different SSRC for each individual media stream, which still is very applicable.

The second argument is advocating against using payload type multiplexing, which still stands as can been seen by the extensive list of issues found in Appendix A.

The third argument is yet another argument against payload type multiplexing.

The fourth is an argument against multiplexing media streams that require different handling into the same session. This is to simplify the processing at any receiver of the media stream. If all media streams that exist in an RTP session are of one media type and one particular purpose, there is no need for deeper inspection of the packets before processing them in both end-points and RTP aware middle nodes.

The fifth argument discusses network aspects that we will discuss more below in Section 6.5. It also goes into aspects of implementation, like decomposed end-points where different processes or inter-connected devices handle different aspects of the whole multi-media session.

A summary of RFC 3550’s view on multiplexing is to use unique SSRCs for anything that is its’ own media/packet stream, and secondly use different RTP sessions for media streams that don’t share media type and purpose, to maximize flexibility when it comes to processing and handling of the media streams.

This mostly agrees with the discussion and recommendations in this document. However, there has been an evolution of RTP since that text was written which needs to be reflected in the discussion. Additional clarifications for specific cases are also needed.

6.2.1.1. Different Media Types Recommendations

The above quote from RTP [RFC3550] includes a strong recommendation:

"For example, in a teleconference composed of audio and video media encoded separately, each medium SHOULD be carried in a separate RTP session with its own destination transport address."

It has been identified in "Why RTP Sessions Should Be Content Neutral" [I-D.alvestrand-rtp-sess-neutral] that the above statement is poorly supported by any of the motivations provided in the RTP specification. This document has a more detailed analysis of



potential issues in having multiple media types in the same RTP session in Section 6.7. An important influence for underlying thinking for the RTP design and likely this statement can be found in the academic paper by David Clark and David Tennenhouse "Architectural considerations for a new generation of protocols" [ALF].

6.2.2. Handling Varying sets of Senders

A potential issue that some application designers may need to consider is the case where the set of simultaneously active sources varies within a larger set of session members. As each media decoding chain may contain state, it is important that this type of usage ensures that a receiver can flush a decoding state for an inactive source and if that source becomes active again, it does not assume that this previous state exists.

This behavior will cause similar issues independent of SSRC or Session multiplexing. It might be possible in certain applications to limit the changes to a subset of communication session participants by have the sub-set use particular RTP Sessions.

6.2.3. Cross Session RTCP Requests

There currently exists no functionality to make truly synchronized and atomic RTCP messages with some type of request semantics across multiple RTP Sessions. Instead, separate RTCP messages will have to be sent in each session. This gives SSRC multiplexed streams a slight advantage as RTCP messages for different streams in the same session can be sent in a compound RTCP packet. Thus providing an atomic operation if different modifications of different streams are requested at the same time.

In Session multiplexed cases, the RTCP timing rules in the sessions and the transport aspects, such as packet loss and jitter, prevents a receiver from relying on atomic operations, forcing it to use more robust and forgiving mechanisms.

6.2.4. Binding Related Sources

A common problem in a number of various RTP extensions has been how to bind related sources together. This issue is common to SSRC multiplexing and Session Multiplexing, and any solution and recommendation related to the problem should work equally well with both methods to avoid creating barriers between using session multiplexing and SSRC multiplexing.

The current solutions do not have these properties. There exists one



solution for grouping RTP session together in SDP [RFC5888] to know which RTP session contains for example the FEC data for the source data in another session. However, this mechanism does not work on individual media flows and is thus not directly applicable to the problem. The other solution is also SDP based and can group SSRCs within a single RTP session [RFC5576]. Thus this mechanism can bind media streams in SSRC multiplexed cases. Both solutions have the shortcoming of being restricted to SDP based signalling and also do not work in cases where the session’s dynamic properties are such that it is difficult or resource consuming to keep the list of related SSRCs up to date.

One possible solution could be to mandate the same SSRC being used in all RTP session in case of session multiplexing. We do note that Section 8.3 of the RTP Specification [RFC3550] recommends using a single SSRC space across all RTP sessions for layered coding. However this recommendation has some downsides and is less applicable beyond the field of layered coding. To use the same sender SSRC in all RTP sessions from a particular end-point can cause issues if an SSRC collision occurs. If the same SSRC is used as the required binding between the streams, then all streams in the related RTP sessions must change their SSRC. This is extra likely to cause problems if the participant populations are different in the different sessions. For example, in case of large number of receivers having selected totally random SSRC values in each RTP session as RFC 3550 specifies, a change due to a SSRC collision in one session can then cause a new collision in another session. This cascading effect is not severe but there is an increased risk that this occurs for well populated sessions. In addition, being forced to change the SSRC affects all the related media streams; instead of having to re-synchronize only the originally conflicting stream, all streams will suddenly need to be re-synchronized with each other. This will prevent also the media streams not having an actual collision from being usable during the re-synchronization and also increases the time until synchronization is finalized. In addition, it requires exception handling in the SSRC generation.

The above collision issue does not occur in case of having only one SSRC space across all sessions and all participants will be part of at least one session, like the base layer in layered encoding. In that case the only downside is the special behavior that needs to be well defined by anyone using this. But, having an exception behavior where the SSRC space is common across all session is an issue as this behavior does not fit all the RTP extensions or payload formats. It is possible to create a situation where the different mechanisms cannot be combined due to the non standard SSRC allocation behavior.

Existing mechanisms with known issues:



RTP Retransmission (RFC4588): Has two modes, one for SSRC multiplexing and one for Session multiplexing. The session multiplexing requires the same CNAME and mandates that the same SSRC is used in both sessions. Using the same SSRC does work but will potentially have issues in certain cases. In SSRC multiplexed mode the CNAME is used to bind media and retransmission streams together. However, if multiple media streams are sent from the same end-point in the same session this does not provide non-ambiguous binding. Therefore when the first retransmission request for a media stream is sent, one must not have another retransmission request outstanding for an SSRC which don’t have a binding between the original SSRC and the retransmission stream’s SSRC. This works but creates some limitations that can be avoided by a more explicit mechanism. The SDP based ssrc-group mechanism is sufficient in this case as long as the application can rely on the signalling based solution.

Scalable Video Coding (RFC6190): As an example of scalable coding, SVC [RFC6190] has various modes. The Multi Session Transmission (MST) uses Session multiplexing to separate scalability layers. However, this specification has failed to be explicit on how these layers are bound together in cases where CNAME is not sufficient. CNAME is no longer sufficient when more than one media source occur within a session that has the same CNAME, for example due to multiple video cameras capturing the same lecture hall. This likely implies that a single SSRC space as recommend by Section 8.3 of RTP [RFC3550] is to be used.

Forward Error Correction: If some type of FEC or redundancy stream is being sent, it needs its own SSRC, with the exception of constructions like redundancy encoding [RFC2198]. Thus in case of transmitting the FEC in the same session as the source data, the inter SSRC relation within a session is needed. In case of sending the redundant data in a separate session from the source, the SSRC in each session needs to be related. This occurs for example in RFC5109 when using session separation of original and FEC data. SSRC multiplexing is not supported, only using redundant encoding is supported.

This issue appears to need action to harmonize and avoid future shortcomings in extension specifications. A proposed solution for handling this issue is [I-D.westerlund-avtext-rtcp-sdes-srcname].

6.2.5. Forward Error Correction

There exist a number of Forward Error Correction (FEC) based schemes for how to reduce the packet loss of the original streams. Most of the FEC schemes will protect a single source flow. The protection is



achieved by transmitting a certain amount of redundant information that is encoded such that it can repair one or more packet loss over the set of packets they protect. This sequence of redundant information also needs to be transmitted as its own media stream, or in some cases instead of the original media stream. Thus many of these schemes create a need for binding the related flows as discussed above. They also create additional flows that need to be transported. Looking at the history of these schemes, there is both SSRC multiplexed and Session multiplexed solutions and some schemes that support both.

Using a Session multiplexed solution provides good support for legacy when deploying FEC or changing the scheme used, in the sense that it supports the case where some set of receivers may not be able to utilize the FEC information. By placing it in a separate RTP session, it can easily be ignored.

In usages involving multicast, having the FEC information on its own multicast group and RTP session allows for flexibility, for example when using Rapid Acquisition of Multicast Groups (RAMS) [RFC6285]. During the RAMS burst where data is received over unicast and where it is possible to combine with unicast based retransmission [RFC4588], there is no need to burst the FEC data related to the burst of the source media streams needed to catch up with the multicast group. This saves bandwidth to the receiver during the burst, enabling quicker catch up. When the receiver has caught up and joins the multicast group(s) for the source, it can at the same time join the multicast group with the FEC information. Having the source stream and the FEC in separate groups allow for easy separation in the Burst/Retransmission Source (BRS) without having to individually classify packets.

6.2.6. Transport Translator Sessions

A basic Transport Translator relays any incoming RTP and RTCP packets to the other participants. The main difference between SSRC multiplexing and Session multiplexing resulting from this use case is that for SSRC multiplexing it is not possible for a particular session participant to decide to receive a subset of media streams. When using separate RTP sessions for the different sets of media streams, a single participant can choose to leave one of the sessions but not the other.

6.3. Interworking

There are several different kinds of interworking, and this section discusses two related ones. The interworking between different applications and the implications of potentially different choices of



usage of RTP’s multiplexing points. The second topic relates to what limitations may have to be considered working with some legacy applications.

6.3.1. Interworking Applications

It is not uncommon that applications or services of similar usage, especially the ones intended for interactive communication, ends up in a situation where one want to interconnect two or more of these applications. From an RTP perspective this could be problem free if all the applications have made the same multiplexing choices, have the same capabilities in number of simultaneous media streams combined with the same set of RTP/RTCP extensions being supported. Unfortunately this may not always be true.

In these cases one ends up in a situation where one might use a gateway to interconnect applications. This gateway then needs to change the multiplexing structure or adhere to limitations in each application. If one’s goal is to make minimal amount of work in such a gateway, there are some multiplexing choices that one should avoid. The lowest amount of work represents solutions where one can take an SSRC from one RTP session in one application and forward it into another RTP session. For example if one has one application that has multiple SSRCs for one media type in one session and another application that instead has chosen to use multiple RTP sessions with only a single SSRC per end-point in each of these sessions. Then mapping an SSRC from the side with one session into an RTP session is possible. However mapping SSRC from different RTP sessions into a single RTP session has the potential of creating SSRC collisions, especially if an end-point has not generated independent random SSRC values in each RTP session. This issue is even more likely in a case where one side uses a single RTP session with multiple media types and the other uses different RTP session for different media or robustness mechanism such as retransmission [RFC4588]. Then it is more likely or maybe even required to use the same SSRC in the different RTP sessions.

In cases where the used structure is incompatible, the gateway will need to make SSRC translation. Thus this incurs overhead and some potential loss of functionality. First of all, if one translates the SSRC in an RTP header then one will be forced to decrypt and re- encrypt if one uses SRTP and thus also needs to be part of the security association. Secondly, changing the SSRC also means that one needs to translate all RTCP messages. This can be more complex, but important so that the gateway does not end up having to terminate the end-to-end RTCP chain. In that case the gateway will need to be able to take the role of a true end-point in each session, which may include functions such as bit-rate adaptation and correctly respond



to whatever RTCP extensions are being used, and then translate them or locally respond to them. Thirdly, an SSRC translation may require that one changes RTP payloads; for example, an RTP retransmission packet contains an original sequence number that must match the sequence number used in for the corresponding packet with the new SSRC. And for FEC packets this is even worse, as the original SSRC is included as part of the data for which FEC redundant data is calculated. A fourth issue is the potential for these gateways to block evolution of the applications by blocking unknown RTP and RTCP extensions that the regular application has been extended with.

If one uses security functions, like SRTP, they can as seen above incur both additional risk due to the gateway needing to be in security association between the end-points, unless the gateway is on the transport level, and additional complexities in form of the decrypt-encrypt cycles needed for each forwarded packet. SRTP, due to its keying structure, also makes it hard to move a flow from one RTP session to another as each RTP session will have one or more different master keys and these must not be the same in multiple RTP sessions as that can result in two-time pads that completely breaks the confidentiality of the packets.

An additional issue around interworking is that for multi-party applications it can be impossible to judge which different RTP multiplexing behaviors that will be used by end-points that attempt to join a session. Thus if one attempts to use a multiplexing choice that has poor interworking, one may have to switch at a later stage when someone wants to participate in a multi-party session using an RTP application supporting only another behavior. It is likely difficult to implement the switch without some media disruption.

To summarize, certain types of applications are likely to be inter- worked. Sets of applications of similar type should strive to use the same multiplexing structure to avoid the need to make an RTP session level gateway. This as it incurs complexity costs, can force the gateway to be part of security associations, force SSRC translation and even payload translation which is also a potential hinder to application evolution.

6.3.2. Multiple SSRC Legacy Considerations

Historically, the most common RTP use cases have been point to point Voice over IP (VoIP) or streaming applications, commonly with no more than one media source per end-point and media type (typically audio and video). Even in conferencing applications, especially voice only, the conference focus or bridge has provided a single stream with a mix of the other participants to each participant. It is also common to have individual RTP sessions between each end-point and the



RTP mixer.

When establishing RTP sessions that may contain end-points that aren’t updated to handle multiple streams following these recommendations, a particular application can have issues with multiple SSRCs within a single session. These issues include:

1. Need to handle more than one stream simultaneously rather than replacing an already existing stream with a new one.

2. Be capable of decoding multiple streams simultaneously.

3. Be capable of rendering multiple streams simultaneously.

RTP Session multiplexing could potentially avoid these issues if there is only a single SSRC at each end-point, and in topologies which appears like point to point as seen the end-point. However, forcing the usage of session multiplexing due to this reason would be a great mistake, as it is likely that a significant set of applications will need a combination of SSRC multiplexing of several media sources and session multiplexing for other aspects such as encoding alternatives, adding robustness or simply to support legacy. However, this issue does need consideration when deploying multiple media streams within an RTP session where legacy end-points may occur.

6.4. Signalling Aspects

There exist various signalling solutions for establishing RTP sessions. Many are SDP [RFC4566] based, however SDP functionality is also dependent on the signalling protocols carrying the SDP. Where RTSP [RFC2326] and SAP [RFC2974] both use SDP in a declarative fashion, while SIP [RFC3261] uses SDP with the additional definition of Offer/Answer [RFC3264]. The impact on signalling and especially SDP needs to be considered as it can greatly affect how to deploy a certain multiplexing point choice.

6.4.1. Session Oriented Properties

One aspect of the existing signalling is that it is focused around sessions, or at least in the case of SDP the media description. There are a number of things that are signalled on a session level/ media description but those are not necessarily strictly bound to an RTP session and could be of interest to signal specifically for a particular media stream (SSRC) within the session. The following properties have been identified as being potentially useful to signal not only on RTP session level:



o Bitrate/Bandwidth exist today only at aggregate or a common any media stream limit

o Which SSRC that will use which RTP Payload Types

Some of these issues are clearly SDP’s problem rather than RTP limitations. However, if the aim is to deploy an SSRC multiplexed solution that contains several sets of media streams with different properties (encoding/packetization parameter, bit-rate, etc), putting each set in a different RTP session would directly enable negotiation of the parameters for each set. If insisting on SSRC multiplexing only, a number of signalling extensions are needed to clarify that there are multiple sets of media streams with different properties and that they shall in fact be kept different, since a single set will not satisfy the application’s requirements.

This does in fact create a strong driver to use RTP session multiplexing for any case where different sets of media streams with different requirements exist.

6.4.2. SDP Prevents Multiple Media Types

SDP encoded in its structure prevention against using multiple media types in the same RTP session. A media description in SDP can only have a single media type; audio, video, text, image, application. This media type is used as the top-level media type for identifying the actual payload format bound to a particular payload type using the rtpmap attribute. Thus a high fence against using multiple media types in the same session was created.

There is an accepted WG item in the MMUSIC WG to define how multiple media lines describe a single underlying transport [I-D.holmberg-mmusic-sdp-bundle-negotiation] and thus it becomes possible in SDP to define one RTP session with multiple media types.

6.4.3. Media Stream Usage

Media streams being transported in RTP has some particular usage in an RTP application. This usage of the media stream is in many applications so far implicitly signalled. For example by having all audio media streams arriving in the only audio RTP session they are to be decoded, mixed and played out. However, in more advanced applications that use multiple media streams there will be more than a single usage or purpose among the set of media streams being sent or received. RTP applications will need to signal this usage somehow. Here the choice of SSRC multiplexing versus session multiplexing will have significant impact. If one uses SSRC multiplexing to its full extent one will have to explicitly indicate



for each SSRC what its’ usage and purpose are using some signalling between the application instances.

This SSRC usage signalling will have some impact on the application and also on any central RTP nodes. It is important in the design to consider the implications of the need for additional signalling between the nodes. One consideration is if a receiver can utilize the media stream at all before it has received the signalling message describing the media stream and its usage. Another consideration is that any RTP central node, like an RTP mixer or translator that selects, mixes or processes streams, in most cases will need to receive the same signalling to know how to treat media streams with different usage in the right fashion.

Application designers should consider putting media streams of the same usage and/or receiving the same treatment in middleboxes in the same RTP sessions and use the RTP session as an explicit indication of how to deal with media streams. By having session level indication of usage and have different RTP sessions for different usages, the need for stream specific signalling can be reduced. Especially signalling of the type that is time critical and needs to be provided prior to the media stream being available.

6.5. Network Aspects

The multiplexing choice has impact on network level mechanisms that need to be considered by the implementor.

6.5.1. Quality of Service

When it comes to Quality of Service mechanisms, they are either flow based or marking based. RSVP [RFC2205] is an example of a flow based mechanism, while Diff-Serv [RFC2474] is an example of a Marking based one. For a marking based scheme, the method of multiplexing will not affect the possibility to use QoS.

However, for a flow based scheme there is a clear difference between the methods. SSRC multiplexing will result in all media streams being part of the same 5-tuple (protocol, source address, destination address, source port, destination port) which is the most common selector for flow based QoS. Thus, separation of the level of QoS between media streams is not possible. That is however possible for session based multiplexing, where each different version can be in a different RTP session that can be sent over different 5-tuples.



6.5.2. NAT and Firewall Traversal

In today’s network there exist a large number of middleboxes. The ones that normally have most impact on RTP are Network Address Translators (NAT) and Firewalls (FW).

Below we analyze and comment on the impact of requiring more underlying transport flows in the presence of NATs and Firewalls:

End-Point Port Consumption: A given IP address only has 65536 available local ports per transport protocol for all consumers of ports that exist on the machine. This is normally never an issue for an end-user machine. It can become an issue for servers that handle large number of simultaneous streams. However, if the application uses ICE to authenticate STUN requests, a server can serve multiple end-points from the same local port, and use the whole 5-tuple (source and destination address, source and destination port, protocol) as identifier of flows after having securely bound them to the remote end-point address using the STUN request. In theory the minimum number of media server ports needed are the maximum number of simultaneous RTP Sessions a single end-point may use. In practice, implementation will probably benefit from using more server ports to simplify implementation or avoid performance bottlenecks.

NAT State: If an end-point sits behind a NAT, each flow it generates to an external address will result in a state that has to be kept in the NAT. That state is a limited resource. In home or Small Office/Home Office (SOHO) NATs, memory or processing are usually the most limited resources. For large scale NATs serving many internal end-points, available external ports are typically the scarce resource. Port limitations is primarily a problem for larger centralized NATs where end-point independent mapping requires each flow to use one port for the external IP address. This affects the maximum number of internal users per external IP address. However, it is worth pointing out that a real-time video conference session with audio and video is likely using less than 10 UDP flows, compared to certain web applications that can use 100+ TCP flows to various servers from a single browser instance.

NAT Traversal Excess Time: Making the NAT/FW traversal takes a certain amount of time for each flow. It also takes time in a phase of communication between accepting to communicate and the media path being established which is fairly critical. The best case scenario for how much extra time it can take following the specified ICE procedures are: 1.5*RTT + Ta*(Additional_Flows-1), where Ta is the pacing timer, which ICE specifies to be no smaller than 20 ms. That assumes a message in one direction, and then an



immediate triggered check back. This as ICE first finds one candidate pair that works prior to establish multiple flows. Thus, there is no extra time until one has found a working candidate pair. Based on that working pair the needed extra time is to in parallel establish the, in most cases 2-3, additional flows.

NAT Traversal Failure Rate: Due to the need to establish more than a single flow through the NAT, there is some risk that establishing the first flow succeeds but that one or more of the additional flows fail. The risk that this happens is hard to quantify, but it should be fairly low as one flow from the same interfaces has just been successfully established. Thus only rare events such as NAT resource overload, or selecting particular port numbers that are filtered etc, should be reasons for failure.

Deep Packet Inspection and Multiple Streams: Firewalls differ in how deeply they inspect packets. There exist some potential that deeply inspecting firewalls will have similar legacy issues with multiple SSRCs as some stack implementations.

SSRC multiplexing keeps additional media streams within one RTP Session and does not introduce any additional NAT traversal complexities per media stream. In contrast, the session multiplexing is using one RTP session per media stream. Thus additional lower layer transport flows will be required, unless an explicit demultiplexing layer is added between RTP and the transport protocol. A proposal for how to multiplex multiple RTP sessions over the same single lower layer transport exist in [I-D.westerlund-avtcore-single-transport-multiplexing].

6.5.3. Multicast

Multicast groups provides a powerful semantics for a number of real- time applications, especially the ones that desire broadcast-like behaviors with one end-point transmitting to a large number of receivers, like in IPTV. But that same semantics do result in a certain number of limitations.

One limitation is that for any group, sender side adaptation to the actual receiver properties causes degradation for all participants to what is supported by the receiver with the worst conditions among the group participants. In most cases this is not acceptable. Instead various receiver based solutions are employed to ensure that the receivers achieve best possible performance. By using scalable encoding and placing each scalability layer in a different multicast group, the receiver can control the amount of traffic it receives. To have each scalability layer on a different multicast group, one



RTP session per multicast group is used.

If instead a single RTP session over multiple transports were to be deployed, i.e. multicast groups with each layer as it’s own SSRC, then very different views of the RTP session would exist. That as one receiver may see only a single layer (SSRC), while another may see three SSRCs if it joined three multicast groups. This would cause disjoint RTCP reports where a management system would not be able to determine if a receiver isn’t reporting on a particular SSRC due to that it is not a member of that multicast group, or because it doesn’t receive it as a result of a transport failure.

Thus it appears easiest and most straightforward to use multiple RTP sessions. In addition, the transport flow considerations in multicast are a bit different from unicast. First of all there is no shortage of port space, as each multicast group has its own port space.

6.5.4. Multiplexing multiple RTP Session on a Single Transport

For applications that doesn’t need flow based QoS and like to save ports and NAT/FW traversal costs and where usage of multiple media types in one RTP session is not suitable, there is a proposal for how to achieve multiplexing of multiple RTP sessions over the same lower layer transport [I-D.westerlund-avtcore-single-transport-multiplexing]. Using such a solution would allow session multiplexing without most of the perceived downsides of additional RTP sessions creating a need for additional transport flows.

6.6. Security Aspects

On the basic level there is no significant difference in security when having one RTP session and having multiple. However, there are a few more detailed considerations that might need to be considered in certain usages.

6.6.1. Security Context Scope

When using SRTP [RFC3711] the security context scope is important and can be a necessary differentiation in some applications. As SRTP’s crypto suites (so far) is built around symmetric keys, the receiver will need to have the same key as the sender. This results in that no one in a multi-party session can be certain that a received packet really was sent by the claimed sender or by another party having access to the key. In most cases this is a sufficient security property, but there are a few cases where this does create situations.



The first case is when someone leaves a multi-party session and one wants to ensure that the party that left can no longer access the media streams. This requires that everyone re-keys without disclosing the keys to the excluded party.

A second case is when using security as an enforcing mechanism for differentiation. Take for example a scalable layer or a high quality simulcast version which only premium users are allowed to access. The mechanism preventing a receiver from getting the high quality stream can be based on the stream being encrypted with a key that user can’t access without paying premium, having the key-management limit access to the key.

In the latter case it is likely easiest from signalling, transport (if done over multicast) and security to use a different RTP session. That way the user(s) not intended to receive a particular stream can easily be excluded. There is no need to have SSRC specific keys, which many of the key-management systems cannot handle.

6.6.2. Key-Management for Multi-party session

Performing key-management for Multi-party session can be a challenge. This section considers some of the issues.

Transport translator based session cannot use Security Description [RFC4568] nor DTLS-SRTP [RFC5764] without an extension as each endpoint provides its set of keys. In centralized conference, the signalling counterpart is a conference server and the media plane unicast counterpart (to which DTLS messages would be sent) is the translator. Thus an extension like Encrypted Key Transport [I-D.ietf-avt-srtp-ekt] is needed or a MIKEY [RFC3830] based solution that allows for keying all session participants with the same master key.

Keying of multicast transported SRTP face similar challenges as the transport translator case.

6.6.3. Complexity Implications

The usage of security functions can surface complexity implications of the choice of multiplexing and topology. This becomes especially evident in RTP topologies having any type of middlebox that processes or modifies RTP/RTCP packets. Where there is very small overhead for a not secured RTP translator or mixer to rewrite an SSRC value in the RTP packet, the cost of doing it when using cryptographic security functions is higher. For example if using SRTP [RFC3711], the actual security context and exact crypto key are determined by the SSRC field value. If one changes it, the encryption and authentication



tag must be performed using another key. Thus changing the SSRC value implies a decryption using the old SSRC and its security context followed by an encryption using the new one.

There exist many valid cases where a middlebox will be forced to perform such cryptographic operations due to the intended purpose of the middlebox, for example a media transcoding RTP translator cannot avoid performing these operations as they will produce a different payload compared to the input. However, there exist some cases where another topology and/or multiplexing choice could avoid the complexities.

6.7. Multiple Media Types in one RTP session

Having different media types, like audio and video, in the same RTP sessions is not forbidden, only recommended against as earlier discussed in Section 6.2.1.1. When using multiple media types, there are a number of considerations:

Payload Type gives Media Type: This solution is dependent on getting the media type from the Payload Type. Thus overloading this demultiplexing point in a receiver making it serve two purposes. First to provide the main media type and determining the processing chain, then later for the exact configuration of the encoder and packetization.

Payload Type field limitations: The total number of Payload Types available to use in an RTP session is fairly limited, especially if Multiplexing RTP Data and Control Packets on a Single Port [RFC5761] is used. For certain applications negotiating a large set of codes and configuration this may become an issue.

An SSRC cannot use two clock rates simultaneously: The used RTP clock rate for an SSRC is determined from the payload type. As discussed in Appendix A it is not possible to simultaneously use two different clock rates for the same SSRC. Even switching clock rate once has potential issues if packet loss occurs at the same time. Different media types commonly have different clock rates preventing or creating issues to use two different media types for the same SSRC.

Do not switch media types for an SSRC: The primary reasons to avoid switching from sending for example audio to sending video using the same SSRC is the implications on a receiver. When this happens, the processing chain in the receiver will have to switch from one media type to another. As the different media type’s entire processing chains are different and are connected to different outputs it is difficult to reuse the decoding chain,



which a normal codec change likely can. Instead the entire processing chain has to be torn down and replaced. In addition, there is likely a clock rate switching problem, possibly resulting in synchronization loss at the point of switching media type if some packet loss occurs. So this is a behavior that shall be avoided.

RTCP Bit-rate Issues: If the media types are significantly different in bit-rate, the RTCP bandwidth rates assigned to each source in a session can result in interesting effects, like that the RTCP bit- rate share for an audio stream is larger than the actual audio bit-rate. In itself this doesn’t cause any conflicts, only potentially unnecessary overhead. It is possible to avoid this using AVPF or SAVPF and setting trr-int parameter, which can bring down unnecessary regular reporting while still allowing for rapid feedback.

De-composite end-points: De-composite nodes that rely on the regular network to separate audio and video to different devices do not work well with this session setup. If they are forced to work, all media receiver parts of a de-composite end-point will receive all media, thus doubling the bit-rate consumption for the endpoint.

Flow based QoS Separation: Flow based QoS mechanisms will see all the media streams in the RTP session as part of a single flow. Therefore there is no possibility to provide separated QoS behavior for the different media types or flows.

RTP Mixers and Translators: An RTP mixer or Media Translator will also have to support this particular session setup, where it before could rely on the RTP session to determine what processing options should be applied to the incoming packets.

Legacy Implementations: The use of multiple media types has the potential for even larger issues with legacy implementations than single media type SSRC multiplexing due to the occurrence of multiple media types among the payload type configurations.

As can be seen, there is nothing in here that prevents using a single RTP session for multiple media types, however it does create a number of limitations and special case implementation requirements. So anyone considering using this setup should carefully review if the reasons for using a single RTP session are sufficient to motivate the needed special handling.



7. Arch-Types

This section discusses some arch-types of how RTP multiplexing can be used in applications to achieve certain goals and a summary of their implications. For each arch-type there is discussion of benefits and downsides.

7.1. Single SSRC per Session

In this arch-type each end-point in a point-to-point session has only a single SSRC, thus the RTP session contains only two SSRCs, one local and one remote. This session can be used both unidirectional, i.e. only a single media stream or bi-directional, i.e. both end- points have one media stream each. If the application needs additional media flows between the end-points, they will have to establish additional RTP sessions.

The Pros:

1. This arch-type has great legacy interoperability potential as it will not tax any RTP stack implementations.

2. The signalling has good possibilities to negotiate and describe the exact formats and bit-rates for each media stream, especially using today’s tools in SDP.

3. It does not matter if usage or purpose of the media stream is signalled on media stream level or session level as there is no difference.

4. It is possible to control security association per RTP session with current key-management.

The Cons:

a. The number of required RTP sessions cannot really be higher, which has the implications:

* Linear growth of the amount of NAT/FW state with number of media streams.

* Increased delay and resource consumption from NAT/FW traversal.

* Likely larger signalling message and signalling processing requirement due to the amount of session related information.



* Higher potential for a single media stream to fail during transport between the end-points.

b. When the number of RTP sessions grows, the amount of explicit state for relating media stream also grows, linearly or possibly exponentially, depending on how the application needs to relate media streams.

c. The port consumption may become a problem for centralized services, where the central node’s port consumption grows rapidly with the number of sessions.

d. For applications where the media streams are highly dynamic in their usage, i.e. entering and leaving, the amount of signalling can grow high. Issues arising from the timely establishment of additional RTP sessions can also arise.

e. Cross session RTCP requests needs is likely to exist and may cause issues.

f. If the same SSRC value is reused in multiple RTP sessions rather than being randomly chosen, interworking with applications that uses another multiplexing structure than this application will have issues and require SSRC translation.

g. Cannot be used with Any Source Multicast (ASM) as one cannot guarantee that only two end-points participate as packet senders. Using SSM, it is possible to restrict to these requirements if no RTCP feedback is used.

h. For most security mechanisms, each RTP session or transport flow requires individual key-management and security association establishment thus increasing the overhead.

i. Does not support multiparty session within a session. Instead each multi-party participant will require an individual RTP session to a given end-point, even if a central node is used.

RTP applications that need to inter-work with legacy RTP applications, like VoIP and video conferencing, can potentially benefit from this structure. However, a large number of media descriptions in SDP can also run into issues with existing implementations. For any application needing a larger number of media flows, the overhead can become very significant. This structure is also not suitable for multi-party sessions, as any given media stream from each participant, although having same usage in the application, must have its own RTP session. In addition, the dynamic behavior that can arise in multi-party applications can tax the



signalling system and make timely media establishment more difficult.

7.2. Multiple SSRCs of the Same Media Type

In this arch-type, each RTP session serves only a single media type. The RTP session can contain multiple media streams, either from a single end-point or due to multiple end-points. This commonly creates a low number of RTP sessions, typically only two one for audio and one for video with a corresponding need for two listening ports when using RTP and RTCP multiplexing.

The Pros:

1. Low number of RTP sessions needed compared to single SSRC case. This implies:

* Reduced NAT/FW state

* Lower NAT/FW Traversal Cost in both processing and delay.

2. Allows for early de-multiplexing in the processing chain in RTP applications where all media streams of the same type have the same usage in the application.

3. Works well with media type de-composite end-points.

4. Enables Flow-based QoS with different prioritization between media types.

5. For applications with dynamic usage of media streams, i.e. they come and go frequently, having much of the state associated with the RTP session rather than an individual SSRC can avoid the need for in-session signalling of meta-information about each SSRC.

6. Low overhead for security association establishment.

The Cons:

a. May have some need for cross session RTCP requests for things that affect both media types in an asynchronous way.

b. Some potential for concern with legacy implementations that does not support the RTP specification fully when it comes to handling multiple SSRC per end-point.

c. Will not be able to control security association for sets of media streams within the same media type with today’s key- management mechanisms, only between SDP media descriptions.



For RTP applications where all media streams of the same media type share same usage, this structure provides efficiency gains in amount of network state used and provides more faith sharing with other media flows of the same type. At the same time, it is still maintaining almost all functionalities when it comes to negotiation in the signalling of the properties for the individual media type and also enabling flow based QoS prioritization between media types. It handles multi-party session well, independently of multicast or centralized transport distribution, as additional sources can dynamically enter and leave the session.

7.3. Multiple Sessions for one Media type

In this arch-type one goes one step further than in the above (Section 7.2) by using multiple RTP sessions also for a single media type. The main reason for going in this direction is that the RTP application needs separation of the media streams due to their usage. Some typical reasons for going to this arch-type are scalability over multicast, simulcast, need for extended QoS prioritization of media streams due to their usage in the application, or the need for fine granular signalling using today’s tools.

The Pros:

1. More suitable for Multicast usage where receivers can individually select which RTP sessions they want to participate in, assuming each RTP session has its own multicast group.

2. Detailed indication of the application’s usage of the media stream, where multiple different usages exist.

3. Less need for SSRC specific explicit signalling for each media stream and thus reduced need for explicit and timely signalling.

4. Enables detailed QoS prioritization for flow based mechanisms.

5. Works well with de-composite end-points.

6. Handles dynamic usage of media streams well.

7. For transport translator based multi-party sessions, this structure allows for improved control of which type of media streams an end-point receives.

8. The scope for who is included in a security association can be structured around the different RTP sessions, thus enabling such functionality with existing key-management.



The Cons:

a. Increases the amount of RTP sessions compared to Multiple SSRCs of the Same Media Type.

b. Increased amount of session configuration state.

c. May need synchronized cross-session RTCP requests and require some consideration due to this.

d. For media streams that are part of scalability, simulcast or transport robustness it will be needed to bind sources, which must support multiple RTP sessions.

e. Some potential for concern with legacy implementations that does not support the RTP specification fully when it comes to handling multiple SSRC per end-point.

f. Higher overhead for security association establishment.

g. If the applications need finer control than on media type level over which session participants that are included in different sets of security associations, most of today’s key-management will have difficulties establishing such a session.

For more complex RTP applications that have several different usages for media streams of the same media type and / or uses scalability or simulcast, this solution can enable those functions at the cost of increased overhead associated with the additional sessions. This type of structure is suitable for more advanced applications as well as multicast based applications requiring differentiation to different participants.

7.4. Multiple Media Types in one Session

This arch-type is to use a single RTP session for multiple different media types, like audio and video, and possibly also transport robustness mechanisms like FEC or Retransmission. Each media stream will use its own SSRC and a given SSRC value from a particular endpoint will never use the SSRC for more than a single media type.

The Pros:

1. Single RTP session which implies:

* Minimal NAT/FW state.



* Minimal NAT/FW Traversal Cost.

* Fate-sharing for all media flows.

2. Enables separation of the different media types based on the payload types so media type specific end-point or central processing can still be supported despite single session.

3. Can handle dynamic allocations of media streams well on an RTP level. Depends on the application’s needs for explicit indication of the stream usage and how timely that can be signalled.

4. Minimal overhead for security association establishment.

The Cons:

a. Not suitable for interworking with other applications that uses individual RTP sessions per media type or multiple sessions for a single media type, due to high risk of forced SSRC translation.

b. Negotiation of bandwidth for the different media types is currently not possible in SDP. This requires SDP extensions to enable payload or source specific bandwidth. Likely to be a problem due to media type asymmetry in required bandwidth.

c. Does enforce higher bandwidth and processing on de-composite end- points.

d. Flow based QoS cannot provide separate treatment to some media streams compared to other in the single RTP session.

e. If there is significant asymmetry between the media streams RTCP reporting needs, there are some challenges in configuration and usage to avoid wasting RTCP reporting on the media stream that does not need that frequent reporting.

f. Not suitable for applications where some receivers like to receive only a subset of the media streams, especially if multicast or transport translator is being used.

g. Additional concern with legacy implementations that does not support the RTP specification fully when it comes to handling multiple SSRC per end-point, as also multiple simultaneous media types needs to be handled.

h. If the applications need finer control over which session participants that are included in different sets of security



associations, most key-management will have difficulties establishing such a session.

The analysis in this document and considerations in Section 6.7 implies that this is suitable only in a set of restricted use cases. The aspect in the above list that can be most difficult to judge long term is likely the potential need for interworking with other applications and services.

7.5. Summary

There are some clear relations between these arch-types. Both the "single SSRC per RTP session" and the "multiple media types in one session" are cases which require full explicit signalling of the media stream relations. However, they operate on two different levels where the first primarily enables session level binding, and the second needs to do it all on SSRC level. From another perspective, the two solutions are the two extreme points when it comes to number of RTP sessions required.

The two other arch-types "Multiple SSRCs of the Same Media Type" and "Multiple Sessions for one Media Type" are examples of two other cases that first of all allows for some implicit mapping of the role or usage of the media streams based on which RTP session they appear in. It thus potentially allows for less signalling and in particular reduced need for real-time signalling in dynamic sessions. They also represent points in between the first two when it comes to amount of RTP sessions established, i.e. representing an attempt to reduce the amount of sessions as much as possible without compromising the functionality the session provides both on network level and on signalling level.

8. Guidelines

This section contains a number of recommendations for implementors or specification writers when it comes to handling multi-stream.

Do not Require the same SSRC across Sessions: As discussed in Section 6.2.4 there exist drawbacks in using the same SSRC in multiple RTP sessions as a mechanism to bind related media streams together. It is instead recommended that a mechanism to explicitly signal the relation is used, either in RTP/RTCP or in the used signalling mechanism that establishes the RTP session(s).



Use SSRC multiplexing for additional Media Sources: In the cases an RTP end-point needs to transmit additional media source(s) of the same media type and purpose in the application, it is recommended to send them as additional SSRCs in the same RTP session. For example a tele-presence room where there are three cameras, and each camera captures 2 persons sitting at the table, sending each camera as its own SSRC within a single RTP session is recommended.

Use additional RTP sessions for streams with different purposes: When media streams have different purpose or processing requirements it is recommended that the different types of streams are put in different RTP sessions.

When using Session Multiplexing use grouping: When using Session Multiplexing solutions, it is recommended to be explicitly group the involved RTP sessions using the signalling mechanism, for example The Session Description Protocol (SDP) Grouping Framework. [RFC5888], using some appropriate grouping semantics.

RTP/RTCP Extensions May Support SSRC and Session Multiplexing: When defining an RTP or RTCP extension, the creator needs to consider if this extension is applicable in both SSRC multiplexed and Session multiplexed usages. Any extension intended to be generic is recommended to support both. Applications that are not as generally applicable will have to consider if interoperability is better served by defining a single solution or providing both options.

Transport Support Extensions: When defining new RTP/RTCP extensions intended for transport support, like the retransmission or FEC mechanisms, they are recommended to include support for both SSRC and Session multiplexing so that application developers can choose freely from the set of mechanisms without concerning themselves with which of the multiplexing choices a particular solution supports.

9. Proposal for Future Work

The above discussion and guidelines indicates that a small set of extension mechanisms could greatly improve the situation when it comes to using multiple streams independently of Session multiplexing or SSRC multiplexing. These extensions are:

Media Source Identification: A Media source identification that can be used to bind together media streams that are related to the same media source. A proposal [I-D.westerlund-avtext-rtcp-sdes-srcname] exist for a new SDES



item SRCNAME that also can be used with the a=ssrc SDP attribute to provide signalling layer binding information.

SSRC limitations within RTP sessions: By providing a signalling solution that allows the signalling peers to explicitly express both support and limitations on how many simultaneous media streams an end-point can handle within a given RTP Session. That ensures that usage of SSRC multiplexing occurs when supported and without overloading an end-point. This extension is proposed in [I-D.westerlund-avtcore-max-ssrc].

10. RTP Specification Clarifications

This section describes a number of clarifications to the RTP specifications that are likely necessary for aligned behavior when RTP sessions contain more SSRCs than one local and one remote.

10.1. RTCP Reporting from all SSRCs

When one have multiple SSRC in an RTP node, all these SSRC must send RTCP SR or RR as long as the SSRC exist. It is not sufficient that only one SSRC in the node sends report blocks on the incoming RTP streams. The reason for this is that a third party monitor may not necessarily be able to determine that all these SSRC are in fact co- located and originate from the same stack instance that gather report data.

10.2. RTCP Self-reporting

For any RTP node that sends more than one SSRC, there is the question if SSRC1 needs to report its reception of SSRC2 and vice versa. The reason that they in fact need to report on all other local streams as being received is report consistency. A third party monitor that considers the full matrix of media streams and all known SSRC reports on these media streams would detect a gap in the reports which could be a transport issue unless identified as in fact being sources from same node.

10.3. Combined RTCP Packets

When a node contains multiple SSRCs, it is questionable if an RTCP compound packet can only contain RTCP packets from a single SSRC or if multiple SSRCs can include their packets in a joint compound packet. The high level question is a matter for any receiver processing on what to expect. In addition to that question there is the issue of how to use the RTCP timer rules in these cases, as the existing rules are focused on determining when a single SSRC can



send.


This document makes no request of IANA.

Note to RFC Editor: this section may be removed on publication as an RFC.


There is discussion of the security implications of choosing SSRC vs Session multiplexing in Section 6.6.


The authors would like to thanks Harald Alvestrand for providing input into the discussion regarding multiple media types in a single RTP session.

14. References





[ALF] Clark, D. and D. Tennenhouse, "Architectural Considerations for a New Generation of Protocols", SIGCOMM Symposium on Communications Architectures and Protocols (Philadelphia, Pennsylvania), pp. 200--208, IEEE Computer Communications Review, Vol. 20(4), September 1990.

[I-D.alvestrand-rtp-sess-neutral] Alvestrand, H., "Why RTP Sessions Should Be Content Neutral", draft-alvestrand-rtp-sess-neutral-00 (work in progress), December 2011.



[I-D.holmberg-mmusic-sdp-bundle-negotiation] Holmberg, C. and H. Alvestrand, "Multiplexing Negotiation Using Session Description Protocol (SDP) Port Numbers", draft-holmberg-mmusic-sdp-bundle-negotiation-00 (work in progress), October 2011.

[I-D.ietf-avt-srtp-ekt] Wing, D., McGrew, D., and K. Fischer, "Encrypted Key Transport for Secure RTP", draft-ietf-avt-srtp-ekt-03 (work in progress), October 2011.

[I-D.ietf-avtext-multiple-clock-rates] Petit-Huguenin, M., "Support for multiple clock rates in an RTP session", draft-ietf-avtext-multiple-clock-rates-02 (work in progress), January 2012.

[I-D.ietf-payload-rtp-howto] Westerlund, M., "How to Write an RTP Payload Format", draft-ietf-payload-rtp-howto-01 (work in progress), July 2011.

[I-D.westerlund-avtcore-max-ssrc] Westerlund, M., Burman, B., and F. Jansson, "Multiple Synchronization sources (SSRC) in RTP Session Signaling", draft-westerlund-avtcore-max-ssrc (work in progress), October 2011.

[I-D.westerlund-avtcore-single-transport-multiplexing] Westerlund, M., "Multiple RTP Session on a Single Lower- Layer Transport", draft-westerlund-avtcore-transport-multiplexing (work in progress), October 2011.

[I-D.westerlund-avtext-rtcp-sdes-srcname] Westerlund, M., Burman, B., and P. Sandgren, "RTCP SDES Item SRCNAME to Label Individual Sources", draft-westerlund-avtext-rtcp-sdes-srcname (work in progress), October 2011.

[RFC2198] Perkins, C., Kouvelas, I., Hodson, O., Hardman, V., Handley, M., Bolot, J., Vega-Garcia, A., and S. Fosse- Parisis, "RTP Payload for Redundant Audio Data", RFC 2198, September 1997.

[RFC2205] Braden, B., Zhang, L., Berson, S., Herzog, S., and S. Jamin, "Resource ReSerVation Protocol (RSVP) -- Version 1 Functional Specification", RFC 2205, September 1997.



[RFC2326] Schulzrinne, H., Rao, A., and R. Lanphier, "Real Time Streaming Protocol (RTSP)", RFC 2326, April 1998.

[RFC2474] Nichols, K., Blake, S., Baker, F., and D. Black, "Definition of the Differentiated Services Field (DS Field) in the IPv4 and IPv6 Headers", RFC 2474, December 1998.

[RFC2974] Handley, M., Perkins, C., and E. Whelan, "Session Announcement Protocol", RFC 2974, October 2000.

[RFC3261] Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston, A., Peterson, J., Sparks, R., Handley, M., and E. Schooler, "SIP: Session Initiation Protocol", RFC 3261, June 2002.


[RFC3389] Zopf, R., "Real-time Transport Protocol (RTP) Payload for Comfort Noise (CN)", RFC 3389, September 2002.

[RFC3551] Schulzrinne, H. and S. Casner, "RTP Profile for Audio and Video Conferences with Minimal Control", STD 65, RFC 3551, July 2003.


[RFC3830] Arkko, J., Carrara, E., Lindholm, F., Naslund, M., and K. Norrman, "MIKEY: Multimedia Internet KEYing", RFC 3830, August 2004.

[RFC4103] Hellstrom, G. and P. Jones, "RTP Payload for Text Conversation", RFC 4103, June 2005.

[RFC4566] Handley, M., Jacobson, V., and C. Perkins, "SDP: Session Description Protocol", RFC 4566, July 2006.

[RFC4568] Andreasen, F., Baugher, M., and D. Wing, "Session Description Protocol (SDP) Security Descriptions for Media Streams", RFC 4568, July 2006.




[RFC4607] Holbrook, H. and B. Cain, "Source-Specific Multicast for IP", RFC 4607, August 2006.

[RFC5104] Wenger, S., Chandra, U., Westerlund, M., and B. Burman, "Codec Control Messages in the RTP Audio-Visual Profile with Feedback (AVPF)", RFC 5104, February 2008.

[RFC5117] Westerlund, M. and S. Wenger, "RTP Topologies", RFC 5117, January 2008.

[RFC5576] Lennox, J., Ott, J., and T. Schierl, "Source-Specific Media Attributes in the Session Description Protocol (SDP)", RFC 5576, June 2009.

[RFC5583] Schierl, T. and S. Wenger, "Signaling Media Decoding Dependency in the Session Description Protocol (SDP)", RFC 5583, July 2009.

[RFC5760] Ott, J., Chesterfield, J., and E. Schooler, "RTP Control Protocol (RTCP) Extensions for Single-Source Multicast Sessions with Unicast Feedback", RFC 5760, February 2010.

[RFC5761] Perkins, C. and M. Westerlund, "Multiplexing RTP Data and Control Packets on a Single Port", RFC 5761, April 2010.

[RFC5764] McGrew, D. and E. Rescorla, "Datagram Transport Layer Security (DTLS) Extension to Establish Keys for the Secure Real-time Transport Protocol (SRTP)", RFC 5764, May 2010.

[RFC5888] Camarillo, G. and H. Schulzrinne, "The Session Description Protocol (SDP) Grouping Framework", RFC 5888, June 2010.

[RFC6190] Wenger, S., Wang, Y., Schierl, T., and A. Eleftheriadis, "RTP Payload Format for Scalable Video Coding", RFC 6190, May 2011.

[RFC6285] Ver Steeg, B., Begen, A., Van Caenegem, T., and Z. Vax, "Unicast-Based Rapid Acquisition of Multicast RTP Sessions", RFC 6285, June 2011.

Appendix A. Dismissing Payload Type Multiplexing

This section documents a number of reasons why using the payload type as a multiplexing point for most things related to multiple streams is unsuitable. If one attempts to use Payload type multiplexing beyond it’s defined usage, that has well known negative effects on RTP. To use Payload type as the single discriminator for multiple



streams implies that all the different media streams are being sent with the same SSRC, thus using the same timestamp and sequence number space. This has many effects:

1. Putting restraint on RTP timestamp rate for the multiplexed media. For example, media streams that use different RTP timestamp rates cannot be combined, as the timestamp values need to be consistent across all multiplexed media frames. Thus streams are forced to use the same rate. When this is not possible, Payload Type multiplexing cannot be used.

2. Many RTP payload formats may fragment a media object over multiple packets, like parts of a video frame. These payload formats need to determine the order of the fragments to correctly decode them. Thus it is important to ensure that all fragments related to a frame or a similar media object are transmitted in sequence and without interruptions within the object. This can relatively simple be solved on the sender side by ensuring that the fragments of each media stream are sent in sequence.

3. Some media formats require uninterrupted sequence number space between media parts. These are media formats where any missing RTP sequence number will result in decoding failure or invoking of a repair mechanism within a single media context. The text/ T140 payload format [RFC4103] is an example of such a format. These formats will need a sequence numbering abstraction function between RTP and the individual media stream before being used with Payload Type multiplexing.

4. Sending multiple streams in the same sequence number space makes it impossible to determine which Payload Type and thus which stream a packet loss relates to.

5. If RTP Retransmission [RFC4588] is used and there is a loss, it is possible to ask for the missing packet(s) by SSRC and sequence number, not by Payload Type. If only some of the Payload Type multiplexed streams are of interest, there is no way of telling which missing packet(s) belong to the interesting stream(s) and all lost packets must be requested, wasting bandwidth.

6. The current RTCP feedback mechanisms are built around providing feedback on media streams based on stream ID (SSRC), packet (sequence numbers) and time interval (RTP Timestamps). There is almost never a field to indicate which Payload Type is reported, so sending feedback for a specific media stream is difficult without extending existing RTCP reporting.



7. The current RTCP media control messages [RFC5104] specification is oriented around controlling particular media flows, i.e. requests are done addressing a particular SSRC. Such mechanisms would need to be redefined to support Payload Type multiplexing.

8. The number of payload types are inherently limited. Accordingly, using Payload Type multiplexing limits the number of streams that can be multiplexed and does not scale. This limitation is exacerbated if one uses solutions like RTP and RTCP multiplexing [RFC5761] where a number of payload types are blocked due to the overlap between RTP and RTCP.

9. At times, there is a need to group multiplexed streams and this is currently possible for RTP Sessions and for SSRC, but there is no defined way to group Payload Types.

10. It is currently not possible to signal bandwidth requirements per media stream when using Payload Type Multiplexing.

11. Most existing SDP media level attributes cannot be applied on a per Payload Type level and would require re-definition in that context.

12. A legacy end-point that doesn’t understand the indication that different RTP payload types are different media streams may be slightly confused by the large amount of possibly overlapping or identically defined RTP Payload Types.


Magnus Westerlund Ericsson Farogatan 6 SE-164 80 Kista Sweden

Phone: +46 10 714 82 87 Email: [email protected]



Bo Burman Ericsson Farogatan 6 SE-164 80 Kista Sweden





Network Working Group M. WesterlundInternet-Draft EricssonIntended status: Standards Track C. PerkinsExpires: September 13, 2012 University of Glasgow March 12, 2012

Multiple RTP Sessions on a Single Lower-Layer Transport draft-westerlund-avtcore-transport-multiplexing-02

Abstract

This document specifies how multiple RTP sessions are to be multiplexed on the same lower-layer transport, e.g. a UDP flow. It discusses various requirements that have been raised and their feasibility, which results in a solution with a certain applicability. A solution is recommended and that solution is provided in more detail, including signalling and examples.

Status of this Memo





Copyright Notice


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

Westerlund & Perkins Expires September 13, 2012 [Page 1]

Internet-Draft Multiple RTP Session on Single Transport March 2012

the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.

Table of Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 2. Conventions . . . . . . . . . . . . . . . . . . . . . . . . . 4 2.1. Terminology . . . . . . . . . . . . . . . . . . . . . . . 4 2.2. Requirements Language . . . . . . . . . . . . . . . . . . 4 3. Requirements . . . . . . . . . . . . . . . . . . . . . . . . . 4 3.1. Support Use of Multiple RTP Sessions . . . . . . . . . . . 5 3.2. Same SSRC Value in Multiple RTP Sessions . . . . . . . . . 5 3.3. SRTP . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 3.4. Don’t Redefine Used Bits . . . . . . . . . . . . . . . . . 7 3.5. Firewall Friendly . . . . . . . . . . . . . . . . . . . . 7 3.6. Monitoring and Reporting . . . . . . . . . . . . . . . . . 7 3.7. Usable Also Over Multicast . . . . . . . . . . . . . . . . 7 3.8. Incremental Deployment . . . . . . . . . . . . . . . . . . 8 4. Possible Solutions . . . . . . . . . . . . . . . . . . . . . . 8 4.1. Header Extension . . . . . . . . . . . . . . . . . . . . . 8 4.2. Multiplexing Shim . . . . . . . . . . . . . . . . . . . . 9 4.3. Single Session . . . . . . . . . . . . . . . . . . . . . . 10 4.4. Use the SRTP MKI field . . . . . . . . . . . . . . . . . . 11 4.5. Use an Octet in the Padding . . . . . . . . . . . . . . . 12 4.6. Redefine the SSRC field . . . . . . . . . . . . . . . . . 12 5. Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . 13 5.1. Support of Multiple RTP Sessions Over Single Transport . . 13 5.2. Enable Same SSRC Value in Multiple RTP Sessions . . . . . 13 5.2.1. Avoid SSRC Translation in Gateways/Translation . . . . 13 5.2.2. Support Existing Extensions . . . . . . . . . . . . . 14 5.3. Ensure SRTP Functions . . . . . . . . . . . . . . . . . . 14 5.4. Don’t Redefine Used Bits . . . . . . . . . . . . . . . . . 15 5.5. Firewall Friendly . . . . . . . . . . . . . . . . . . . . 16 5.6. Monitoring and Reporting . . . . . . . . . . . . . . . . . 17 5.7. Usable over Multicast . . . . . . . . . . . . . . . . . . 18 5.8. Incremental Deployment . . . . . . . . . . . . . . . . . . 18 5.9. Summary and Conclusion . . . . . . . . . . . . . . . . . . 19 6. Specification . . . . . . . . . . . . . . . . . . . . . . . . 20 6.1. Shim Layer . . . . . . . . . . . . . . . . . . . . . . . . 21 6.2. Signalling . . . . . . . . . . . . . . . . . . . . . . . . 24 6.3. SRTP Key Management . . . . . . . . . . . . . . . . . . . 25 6.3.1. Security Description . . . . . . . . . . . . . . . . . 25 6.3.2. DTLS-SRTP . . . . . . . . . . . . . . . . . . . . . . 26 6.3.3. MIKEY . . . . . . . . . . . . . . . . . . . . . . . . 26 6.4. Examples . . . . . . . . . . . . . . . . . . . . . . . . . 26 6.4.1. RTP Packet with Transport Header . . . . . . . . . . . 26 6.4.2. SDP Offer/Answer example . . . . . . . . . . . . . . . 27



7. Open Issues . . . . . . . . . . . . . . . . . . . . . . . . . 29 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 30 9. Security Considerations . . . . . . . . . . . . . . . . . . . 30 10. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 30 11. References . . . . . . . . . . . . . . . . . . . . . . . . . . 31 11.1. Normative References . . . . . . . . . . . . . . . . . . . 31 11.2. Informational References . . . . . . . . . . . . . . . . . 31 Authors’ Addresses . . . . . . . . . . . . . . . . . . . . . . . . 32



1. Introduction

There has been renewed interest for having a solution that allows multiple RTP sessions [RFC3550] to use a single lower layer transport, such as a bi-directional UDP flow. The main reason is the cost of doing NAT/FW traversal for each individual flow. ICE and other NAT/FW traversal solutions are clearly capable of attempting to open multiple flows. However, there is both increased risk for failure and an increased cost in the creation of multiple flows. The increased cost comes as slightly higher delay in establishing the traversal, and the amount of consumed NAT/FW resources. The latter might be an increasing problem in the IPv4 to IPv6 transition period.

This document draws up some requirements for consideration on how to transport multiple RTP sessions over a single lower-layer transport. These requirements will have to be weighted as the combined set of requirements result in that no known solution exist that can fulfill them completely.

A number of possible solutions are then considered and discussed with respect to their properties. Based on that, the authors recommends a shim layer variant as single solution, which is described in more detail including signalling solution and examples.

2. Conventions

2.1. Terminology

Some terminology used in this document.

Multiplexing: Unless specifically noted, all mentioning of multiplexing in this document refer to the multiplexing of multiple RTP Sessions on the same lower layer transport. It is important to make this distinction as RTP does contain a number of multiplexing points for various purposes, such as media formats (Payload Type), media sources (SSRC), and RTP sessions.



3. Requirements

This section lists and discusses a number of potential requirements.



However, it is not difficult to realize that it is in fact possible to put requirements that makes the set of feasible solutions an empty set. It is thus necessary to consider which requirements that are essential to fulfill and which can be compromised on to arrive at a solution.

3.1. Support Use of Multiple RTP Sessions

This may at first glance appear to be an obvious requirement. Although the authors are convinced it is a mandatory requirement for a solution, it warrants some discussion around the implications of not having multiple RTP sessions and instead use a single RTP session.

The usage of multiple RTP sessions allow separation of media streams that have different usages or purposes in an RTP based application, for example to separate the video of a presenter or most important current talker from those of the listeners that not all end-points receiver. Also separation for different processing based on media types such as audio and video in end-points and central nodes. Thus providing the node with the knowledge that any SSRC within the session is supposed to be processed in a similar or same way.

For simpler cases, where the streams within each media type need the same processing, it is clearly possible to find other multiplex solutions, for example based on the Payload Type and the differences in encoding that the payload type allows to describe. This may anyhow be insufficient when you get into more advanced usages where you have multiple sources of the same media type, but for different usages or as alternatives. For example when you have one set of video sources that shows session participants and another set of video sources that shares an application or slides, you likely want to separate those streams for various reasons such as control, prioritization, QoS, methods for robustification, etc. In those cases, using the RTP session for separation of properties is a powerful tool. A tool with properties that need to be preserved when providing a solution for how to use only a single lower-layer transport.

For more discussion of the usage of RTP sessions verses other multiplexing we recommend RTP Multiplexing Architecture [I-D.westerlund-avtcore-multiplex-architecture].

3.2. Same SSRC Value in Multiple RTP Sessions

Two different RTP sessions being multiplexed on the same lower layer transport need to be able to use the same SSRC value. This is a strong requirement, for two reasons:



1. To avoid mandating SSRC assignment rules that are coordinated between the sessions. If the RTP sessions multiplexed together must have unique SSRC values, then additional code that works between RTP Sessions is needed in the implementations. Thus raising the bar for implementing this solution. In addition, if one gateways between parts of a system using this multiplexing and parts that aren’t multiplexing, the part that isn’t multiplexing must also fulfill the requirements on how SSRC is assigned or force the gateway to translate SSRCs. Translating SSRC is actually hard as it requires one to understand the semantics of all current and future RTP and RTCP extensions. Otherwise a barrier for deploying new extensions is created.

2. There are some few RTP extensions that currently rely on being able to use the same SSRC in different RTP sessions:

* XOR FEC (RFC5109)

* RTP Retransmission in session mode (RFC4588)

* Certain Layered Coding

3.3. SRTP

SRTP [RFC3711] is one of the most commonly used security solutions for RTP. In addition, it is the only one recommended by IETF that is integrated into RTP. This integration has several aspects that needs to be considered when designing a solution for multiplexing RTP sessions on the same lower layer transport.

Determining Crypto Context: SRTP first of all needs to know which session context a received or to-be-sent packet relates to. It also normally relies on the lower layer transport to identify the session. It uses the MKI, if present, to determine which key set is to be used. Then the SSRC and sequence number are used by most crypto suites, including the most common use of AES Counter Mode, to actually generate the correct cipher stream.

Unencrypted Headers: SRTP has chosen to leave the RTP headers and the first two 32-bit words of the first RTCP header unencrypted, to allow for both header compression and monitoring to work also in the presence of encryption. As these fields are in clear text they are used in most crypto suites for SRTP to determine how to protect or recover the plain text.

It is here important to contrast SRTP against a set of other possible protection mechanisms. DTLS, TLS, and IPsec are all protecting and encapsulating the entire RTP and RTCP packets. They don’t perform



any partial operations on the RTP and RTCP packets. Any change that is considered to be part of the RTP and RTCP packet is transparent to them, but possibly not to SRTP. Thus the impact on SRTP operations must be considered when defining a mechanism.

3.4. Don’t Redefine Used Bits

As the core of RTP is in use in many systems and has a really large deployment story and numerous implementations, changing any of the field definitions is highly problematic. First of all, the implementations need to change to support this new semantics. Secondly, you get a large transition issue when you have some session participants that support the new semantics and some that don’t. Combing the two behaviors in the same session can force the deployment of costly and less than perfect translation devices.

3.5. Firewall Friendly

It is desirable that current firewalls will accept the solutions as normal RTP packets. However, in the authors’ opinion we can’t let the firewall stifle invention and evolution of the protocol. It is also necessary to be aware that a change that will make most deep inspecting firewall consider the packet as not valid RTP/RTCP will have more difficult deployment story.

3.6. Monitoring and Reporting

It is desirable that a third party monitor can still operate on the multiplexed RTP Sessions. It is however likely that they will require an update to correctly monitor and report on multiplexed RTP Sessions.

Another type of function to consider is packet sniffers and their selector filters. These may be impacted by a change of the fields. An observation is that many such systems are usually quite rapidly updated to consider new types of standardized or simply common packet formats.

3.7. Usable Also Over Multicast

It is desirable that a solution should be possible to use also when RTP and RTCP packets are sent over multicast, both Any Source Multicast (ASM) and Single Source Multicast (SSM). The reason for this requirement is to allow a system using RTP to use the same configuration regardless of the transport being done over unicast or multicast. In addition, multicast can’t be claimed to have an issue with using multiple ports, as each multicast group has a complete port space scoped by address.



3.8. Incremental Deployment

A good solution has the property that in topologies that contains RTP mixers or Translators, a single session participant can enable multiplexing without having any impact on any other session participants. Thus a node should be able to take a multiplexed packet and then easily send it out with minimal or no modification on another leg of the session, where each RTP session is transported over its own lower-layer transport. It should also be as easy to do the reverse forwarding operation.

4. Possible Solutions

This section looks at a few possible solutions and discusses their feasibility.

4.1. Header Extension

One proposal is to define an RTP header extension [RFC5285] that explicitly enumerates the session identifier in each packet. This proposal has some merits regarding RTP, since it uses an existing extension mechanism; it explicitly enumerates the session allowing for third parties to associate the packet to a given RTP session; and it works with SRTP as currently defined since a header extension is by default not encrypted, and is thus readable by the receiving stack without needing to guess which session it belongs to and attempt to decrypt it. This approach does, however, conflict with the requirement from [RFC5285] that "header extensions using this specification MUST only be used for data that can be safely ignored by the recipient", since correct processing of the received packet depends on using the header extension to demultiplex it to the correct RTP session.

Using a header extension also result in the session ID is in the integrity protected part of the packet. Thus a translator between multiplexed and non-multiplexed has the options:

1. to be part of the security context to verify the field

2. to be part of the security context to verify the field and remove it before forwarding the packet

3. to be outside of the security context and leave the header extension in the packet. However, that requires successful negotiation of the header extension, but not of the functionality, with the receiving end-points.



The biggest existing hurdle for this solution is that there exist no header extension field in the RTCP packets. This requires defining a solution for RTCP that allows carrying the explicit indicator, preferably in a position that isn’t encrypted by SRTCP. However, the current SRTCP definition does not offer such a position in the packet.

Modifying the RR or SR packets is possible using profile specific extensions. However, that has issues when it comes to deployability and in addition any information placed there would end up in the encrypted part.

Another alternative could be to define another RTCP packet type that only contains the common header, using the 5 bits in the first byte of the common header to carry a session id. That would allow SRTCP to work correctly as long it accepts this new packet type being the first in the packet. Allowing a non-SR/RR packet as the first packet in a compound RTCP packet is also needed if an implementation is to support Reduced Size RTCP packets [RFC5506]. The remaining downside with this is that all stack implementations supporting multiplexing would need to modify its RTCP compound packet rules to include this packet type first. Thus a translator box between supporting nodes and non-supporting nodes needs to be in the crypto context.

This solution’s per packet overhead is expected to be 64-bits for RTCP. For RTP it is 64-bits if no header extension was otherwise used, and an additional 16 bits (short header), or 24 bits plus (if needed) padding to next 32-bits boundary if other header extensions are used.

4.2. Multiplexing Shim

This proposal is to prefix or postfix all RTP and RTCP packets with a session ID field. This field would be outside of the normal RTP and RTCP packets, thus having no impact on the RTP and RTCP packets and their processing. An additional step of demultiplexing processing would be added prior to RTP stack processing to determine in which RTP session context the packet shall be included. This has also no impact on SRTP/SRTCP as the shim layer would be outside of its protection context. The shim layer’s session ID is however implicitly integrity protected as any error in the field will result in the packet being placed in the wrong or non-existing context, thus resulting in a integrity failure if processed by SRTP/SRTCP.

This proposal is quite simple to implement in any gateway or translating device that goes from a multiplexed to a non-multiplexed domain or vice versa, as only an additional field needs to be added to or removed from the packet.



The main downside of this proposal is that it is very likely to trigger a firewall response from any deep packet inspection device. If the field is prefixed, the RTP fields are not matching the heuristics field (unless the shim is designed to look like an RTP header, in which case the payload length is unlikely to match the expected value) and thus are likely preventing classification of the packet as an RTP packet. If it is postfixed, it is likely classified as an RTP packet but may not correctly validate if the content validation is such that the payload length is expected to match certain values. It is expected that a postfixed shim will be less problematic than a prefixed shim in this regard, but we are lacking hard data on this.

This solution’s per packet overhead is 1 byte.

4.3. Single Session

Given the difficulty of multiplexing several RTP sessions onto a single lower-layer transport, it’s tempting to send multiple media streams in a single RTP session. Doing this avoids the need to demultiplex several sessions on a single transport, but at the cost of losing the RTP session as a separator for different type of streams. Lacking different RTP sessions to demultiplex incoming packets, a receiver will have to dig deeper into the packet before determining what to do with it. Care must be taken in that inspection. For example, you must be careful to ensure that each real media source uses its own SSRC in the session and that this SSRC doesn’t change media type.

The loss of the RTP session as a separator for different usages or purpose would be an minor issue if the only difference between the RTP sessions is the media type. In this case, the application could use the Payload Type field to identify the media type. The loss of the RTP Session functionality is however severe, if the application uses the RTP Session for separating different treatments, contexts etc. Then you would need additional signalling to bind the different sources to groups which can help make the necessary distinctions.

However, the loss of the RTP session as separator is not the only issue with this approach. The RTP Multiplexing Architecture [I-D.westerlund-avtcore-multiplex-architecture] discusses a number of issues in Section 6.7. These include RTCP bandwidth differences, limitations in the number of payload types, media aware RTP mixers and interactions with Legacy end-points.

Additional attention should be place on this important aspect. In multi-party situations using central nodes there exist some difficulties in having a legacy implementation using multiple RTP



sessions interworking with an end-point having only a single RTP session across the central node. The main reason is the fact that the one using single session with multiple media types has only one SSRC space, while the other end-points have multiple spaces. Thus translation may have to occur because there is several RTP sessions using the same SSRC value. This has both limitations, processing overhead and the possibility of becoming an deployment obstacle for new RTP/RTCP extensions.

This approach has been proposed in the RTCWeb context in [I-D.lennox-rtcweb-rtp-media-type-mux] and [I-D.ietf-mmusic-sdp-bundle-negotiation]. These drafts describe how to signal multiple media streams multiplexed into a single RTP session, and address some of the issues raised here and in Section 6.7 of the RTP Multiplexing Architecture [I-D.westerlund-avtcore-multiplex-architecture] draft.

This method has several limitations that limits its usage as solution in providing multiple RTP sessions on the same lower layer transport. However, we acknowledge that there are some uses for which this method may be sufficient and which can accept the methods limitations and downsides. The RTCWEB WG has a working assumption to support this method. For more details of this method, see the relevant drafts under development. We do include this method in the comparison to provide a more complete picture of the pro and cons of this method.

This solution has no per packet overhead. The signalling overhead will be a different question.

4.4. Use the SRTP MKI field

This proposal is to overload the MKI SRTP/SRTCP identifier to not only identify a particular crypto context, but also identify the actual RTP Session. This clearly is a miss use of the MKI field, however it appears to be with little negative implications. SRTP already supports handling of multiple crypto contexts.

The two major downsides with this proposal is first the fact that it requires using SRTP/SRTCP to multiplex multiple sessions on a single lower layer transport. The second issue is that the session ID parameter needs to be put into the various key-management schemes and to make them understand that the reason to establish multiple crypto contexts is because they are connected to various RTP Sessions. Considering that SRTP have at least 3 used keying mechanisms, DTLS- SRTP [RFC5764], Security Descriptions [RFC4568], and MIKEY [RFC3830], this is not an insignificant amount of work.



This solution has 32-bit per packet overhead, but only if the MKI was not already used.

4.5. Use an Octet in the Padding

The basics of this proposal is to have the RTP packet and the last (required by RFC3550) RTCP packet in a compound to include padding, at least 2 bytes. One byte for the padding count (last byte) and one byte just before the padding count containing the session ID.

This proposal uses bytes to carry the session ID that have no defined value and is intended to be ignored by the receiver. From that perspective it only causes packet expansion that is supported and handled by all existing equipment. If an implementation fails to understand that it is required to interpret this padding byte to learn the session ID, it will see a mostly coherent RTP session except where SSRCs overlap or where the payload types overlap. However, reporting on the individual sources or forwarding the RTCP RR are not completely without merit.

There is one downside of this proposal and that has to do with SRTP. To be able to determine the crypto context, it is necessary to access to the encrypted payload of the packet. Thus, the only mechanism available for a receiver to solve this issue is to try the existing crypto contexts for any session on the same lower layer transport and then use the one where the packet decrypts and verifies correctly. Thus for transport flows with many crypto contexts, an attacker could simply generate packets that don’t validate to force the receiver to try all crypto contexts they have rather than immediately discard it as not matching a context. A receiver can mitigate this somewhat by using heuristics based on the RTP header fields to determine which context applies for a received packet, but this is not a complete solution.

This solution has a 16-bit per packet overhead.

4.6. Redefine the SSRC field

The Rosenberg et. al. Internet draft "Multiplexing of Real-Time Transport Protocol (RTP) Traffic for Browser based Real-Time Communications (RTC)" [I-D.rosenberg-rtcweb-rtpmux] proposed to redefine the SSRC field. This has the advantage of no packet expansion. It also looks like regular RTP. However, it has a number of implications. First of all it prevents any RTP functionality that require the same SSRC in multiple RTP sessions.

Secondly its interoperability with end-point using multiple RTP sessions are problematic. Such interoperability will requires an



SSRC translator function in the gatewaying node to ensure that the SSRCs fulfill the semantic rules of the different domains. That translator is actually far from easy as it needs to understand the semantics of all RTP and RTCP extensions that include SSRC/CSRC. This as it is necessary to know when a particular matching 32-bit pattern is an SSRC field and when the field is just a combination of other fields that create the same matching 32-bit pattern. Thus there is a possibility that such a translator becomes a obstacle in deploying future RTP/RTCP extensions. In addition the translator actually have significant overhead when SRTP are in use. This as a verification that the packet is authentic, decryption, SSRC translation, encryption and finally generation of authentication tags are required. In addition the translator must be part of the security context.

This solution has no per packet overhead.

5. Comparison

This section compares the above potential solutions with the requirements. Motivations are provided in addition to a high level metric of successfully, partially and failing to meet requirement. In the end a summary table (Figure 1) of the high level value are provided.

5.1. Support of Multiple RTP Sessions Over Single Transport

This one is easy to determine. Only the single session proposal fails this requirement as it is not at all designed to meet it. The rest fully support this requirement. The main question around this requirement is how important it is to have as discussed in Section 3.1.

5.2. Enable Same SSRC Value in Multiple RTP Sessions

Based on the discussion in Section 3.2 two sub-requirements have been derived.

5.2.1. Avoid SSRC Translation in Gateways/Translation

This sub-requirement is derived based on the desire to avoid having gateways or translators perform full SSRC translation to minimize complexity, avoid the requirement to have gateways in security context, and as a hinder to long-term evolution. Two of the proposals have issues with this, due to their lack of support for multiple 32-bit SSRC spaces and lacking possibility to have the same SSRC value in multiple RTP sessions. The proposals that have these



properties and thus are marked as failing are the Single Session and Redefine the SSRC field. The other proposals are all succcesful in meeting this requirement.

5.2.2. Support Existing Extensions

The second sub-requirement is how well the proposals support using the existing RTP mechanisms. Here both Single Session and Redefine the SSRC field will have clear issues as they cannot support the same full 32-bit SSRC value in two different RTP sessions. This is clearly an issue for the XOR based FEC. RTP retransmission and scalable encoding are minor issues as there exist alternatives to those mechanisms that works with the structure of these two proposals. Thus we give them a fail. The Header Extension gets a partial due to unclear interaction between putting in an header extension and these mechanisms.

5.3. Ensure SRTP Functions

This requirement is about ensuring both secure and efficient usage of SRTP. The Octet in Padding field proposal gets a fail as the receiving end-point cannot determine the intended RTP session prior to de-encryption of the padding field. Thus a catch-22 arises which can only be resolved by trying all session contexts and see what decrypts. This causes a security vulnerability as an attacker can inject a packet which does not meet any of the session contexts. The receiver will then attempt decryption and authentication of it using all its session contexts, increasing the amount of wasted resources by a factor equal to the number of multiplexed sessions. Thus this proposal gets a fail.

The proposal of Overloading the SRTP MKI field as session identifier gets a partial due to the fact that it cannot use SRTP’s key- management mechanism out of the box. It forces the key-management mechanism and the SRTP implementations to maintain the MKI-to-RTP session bindings to maintain secure and correct function.

The Redefine the SSRC field gets a partial due to its need to modify the key-management mechanisms to correctly identify the partial SSRC space the parameters applies to. Similarly, the SRTP implementation also needs to be updated to correctly support this security context differentiation.

The header extension based solution gets a less severe partial than Redefine the SSRC and the MKI. It will however have an issue when being gatewayed to a domain that does not multiplex multiple RTP sessions over the same transport. Then the gateway will require to be in the security context to be able to add or remove the header



extension as it is in the part of the packet that is integrity protected by SRTP.

The remaining two proposals do not affect SRTP mechanisms and thus successfully meet this requirement.

5.4. Don’t Redefine Used Bits

This requirement is all about RTP and RTCP header fields having a given definition should not be changed as it can cause interoperability problems between modified and non-modified implementations. This becomes especially problematic in RTP sessions used for multi-party sessions.

Redefine the SSRC field gets a big fail on this as it redefines the SSRC field, a core field in RTP. It has been identified that such a change will have issues since if it gets connected to a non-modified end-point that randomly assigns the SSRC, as supposed by RFC 3550, those SSRCs will be distributed over different RTP sessions at the modified end-point. Also other functions using the SSRC field, not understanding the additional semantics of the SSRC field, is likely to have issues.

Using the SRTP MKI field to identify a session is overloading that field with double semantics. This likely has minimal negative impact in RTP since it should be possible to have the SRTP stack use the MKI field to both look up the security context and which output RTP session the processed packet belongs to. However, this redefinition clearly creates issues with the key-management scheme. That will have to be modified to handle both this change and deal with the interoperability issues when negotiating its usage. This gets a full fail due to that it makes the problem someone else’s, namely the RTP implementors.

Defining an Octet in the Padding field redefines a field, whose definition is to have zero value and is expected to be ignored by the receiver according to the original semantics. Thus this is one of the more benign modifications one can do, however this can still cause issues in implementations that unnecessarily check the field values, or in firewalls. This is judged to be partially meeting the requirement.

The Header Extension proposal does in fact not redefine any currently used bits in RTP. The header extension would be a correctly identified extension with its own definition. However, it does redefine a rule on what header extensions are for. The RTCP solution however would have more severe impact as it would need to redefine the standard meaning of an RTCP packet header in addition to the



default compound packet rules. Due to these issues the proposal fails to meet this requirement.

The multiplexing shim and the single session both successfully meet this requirement.

5.5. Firewall Friendly

This requirement is clearly difficult to judge as firewall implementations are highly different in both implementation, scope of what it investigates in packets, and set policies. A reasonable goal is to minimize the likeliness that rules and policies intended to let RTP media streams pass, will also let these streams through when multiplexing RTP sessions over a single transport. The below analysis shows that no solution is truly firewall friendly and all are judged as being partially meeting this goal. However, the reason why it is believed that a firewall might react to the streams are quite different.

The Single Session and Redefine the SSRC field are likely the least suspect solutions from a firewall perspective. However, as their transport flows contain multiple SSRCs with payloads that indicate likely multiple different media types they are still likely to make a picky firewall block the transport. This is especially true for firewalls that take signalling messages into account where it will expect a particular media type in a given context. A non upgraded firewall might in fact produce two different contexts with overlapping transport parameters where both rules will receive media streams of the other media type that are outside of the allowed rule. However, to be clear if these proposals doesn’t get through, none of the other will either as they all will have this behavior.

The header extension proposal is potentially problematic for two reasons. The first reason, which also other proposals has, is related to that the same SSRC value can exist in two RTP sessions over the same underlying flow. Anyone tracking the sequence number and timestamp will react badly as the second media stream with the same SSRC causes constant jumps back and forth in these fields compared to the first stream, if packets are transmitted simultaneously for both SSRCs. This issue can likely only be solved by having the firewalls that like to track flows to also use the session identifier to create context. This is possible as the header extension will be in the clear and in the front. The second issue is that the header extension itself may get the firewall to react. Especially very picky ones that expect packets with certain media types to have certain packet lengths. They are not compatible with a header extension.



The Multiplexing Shim shares the issue with multiple flows for the same SSRC. Firewalls and deep packet inspection cause the shim placement to be in question. If it is a pre-fixed shim, it prevents the packet from looking like regular IP/UDP/RTP packets and be correctly classified in firewalls and DPI engines. However, if one puts it last, it is unlikely that any firewall or DPI ever will be able to take the session context into account as it is at the end of the packet. This as many line rate processing devices only take a certain amount of the the headers into account.

The SRTP MKI field is likely the solution that has least firewall and DPI issues, after the single RTP session. There is no additional suspect field. The only difference from a single RTP session in the transport flow is the fact that multiple MKI are guaranteed to be used. However, that may occur also in a single RTP session usage. Thus the only issues are the one shared with single session and the one that several RTP media streams may use the same SSRC.

The octet in the padding field has, in addition to the issues the SRTP MKI field has, the single issue that it redefines something that is supposed to be zero into a value. Thus potentially causing a deeply inspecting firewall to clamp the flow in fear of covert channel or non-compliance.

5.6. Monitoring and Reporting

The monitoring and reporting requirement considers several aspects. How useful monitoring can one get from an existing legacy monitor, and secondary any issues in upgrading them to handle the selected solution. Thirdly, packet selector filters and packet sniffers concerns are considered.

In general one can expect the proposals that have only a single SSRC space to work better with legacy. Thus both Single Session and Redefine SSRC space can gather and report data on media flows most likely. The only potential issue is that due to the different media types and clock rates, some failure may occur. In particular a third party monitor may be targeted to a specific media type, like monitoring VoIP. That monitor will have problems processing any video packets correctly and generate the VoIP specific metrics for any video sending SSRC. In general, no legacy solution for monitoring will be able to correctly create the sub-contexts that each RTP session has in the solutions, without update to handle the new semantics. Also when it comes to the packet filtering and selector filters, fine grained control can only be accomplished implementing the new semantics. Therefore only the Single Session meets this requirement fully.



Redefine the SSRC field is close to fully meeting the requirement, however due to that there exist a session structure that is hidden to anyone that is not upgraded to understand the semantics, this only gets a partial.

The other proposals all can have multiple RTP sessions using the same SSRC. This will create significant issues for any legacy third party monitor. Only an updated monitor, or for that matter packet selector, can pick out the individual media streams and their associated RTCP traffic. Thus all these proposals gets a failure to meet the requirement.

5.7. Usable over Multicast

As discussed earlier the goal with having the option usable also over multicast is to remove the need to produce different media streams for transport over unicast and multicast. All of the proposals successfully meet the requirement.

5.8. Incremental Deployment

The possibility to deploy the usage of the multiplexing of multiple RTP sessions over a single transport, especially in the context of multi-party sessions, is a great benefit for any of the proposals. Thus not all end-point implementations needs to be upgraded before one start enabling it in the central node and any signalling.

Considering a centralized multi-party application where some participants are using multiple transport flows and you want to enable one particular participant to use the single transport to the central node, one criteria stands out. The possibility to have one RTP session per transport in one leg, and in the next multiplex them together with minimal complexity and packet changes. Here there are significant differences.

The Multiplexing Shim has the least overhead for this. As the central node or gateway between deployments only needs to either add or remove the shim identifier and then forward the packet over the corresponding transport, either a joint one on the single transport side, or over the individual one on the multiple transport side.

The SRTP MKI field proposal is almost as good, as the only main difference is the need to coordinate the used MKIs on the non- multiplexed legs so that there is no overlap between the RTP sessions. And if there is, the MKI can be translated in gateway as SRTP has no integrity protection over the MKI. Thus both multiplexing shim and SRTP MKI field does successfully meet this requirement.



The Header Extension supports multiple full 32-bit SSRC spaces and can thus handle all the RTP sessions without need for any SSRC translation, however this proposal does run into the problem that the gateway needs to be in the security context to be able to add or remove the header extension when SRTP is used. In addition to the security implications of that, there is a complexity overhead due to the need to redo the authentication tags on all RTP/RTCP packets. Thus it gets a partial.

The Octet in the Padding field share issues with the header extension but have even higher complexities for this. The reason is that the padding field is also encrypted. Thus to add or remove it (although removing it may be unncessary) forces the end-point to encrypt at least that byte also, and for ciphers that are not stream-ciphers, the whole packet needs to be re-encrypted. Thus this proposal gets a very weak partially meeting the requirement.

The Single Session and Redefine the SSRC field do not allow several vanilla RTP sessions to be connected to these proposals. The reason is the single 32-bit SSRC space they have. Single Session only has one session and the Redefine the SSRC fields uses some of the bits as session identifier. This forces the gateway to translate the SSRC whenever it does not fulfill the rules or semantics of the multiplexed side. For Redefine SSRC field this becomes almost constant as the session identifier part of the SSRC must be the same over all SSRCs from the same session. For Single Session it may only be needed when there otherwise would be an SSRC collision between the sessions. This further assumes that the non-multiplexed side would never use any of the RTP mechanisms that require the same SSRC in multiple RTP sessions, as they cannot be gatewayed at all. When translating an SSRC there is first of all an overhead, with SRTP that includes a complete authenticate, decrypte, encrypt and create a new authentication tag cycle. In addition, the SSRC translation could potentially be a deployment obstacle for new RTP/RTCP extensions required to be understood by the translator to be correctly translated. Therefore these two proposals gets a fail to meet the requirements.

5.9. Summary and Conclusion

This section contains a summary table of the high level outcome against the different requirements.

A table mapping the requirements against the ID numbers used in the table is the following:



1: Support multiple RTP sessions over one transport flow

2: Enable same SSRC value in multiple RTP sessions

2.1: Avoid SSRC translation in gateways/translators

2.2: Support existing extensions

3: Ensure SRTP functions

4: Don’t Redefine used bits

5: Firewall Friendly

6: Monitoring and Reporting should still function

7: Usable over Multicast

8: Incremental deployment

OH: Overhead in Bytes. + means variable

---------------+---+---+---+---+---+---+---+---+---+---- Solution | 1 |2.1|2.2| 3 | 4 | 5 | 6 | 7 | 8 | OH ---------------+---+---+---+---+---+---+---+---+---+---- Header Ext. | S | S | P | P | F | P | F | S | P | 8+ Multiplex Shim | S | S | S | S | S | P | F | S | S | 1 Single Session | F | F | F | S | S | P | S | S | F | 0 SRTP MKI Field | S | S | S | P | F | P | F | S | S | 4 Padding Field | S | S | S | F | P | P | F | S | P | 2 Redefine SSRC | S | F | F | P | F | P | P | S | S | 0 ---------------+---+---+---+---+---+---+---+---+---+----

Figure 1: Summary Table of Evaluation (Successfully (S), Partially (P) or Fails (F) to meet requirement)

Considering these options, the authors would recommend that AVTCORE standardize a solution based on a postfixed multiplexing field, i.e. a shim approach combined with the appropriate signalling as described in Section 4.2.

6. Specification

This section contains the specification of the solution based on a SHIM, with the explicit session identifier at the end of the encapsulated payload.



6.1. Shim Layer

This solution is based on a shim layer that is inserted in the stack between the regular RTP and RTCP packets and the transport layer being used by the RTP sessions. Thus the layering looks like the following:

+---------------------+ | RTP / RTCP Packet | +---------------------+ | Session ID Layer | +---------------------+ | Transport layer | +---------------------+

Stack View with Session ID SHIM

The above stack is in fact a layered one as it does allow multiple RTP Sessions to be multiplexed on top of the Session ID shim layer. This enables the example presented in Figure 2 where four sessions, S1-S4 is sent over the same Transport layer and where the Session ID layer will combine and encapsulate them with the session ID on transmission and separate and decapsulate them on reception.

+-------------------+ | S1 | S2 | S3 | S4 | +-------------------+ | Session ID Layer | +-------------------+ | Transport layer | +-------------------+

Figure 2: Multiple RTP Session On Top of Session ID Layer

The Session ID layer encapsulates one RTP or RTCP packet from a given RTP session and postfixes a one byte Session ID (SID) field to the packet. Each RTP session being multiplexed on top of a given transport layer is assigned either a single or a pair of unique SID in the range 0-255. The reason for assigning a pair of SIDs to a given RTP session are for RTP Sessions that doesn’t support "Multiplexing RTP Data and Control Packets on a Single Port" [RFC5761] to still be able to use a single 5-tuple. The reasons for supporting this extra functionality is that RTP and RTCP multiplexing based on the payload type/packet type fields enforces certain restrictions on the RTP sessions. These restrictions may not be acceptable. As this solution does not have these restrictions, performing RTP and RTCP multiplexing in this way has benefits.



Each Session ID value space is scoped by the underlying transport protocol. Common transport protocols like UDP, DCCP, TCP, and SCTP can all be scoped by one or more 5-tuple (Transport protocol, source address and port, destination address and port). The case of multiple 5-tuples occur in the case of multi-unicast topologies, also called meshed multiparty RTP sessions or in case any application would need more than 128 RTP sessions.

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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<+ |V=2|P|X| CC |M| PT | sequence number | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | timestamp | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | synchronization source (SSRC) identifier | | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ | | contributing source (CSRC) identifiers | | | .... | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | RTP extension (OPTIONAL) | | +>+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | | payload ... | | | | +-------------------------------+ | | | | RTP padding | RTP pad count | | +>+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<+ | ˜ SRTP MKI (OPTIONAL) ˜ | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | : authentication tag (RECOMMENDED) : | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | | Session ID | | | +---------------+ | +- Encrypted Portion* Authenticated Portion ---+

Figure 3: SRTP Packet encapsulated by Session ID Layer



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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<+ |V=2|P| RC | PT=SR or RR | length | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | SSRC of sender | | +>+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ | | ˜ sender info ˜ | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | ˜ report block 1 ˜ | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | ˜ report block 2 ˜ | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | ˜ ... ˜ | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | |V=2|P| SC | PT=SDES=202 | length | | | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ | | | SSRC/CSRC_1 | | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | ˜ SDES items ˜ | | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ | | ˜ ... ˜ | +>+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ | | |E| SRTCP index | | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<+ | ˜ SRTCP MKI (OPTIONAL) ˜ | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | : authentication tag : | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | | Session ID | | | +---------------+ | +-- Encrypted Portion Authenticated Portion -----+

Figure 4: SRTCP packet encapsulated by Session ID layer

The processing in a receiver when the Session ID layer is present will be to

1. Pick up the packet from the lower layer transport

2. Inspect the SID field value

3. Strip the SID field from the packet

4. Forward it to the (S)RTP Session context identified by the SID value



6.2. Signalling

The use of the Session ID layer needs to be explicitly agreed on between the communicating parties. Each RTP Session the application uses must in addition to the regular configuration such as payload types, RTCP extension etc, have both the underlying 5-tuple (source address and port, destination address and port, and transport protocol) and the Session ID used for the particular RTP session. The signalling requirement is to assign unique Session ID values to all RTP Sessions being sent over the same 5-tuple. The same Session ID shall be used for an RTP session independently of the traffic direction. Note that nothing prevents a multi-media application from using multiple 5-tuples if desired for some reason, in which case each 5-tuple has its own session ID value space.

This section defines how to negotiate the use of the Session ID layer, using the Session Description Protocol (SDP) Offer/Answer mechanism [RFC3264]. A new media-level SDP attribute, ’session-mux-id’, is defined, in order to be used with the media BUNDLE mechanism defined in [I-D.ietf-mmusic-sdp-bundle-negotiation]. The attribute allows each media description ("m=" line) associated with a ’BUNDLE’ group to form separate RTP sessions.

The ’session-mux-id’ attribute is included for a media description, in order to indicate the Session ID for that particular media description. Every media description that shares a common attribute value is assumed to be part of a single RTP session. An SDP Offerer MUST include the ’session-mux-id’ attribute for every media description associated with a ’BUNDLE’ group. If the SDP Answer does not contain ’session-mux-id’ attributes, the SDP Offerer MUST NOT assume that separate RTP sessions will be used. If the SDP Answer still describes a ’BUNDLE’ group, the procedures in [I-D.ietf-mmusic-sdp-bundle-negotiation] apply.

An SDP Answerer MUST NOT include the ’session-mux-id’ attribute in an SDP Answer, unless included in the SDP Offer.

The attribute has the following ABNF [RFC5234] definition.

Session-mux-id-attr = "a=session-mux-id:" SID *SID-prop SID = SID-value / SID-pairs SID-value = 1*3DIGIT / "NoN" SID-pairs = SID-value "/" SID-value ; RTP/RTCP SIDs SID-prop = SP assignment-policy / prop-ext prop-ext = token "=" value assignment-policy = "policy=" ("tentative" / "fixed")

The following parameters MUST be configured as specified:



o RTP Profile SHOULD be the same, but MUST be compatible, like AVP and AVPF.

o RTCP bandwidth parameters are the same

o RTP Payload type values are not overlapping

In declarative SDP usage, there is clearly no method for fallback unless some other negotiation protocol is used.

The SID property "policy" is used in negotiation by an end-point to indicate if the session ID values are merely a tentative suggestion or if they must have these values. This is used when negotiating SID for multi-party RTP sessions to support shared transports such as multicast or RTP translators that are unable to produce renumbered SIDs on a per end-point basis. The normal behavior is that the offer suggest a tentative set of values, indicated by "policy=tentative". These SHOULD be accepted by the peer unless that peer negotiate session IDs on behalf of a centralized policy, in which case it MAY change the value(s) in the answer. If the offer represents a policy that does not allow changing the session ID values, it can indicate that to the answerer by setting the policy to "fixed". This enables the answering peer to either accept the value or indicate that there is a conflict in who is performing the assignment by setting the SID value to NoN (Not a Number). Offerer and answerer SHOULD always include the policy they are operating under. Thus, in case of no centralized behaviors, both offerer and answerer will indicate the tentative policy.

6.3. SRTP Key Management

Key management for SRTP do needs discussion as we do cause multiple SRTP sessions to exist on the same underlying transport flow. Thus we need to ensure that the key management mechanism still are properly associated with the SRTP session context it intends to key. To ensure that we do look at the three SRTP key management mechanism that IETF has specified, one after another.

6.3.1. Security Description

Session Description Protocol (SDP) Security Descriptions for Media Streams [RFC4568] as being based on SDP has no issue with the RTP session multiplexing on lower layer specified here. The reason is that the actual keying is done using a media level SDP attribute. Thus the attribute is already associated with a particular media description. A media description that also will have an instance of the "a=session-mux-id" attribute carrying the SID value/pair used with this particular crypto parameters.



6.3.2. DTLS-SRTP

Datagram Transport Layer Security (DTLS) Extension to Establish Keys for the Secure Real-time Transport Protocol (SRTP) [RFC5764] is a keying mechanism that works on the media plane on the same lower layer transport that SRTP/SRTCP will be transported over. Thus each DTLS message must be associated with the SRTP and/or SRTCP flow it is keying.

The most direct solution is to use the SHIM and the SID context identifier to be applied also on DTLS packets. Thus using the same SID that is used with RTP and/or RTCP also for the DTLS message intended to key that particular SRTP and/or SRTCP flow(s).

6.3.3. MIKEY

MIKEY: Multimedia Internet KEYing [RFC3830] is a key management protocol that has several transports. In some cases it is used directly on a transport protocol such as UDP, but there is also a specification for how MIKEY is used with SDP "Key Management Extensions for Session Description Protocol (SDP) and Real Time Streaming Protocol (RTSP)" [RFC4567].

Lets start with the later, i.e. the SDP transport, which shares the properties with Security Description in that is can be associated with a particular media description in a SDP. As long as one avoids using the session level attribute one can be certain to correctly associate the key exchange with a given SRTP/SRTCP context.

It does appear that MIKEY directly over a lower layer transport protocol will have similar issues as DTLS.

6.4. Examples

6.4.1. RTP Packet with Transport Header

The below figure contains an RTP packet with SID field encapsulated by a UDP packet (added UDP header).



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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Source Port | Destination Port | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Length | Checksum | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<+ |V=2|P|X| CC |M| PT | sequence number | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | timestamp | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | synchronization source (SSRC) identifier | | +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ | | contributing source (CSRC) identifiers | | | .... | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | RTP extension (OPTIONAL) | | +>+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | | payload ... | | | | +-------------------------------+ | | | | RTP padding | RTP pad count | | +>+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<+ | ˜ SRTP MKI (OPTIONAL) ˜ | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | : authentication tag (RECOMMENDED) : | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | | Session ID | | | +---------------+ | +- Encrypted Portion* Authenticated Portion ---+

SRTP Packet Encapsulated by Session ID Layer

6.4.2. SDP Offer/Answer example

This section contains SDP offer/answer examples. First one example of successful BUNDLEing, and then two where fallback occurs.

In the below SDP offer, one audio and one video is being offered. The audio is using SID 0, and the video is using SID 1 to indicate that they are different RTP sessions despite being offered over the same 5-tuple.



v=0 o=alice 2890844526 2890844526 IN IP4 atlanta.example.com s= c=IN IP4 atlanta.example.com t=0 0 a=group:BUNDLE foo bar m=audio 10000 RTP/AVP 0 8 97 b=AS:200 a=mid:foo a=session-mux-id:0 policy=tentative a=rtpmap:0 PCMU/8000 a=rtpmap:8 PCMA/8000 a=rtpmap:97 iLBC/8000 m=video 10000 RTP/AVP 31 32 b=AS:1000 a=mid:bar a=session-mux-id:1 policy=tentative a=rtpmap:31 H261/90000 a=rtpmap:32 MPV/90000

The SDP answer from an end-point that supports this BUNDLEing: v=0 o=bob 2808844564 2808844564 IN IP4 biloxi.example.com s= c=IN IP4 biloxi.example.com t=0 0 a=group:BUNDLE foo bar m=audio 20000 RTP/AVP 0 b=AS:200 a=mid:foo a=session-mux-id:0 policy=tentative a=rtpmap:0 PCMU/8000 m=video 20000 RTP/AVP 32 b=AS:1000 a=mid:bar a=session-mux-id:1 policy=tentative a=rtpmap:32 MPV/90000

The SDP answer from an end-point that does not support this BUNDLEing or the general signalling of [I-D.ietf-mmusic-sdp-bundle-negotiation].



v=0 o=bob 2808844564 2808844564 IN IP4 biloxi.example.com s= c=IN IP4 biloxi.example.com t=0 0 m=audio 20000 RTP/AVP 0 b=AS:200 a=rtpmap:0 PCMU/8000 m=video 30000 RTP/AVP 32 b=AS:1000 a=rtpmap:32 MPV/90000

The SDP answer of a client supporting [I-D.ietf-mmusic-sdp-bundle-negotiation] but not this BUNDLEing would look like this: v=0 o=bob 2808844564 2808844564 IN IP4 biloxi.example.com s= c=IN IP4 biloxi.example.com t=0 0 a=group:BUNDLE foo bar m=audio 20000 RTP/AVP 0 a=mid:foo b=AS:200 a=rtpmap:0 PCMU/8000 m=video 20000 RTP/AVP 32 a=mid:bar b=AS:1000 a=rtpmap:32 MPV/90000

In this last case, the result is a sing RTP session with both media types being established. If that isn’t supported or desired, the offerer will have to either re-invite without the BUNDLE grouping to force different 5-tuples, or simply terminate the session.

7. Open Issues

This work is still in the early phase of specification. This section contains a list of open issues where the author desires some input.

1. Should RTP and RTCP multiplexing without RFC 5761 support be included?

2. In Section 6.2 there is a discussion of which parameters that must be configured. The scope of these rules and if they do make sense needs additional discussion.



3. Can we provide better control so that applications that doesn’t desire fallback to single RTP session when Multiplexing shim fails to be supported but Bundle is supported ends up with a better alternative?


This document request the registration of one SDP attribute. Details of the registration to be filled in.


The security properties of the Session ID layer is depending on what mechanism is used to protect the RTP and RTCP packets of a given RTP session. If IPsec or transport layer security solutions such as DTLS or TLS are being used then both the encapsulated RTP/RTCP packets and the session ID layer will be protected by that security mechanism. Thus potentially providing both confidentiality, integrity and source authentication. If SRTP is used, the session ID layer will not be directly protected by SRTP. However, it will be implicitly integrity protected (assuming the RTP/RTCP packet is integrity protected) as the only function of the field is to identify the session context. Thus any modification of the SID field will attempt to retrieve the wrong SRTP crypto context. If that retrieval fails, the packet will be anyway be discarded. If it is successful, the context will not lead to successful verification of the packet.


This document is based on the input from various people, especially in the context of the RTCWEB discussion of how to use only a single lower layer transport. The RTP and RTCP packet figures are borrowed from RFC3711. The SDP example is extended from the one present in [I-D.ietf-mmusic-sdp-bundle-negotiation]. The authors would like to thank Christer Holmberg for assistance in utilizing the BUNDLE grouping mechanism.

The proposal in Section 4.5 is original suggested by Colin Perkins. The idea in Section 4.6 is from an Internet Draft [I-D.rosenberg-rtcweb-rtpmux] written by Jonathan Rosenberg et. al. The proposal in Section 4.3 is a result of discussion by a group of people at IETF meeting #81 in Quebec.

11. References




[I-D.ietf-mmusic-sdp-bundle-negotiation] Holmberg, C. and H. Alvestrand, "Multiplexing Negotiation Using Session Description Protocol (SDP) Port Numbers", draft-ietf-mmusic-sdp-bundle-negotiation-00 (work in progress), February 2012.




[RFC5234] Crocker, D. and P. Overell, "Augmented BNF for Syntax Specifications: ABNF", STD 68, RFC 5234, January 2008.

11.2. Informational References

[I-D.lennox-rtcweb-rtp-media-type-mux] Rosenberg, J. and J. Lennox, "Multiplexing Multiple Media Types In a Single Real-Time Transport Protocol (RTP) Session", draft-lennox-rtcweb-rtp-media-type-mux-00 (work in progress), October 2011.

[I-D.rosenberg-rtcweb-rtpmux] Rosenberg, J., Jennings, C., Peterson, J., Kaufman, M., Rescorla, E., and T. Terriberry, "Multiplexing of Real- Time Transport Protocol (RTP) Traffic for Browser based Real-Time Communications (RTC)", draft-rosenberg-rtcweb-rtpmux-00 (work in progress), July 2011.

[I-D.westerlund-avtcore-multiplex-architecture] Westerlund, M., Burman, B., and C. Perkins, "RTP Multiplexing Architecture", draft-westerlund-avtcore-multiplex-architecture-01 (work in progress), March 2012.




[RFC3830] Arkko, J., Carrara, E., Lindholm, F., Naslund, M., and K. Norrman, "MIKEY: Multimedia Internet KEYing", RFC 3830, August 2004.

[RFC4567] Arkko, J., Lindholm, F., Naslund, M., Norrman, K., and E. Carrara, "Key Management Extensions for Session Description Protocol (SDP) and Real Time Streaming Protocol (RTSP)", RFC 4567, July 2006.

[RFC4568] Andreasen, F., Baugher, M., and D. Wing, "Session Description Protocol (SDP) Security Descriptions for Media Streams", RFC 4568, July 2006.

[RFC5285] Singer, D. and H. Desineni, "A General Mechanism for RTP Header Extensions", RFC 5285, July 2008.

[RFC5506] Johansson, I. and M. Westerlund, "Support for Reduced-Size Real-Time Transport Control Protocol (RTCP): Opportunities and Consequences", RFC 5506, April 2009.

[RFC5761] Perkins, C. and M. Westerlund, "Multiplexing RTP Data and Control Packets on a Single Port", RFC 5761, April 2010.

[RFC5764] McGrew, D. and E. Rescorla, "Datagram Transport Layer Security (DTLS) Extension to Establish Keys for the Secure Real-time Transport Protocol (SRTP)", RFC 5764, May 2010.


Magnus Westerlund Ericsson Farogatan 6 SE-164 80 Kista Sweden





Audio/Video Transport Core A. WilliamsMaintenance AudinateInternet-Draft R. van BrandenburgIntended status: Standards Track TNOExpires: August 31, 2012 K. Gross AVA Networks February 28, 2012

RTP Clock Source Signalling draft-williams-avtcore-clksrc-00

Abstract

NTP timestamps are used by several RTP protocols for synchronisation and statistical measurement. This memo specificies SDP signalling identifying NTP timestamp clock sources and SDP signalling identifying the media clock sources in a multimedia session.

Requirements Language

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119 [1].

Status of this Memo




This Internet-Draft will expire on August 31, 2012.

Copyright Notice


This document is subject to BCP 78 and the IETF Trust’s Legal

Williams, et al. Expires August 31, 2012 [Page 1]

Internet-Draft RTP Clock Source Signalling February 2012

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.

Table of Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 2. Applications . . . . . . . . . . . . . . . . . . . . . . . . . 3 3. Definitions . . . . . . . . . . . . . . . . . . . . . . . . . 4 4. Timestamp Reference Clock Source Signalling . . . . . . . . . 5 4.1. Equivalent Timestamp Clocks . . . . . . . . . . . . . . . 5 4.2. Identifying NTP Reference Clocks . . . . . . . . . . . . . 6 4.3. Identifying PTP Reference Clocks . . . . . . . . . . . . . 6 4.4. Identifying Global Reference Clocks . . . . . . . . . . . 8 4.5. Other Reference Clocks . . . . . . . . . . . . . . . . . . 8 4.6. Traceable Reference Clocks . . . . . . . . . . . . . . . . 8 4.7. Synchronisation Confidence . . . . . . . . . . . . . . . . 8 4.8. SDP Signalling of Timestamp Clock Source . . . . . . . . . 9 4.8.1. Examples . . . . . . . . . . . . . . . . . . . . . . . 11 5. Timescales, UTC TAI and leap seconds . . . . . . . . . . . . . 12 6. Media Clock Source Signalling . . . . . . . . . . . . . . . . 13 7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 14 8. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 14 9. References . . . . . . . . . . . . . . . . . . . . . . . . . . 14 9.1. Normative References . . . . . . . . . . . . . . . . . . . 14 9.2. Informative References . . . . . . . . . . . . . . . . . . 15 Appendix A. An Appendix . . . . . . . . . . . . . . . . . . . . . 15 Authors’ Addresses . . . . . . . . . . . . . . . . . . . . . . . . 16



1. Introduction

RTP protocols use NTP format timestamps to facilitate media stream synchronisation and for providing estimates of round trip time (RTT) and other statistical parameters.

Information about media clock timing exchanged in NTP format timestamps may come from a clock which is synchronised to a global time reference, but this cannot be assumed nor is there a standardised mechanism available to indicate that timestamps are derived from a common reference clock. Therefore, RTP implementations typically assume that NTP timestamps are taken using unsynchronised clocks and must compensate for absolute time differences and rate differences. Without a shared reference clock, RTP can time align flows from the same source at a given receiver using relative timing, however tight synchronisation between two or more different receivers (possibly with different network paths) or between two or more senders is not possible.

High performance AV systems often use a reference media clock distributed to all devices in the system. The reference media clock is often distinct from the the reference clock used to provide timestamps. A reference media clock may be provided along with a audio or video signal interface, or via a dedicated clock signal (e.g. genlock [9] or audio word clock [10]. If sending and receiving media clocks are known to be synchronised to a common reference clock, performance can improved by minimising buffering and avoiding rate conversion.

This specification defines SDP signalling of timestamp clock sources and media reference clock sources.

2. Applications

Timestamp clock source and reference media clock signalling benefit applications requiring synchronised media capture or playout and low latency operation.

Exmaples include, but are not limited to:

Social TV RTCP for inter-destination media synchronization [4] defines social TV as the combination of media content consumption by two or more users at different devices and locations and real- time communication between those users. An example of Social TV, is when two or more users are watching the same television broadcast at different devices and locations, while communicating with each other using text, audio and/or video. A skew in the



media play-out of the two or more users can have adverse effects on their experience. A well-known use case here is one friend experiencing a goal in a football match well before or after other friend(s).

Video Walls A video wall consists of multiple computer monitors, video projectors, or television sets tiled together contiguously or overlapped in order to form one large screen. Each of the screens reproduces a portion of the larger picture. In some implementations, each screen may be individually connected to the network and receive its portion of the overall image from a network-connected video server or video scaler. Screens are refreshed at 60 hertz (every 16-2/3 milliseconds) or potentially faster. If the refresh is not synchronized, the effect of multiple screens acting as one is broken.

Netwoked Audio Networked loudspeakers, amplifiers and analogue I/O devices transmitting or receiving audio signals via RTP can be connected to various parts of a building or campus network. Such situations can for example be found in large conference rooms, legislative chambers, classrooms (especially those supporting distance learning) and other large-scale environments such as stadiums. Since humans are more susceptible to differences in audio delay, this use case needs even more accuracy than the video wall use case. Depending on the exact application, the need for accuracy can then be in the range of microseconds [11].

Sensor Arrays Sensor arrays contain many synchronised measurement elements producing signals which are then combined to form an overall measurement. Accurate capture of the phase relationships between the various signals arriving at each element of the array is critically important for proper operation. Examples include towed or fixed sonar arrays, seismic arrays and phased arrays.

3. Definitions

The definitions of streams, sources and levels of information in SDP descriptions follow the definitions found in Source-Specific Media Attributes in the Session Description Protocol (SDP) [2].

multimedia session A set of multimedia senders and receivers as well as the data streams flowing from senders to receivers. The Session Description Protocol (SDP) [3] describes multimedia sessions.



media stream An RTP session potentially containing more than one RTP source. SDP media descriptions beginning with an "m"-line define the parameters of a media stream.

media source A media source is single stream of RTP packets, identified by an RTP SSRC.

session-level Session-level information applies to an entire multimedia session. In an SDP description, session-level information appears before the first "m"-line.

media-level Media-level information applies to a single media stream (RTP session). In an SDP description, media-level information appears after each "m"-line.

source-level Source-level information applies to a single stream of RTP packets, identified by an RTP SSRC Source-Specific Media Attributes in the Session Description Protocol (SDP) [2] defines how source-level information is included into an SDP session description.

traceable time A clock is considered to provide traceable time if it can be proven to be synchronised to a global time reference. GPS XXX is commonly used to provide a traceable time reference. Some network time synchronisation protocols (e.g. XXX PTP) can explicitly indicate that the master clock is providing a traceable time reference over the network.

4. Timestamp Reference Clock Source Signalling

The NTP timestamps used by RTP are taking by reading a local clock at the sender or receiver. This local clock may be synchronised to another clock (time source) by some means or it may be unsynchronised. A variety of methods are available to synchronise local clocks to a reference time source, including network time protocols (e.g. NTP [5]) and radio clocks like GPS [XXX].

The following sections describe and define SDP signalling indicating whether and how the local timestamping clock in an RTP sender/ receiver is synchronised to a reference clock.

4.1. Equivalent Timestamp Clocks

Two or more local clocks that are sufficiently synchronised will produce timestamps for a given event which are effectively identical for the purposes of RTP. A local clock in one RTP sender/receiver can be considered equivalent to a local clock in another RTP sender/



receiver providing they are sufficiently synchronised such that timestamps produced by one clock are indistinguishable from timestamps produced by the other. The timestamps produced by equivalent local clocks in two or more RTP senders/receivers receivers can be directly compared.

One or more local clocks are equivalent if they are synchronised to a single master clock via a network time protocol (e.g. XXX NTP, 802.1AS, IEEE1588v2).

One or more local clocks are equivalent if they are synchronised to any member of a set of master clocks provided that the set of master clocks are synchronised.

One or more local clocks are equivalent if they are synchronised to a clock master providing a global time reference (e.g. XXX GPS, Gallileo). Some network time protocols (e.g. XXX PTP) may allow master clocks to explicitly indicate that they are "traceable" back to a global time reference.

4.2. Identifying NTP Reference Clocks

A single NTP server is identified by identified by hostname (or IP address) and an optional port number. If the port number is not indicated, it is assumed to be the standard NTP port (123) XXX.

Two or more NTP servers may be listed to indicate that they are interchangeable. RTP senders/receivers can use any of the listed NTP servers to govern a local clock that is equivalent to a local clock slaved to a difference server.

XXX Question: Does NTP carry traceability information? Or is this implicit somehow in the stratum? Apparently there are some bits in the leap seconds functionality which talk about "tracking"..

4.3. Identifying PTP Reference Clocks

The IEEE1588 Precision Time Protocol (PTP) family of clock synchronisation protocols provide a shared reference clock in an network - typically a LAN. IEEE1588 provides sub-microsecond synchronisation between devices on a LAN and typically locks within seconds at startup. With support from Ethernet switches, IEEE1588 protocols can achieve nanosecond timing accuracy in LANs. Network interface chips and cards supporting hardware time-stamping of timing critical protocol messages are also available.

When using IEEE1588 clock synchronisation, networked AV systems can achieve sub 1 microsecond time alignment accuracy when rendering AV



signals and can support latencies less than 1ms through a gigabit LAN.

Three flavours of IEEE1588 are in use today:

o IEEE 1588-2002 [6]: the original "Standard for a Precision Clock Synchronization Protocol for Networked Measurement and Control Systems". This is often called IEEE1588v1 or PTPv1.

o IEEE 1588-2008 [7]: the second version of the "Standard for a Precision Clock Synchronization Protocol for Networked Measurement and Control Systems". This is a revised version of the original IEEE1588-2002 standard and is often called IEEE1588v2 or PTPv2.

o IEEE 802.1AS [8]: "Timing and Synchronization for Time Sensitive Applications in Bridged Local Area Networks". This is a Layer-2 only profile of IEEE 1588-2008 for use in Audio/Video Bridged LANs.

Each IEEE1588 clock is identified by a globally unique EUI-64 called a "ClockIdentity". A slave clock using one of the IEEE1588 family of network time protocols acquires the ClockIdentity/EUI-64 of the grandmaster clock that is the ultimate source of timing information. A master clock which is itself slaved to another master clock passes the grand master clock identity through to its slaves.

Several instances of the IEEE1588v1/v2 protocol may operate independently on a single network, forming distinct PTP network protocol domains each of which may have a different master clock. As the IEEE1588 standards have developed, the definition of PTP domains has changed. IEEE1588v1 identifies protocol subdomains by a textual name and IEEE1588v2 identifies protocol domains using a numeric domain number. 802.1AS is a Layer2 profile of IEEE1588v2 supporting a single numeric clock domain (0). This specification assumes that an IEEE1588 clock master for multiple domains will provide the same timing information to all domains or that each clock domain has a different master. In other words, this specification assumes that a timing domain can be uniquely identified using the ClockIdentity of the grandmaster clock alone.

The PTP family of protocols employ a distributed election protocol called the "Best Master Clock Algorithm (BMCA) to determine the active clock master. The clock master choices available to BMCA can be restricted or favourably biased by setting stratum values, preffered master clock bits, or other parameters to influence the election process. In some systems it may be desirable to limit the number of possible PTP clock masters to avoid re-signalling timestamp clock sources when the clock master changes.



4.4. Identifying Global Reference Clocks

Global reference clocks provide a source of tracable time, typically via a hardware radio receiver interface. Examples include GPS and Gallileo. Apart from the name of the reference clock system, no further identification is required.

4.5. Other Reference Clocks

At the time of writing, it is common for RTP senders/receivers not to synchronise their local timestamp clock to a master. An unsynchronised clock such as a quartz oscillator is identified as a "local" reference clock.

In some systems, all RTP senders/receivers may use a timetsamp clock synchronised to a reference clock that is not provided by one of the methods listed above. Examples may include the reference time information provided by digital television or cellular services. These sources are identified as "private" reference clocks. All RTP senders/receivers in a session using a private reference clock are assumed to have a mechanism outside this specification confirming that their local timestamp clocks are equivalent.

4.6. Traceable Reference Clocks

A timestamp clock source may be labelled "traceable" if it is known to be sourced from a global time reference such as TAI or UTC XXX. Providing adjustments are made for differing time bases, timestamps taken using a clocks synchronised to a traceable time source can be directly compared even if the clocks are synchronised to different servers or via different mechanisms. Any traceable timestamp clock source can be considered equivalent to another traceable timestamp clock source and the timestamps may be directly compared.

Since any NTP or PTP server providing traceable time can be considered equivalent, it is not necessary to identify traceable time servers by protocol address.

4.7. Synchronisation Confidence

Network time protocols periodically exchange timestamped messages between servers and clients. Assuming RTP sender/receiver clocks are based on commonly available quartz crystal hardware, tight synchronisation requires frequent exchange of synchronisation messages.

Unfortunately, in some implementations, it is not possible to control the frequency of synchronisation messages nor is it possible to



discover when the last sychronisation message occured. In order to provide a measure of confidence that the timestamp clock is sufficiently synchronised, an optional timestamp may be included in the SDP clock source signalling. In addition, the frequency of synchronisation message may also optionally be provided.

The optional timestamp and synchronisation frequency parameters provide an indication of synchronisation quality to the receiver of those parameters. If the synchronisation confidence timestamp is far from the timestamp clock at the receiver of the parameters, it can be assumed that synchronisation has not occured recently or the timestamp reference clock source is wrongly configured or cannot be contacted. In this case, the receiver can take action to prevent unsynchronised playout or may fall back to assuming that the timestamp clocks are not synchronised.

Synchronisation frequency is expressed as an 8-bit excess-127 field which is the base 2 logarithm of the frequency in HZ. The synchronisation frequencies represented by this field range from 2^-127 Hz to 2^+128 Hz. The field value of 127 corresponds to an update frequency of 1 Hz.

4.8. SDP Signalling of Timestamp Clock Source

Specification of the timestamp reference clock source may at all levels of an SDP description (see level definitions (Section 3) earlier in this document for more information).

Timestamp clock source signalling included at session-level provides default parameters for all RTP sessions and sources in the session description. More specific signalling included at the media-level overrides default session-level signalling. Further, source-level signalling overrides timestamp clock source signalling at the enclosing media-level and session-level.

If timestamp clock source signalling is included anywhere in an SDP description, it must be properly defined for all levels in the description. This may simply be achieved by providing default signalling at the session level.

Timestamp reference clock parameters may be repeated at a given level (i.e. for a session or source) to provide information about additional servers or clock sources. If the attribute is repeated at a given level, all clocks described at that level are assumed to be equivalent. Traceable clock sources MUST NOT be mixed with clock sources having explicit addresses for a given source or session. Unless synchronisation confidence information is available for each of the reference clocks listed at a given level, it SHOULD only be



included with the first reference clock source attribute at that level.

Note that clock source parameters may change from time to time, for example, as a result of a PTP clock master election. The SIP XXX protocol supports re-signalling of updated SDP information, however other protocols may require additional notification mechanisms.

timestamp-refclk = "a=ts-refclk:" clksrc SP [sync-confidence] CRLF

clksrc = ntp / ptp / gps / gal / local / private

ntp = "ntp=" ntp-server-addr ntp-server-addr = host [ ":" port ] ntp-server-addr =/ "traceable" )

ptp = "ptp=" ptp-version ":" ptp-gmid ptp-version = "IEEE1588-2002" ptp-version =/ "IEEE1588-2008" ptp-version =/ "IEEE802.1AS-2011" ptp-gmid = EUI64 ptp-gmid =/ "traceable"

gps = "gps" gal = "gal" local = "local" private = "private" [ ":" "traceable" ]

sync-confidence = sync-timestamp [SP sync-frequency]

sync-timestamp = sync-date SP sync-time SP sync-UTCoffset

sync-date = 4DIGIT "-" 2DIGIT "-" 2DIGIT ; yyyy-mm-dd (e.g., 1982-12-02)

sync-time = 2DIGIT ":" 2DIGIT ":" 2DIGIT "." 3DIGIT ; 00:00:00.000 - 23:59:59.999

sync-UTCoffset = ( "+" / "-" ) 2DIGIT ":" 2DIGIT ; +HH:MM or -HH:MM

sync-frequency = 2HEXDIG ; If N is the field value, HZ=2^(N-127)

host = hostname / IPv4address / IPv6reference



hostname = *( domainlabel "." ) toplabel [ "." ] toplabel = ALPHA / ALPHA *( alphanum / "-" ) alphanum domainlabel = alphanum =/ alphanum *( alphanum / "-" ) alphanum

IPv4address = 1*3DIGIT "." 1*3DIGIT "." 1*3DIGIT "." 1*3DIGIT IPv6reference = "[" IPv6address "]" IPv6address = hexpart [ ":" IPv4address ] hexpart = hexseq / hexseq "::" [ hexseq ] / "::" [ hexseq ] hexseq = hex4 *( ":" hex4) hex4 = 1*4HEXDIG

port = 1*DIGIT

EUI-64 = 7(HEXDIG "-") 2HEXDIG

Figure 1: Timestamp Reference Clock Source Signalling

4.8.1. Examples

Figure 2 shows an example SDP description with a timestamp reference clock source defined at the session-level.

v=0 o=jdoe 2890844526 2890842807 IN IP4 10.47.16.5 s=SDP Seminar i=A Seminar on the session description protocol u=http://www.example.com/seminars/sdp.pdf [email protected] (Jane Doe) c=IN IP4 224.2.17.12/127 t=2873397496 2873404696 a=recvonly a=ts-refclk:ntp=traceable m=audio 49170 RTP/AVP 0 m=video 51372 RTP/AVP 99 a=rtpmap:99 h263-1998/90000

Figure 2: Timestamp reference clock defintion at the session level

Figure 3 shows an example SDP description with timestamp reference clock definitions at the media-level overriding the session-level defaults. Note that the synchronisation confidence timestamp appears on the first attribute at the media-level only.



v=0 o=jdoe 2890844526 2890842807 IN IP4 10.47.16.5 s=SDP Seminar i=A Seminar on the session description protocol u=http://www.example.com/seminars/sdp.pdf [email protected] (Jane Doe) c=IN IP4 224.2.17.12/127 t=2873397496 2873404696 a=recvonly a=ts-refclk:local m=audio 49170 RTP/AVP 0 a=ts-refclk:ntp=203.0.113.10 2011-02-19 21:03:20.345+01:00 a=ts-refclk:ntp=198.51.100.22 m=video 51372 RTP/AVP 99 a=rtpmap:99 h263-1998/90000 a=ts-refclk:ptp=IEEE802.1AS-2011:39-A7-94-FF-FE-07-CB-D0

Figure 3: Timestamp reference clock definition at the media-level

Figure 4 shows an example SDP description with a timestamp reference clock definition at the source-level overriding the session-level default.

v=0 o=jdoe 2890844526 2890842807 IN IP4 10.47.16.5 s=SDP Seminar i=A Seminar on the session description protocol u=http://www.example.com/seminars/sdp.pdf [email protected] (Jane Doe) c=IN IP4 224.2.17.12/127 t=2873397496 2873404696 a=recvonly a=ts-refclk:local m=audio 49170 RTP/AVP 0 m=video 51372 RTP/AVP 99 a=rtpmap:99 h263-1998/90000 a=ssrc:12345 ts-refclk:ptp=IEEE802.1AS-2011:39-A7-94-FF-FE-07-CB-D0

Figure 4: Timestamp reference clock signalling at the source level

5. Timescales, UTC TAI and leap seconds

RTP implementation is simplified by using a clock reference with a timescale which does not include leap seconds. IEEE 1588, GPS and other TAI (Inernational Atomic Time) references do not include leap seconds. NTP time, operating system clocks and other UTC (Coordinated Universal Time) references include leap seconds (though



the ITU is studying a proposal which could eventually eliminate leap seconds from UTC).

Leap seconds are potentially scheduled at the end of the last day of December and June each year. NTP inserts a leap second at the beginning of the last second of the day. This results in the clock freezing for one second immediately prior to the last second of the affected day. Most system clocks insert the leap second at the end of the last second. This results in repetition of the last second of the day. Generating or using timestamps during the entire last second of a day on which a leap second has been scheduled should therefore be avoided. Note that the period to be avoided has a real- time duration of two seconds.

It is also important that all participants correctly implement leap seconds and have a working communications channel to receive notification of leap second scheduling. Without prior knowledge of leap second schedule, NTP servers and clients may be offset by exactly one second with respect to their UTC reference. This potential discrepancy begins when a leap second occurs and ends when all participants receive a time update from a server or peer (which, depending on the operating system and/or implementation, could be anywhere from a few minutes to a week). Such a long-lived discrepancy can be particularly disruptive to RTP operation.

Apart from the long-lived discrepancy due to dependence on both timing (e.g. NTP) updates and leap seconds scheduling updates, there is also the potential for a short-lived timing discontinuity having an effect on RTP playout timing (even though leap seconds are quite rare).

If a timescale with leap seconds is used for RTP:

o RTP Senders using a leap-second-bearing reference must not generate sender reports (SR) containing an originating NTP timestamp in the vicinity of a leap second. Receivers should ignore timestamps in any such reports inadvertently generated.

o Receivers working to a leap-second-bearing reference must be careful to take leap seconds into account if a leap second occurs between the time a RTP packet is originated and when it is to be presented.

6. Media Clock Source Signalling

A timestamp clock source (ie media clock is locked to a reference clock like NTP, GPS, etc) Reference clock source..



An RTP session.. This should be an SSRC within an RTP session. Include IP address and port.

An IEEE 1722 stream, identified by a Stream ID.


The SDP attribute "ts-clksrc" defined by this document is registered with the IANA registry of SDP Parameters as follows:

SDP Attribute ("att-field"):

Attribute name: ts-refclk

Long form: Timestamp reference clock source

Type of name: att-field

Type of attribute: session, media and source level

Subject to charset: no

Purpose: See sections 1-4 of this document

Reference: This document

Values: see this document and registrations below

The attribute has an extensible parameter field and therefore a registry for these parameters is required. This document creates an IANA registry called the Timestamp Reference Clock Source Parameters Registry. It contains the six parameters defined in Figure 1: "ntp", "ptp", "gps", "gal", "local", "private".

8. Acknowledgements

9. References


[1] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997.

[2] Lennox, J., Ott, J., and T. Schierl, "Source-Specific Media



Attributes in the Session Description Protocol (SDP)", RFC 5576, June 2009.

[3] Handley, M., Jacobson, V., and C. Perkins, "SDP: Session Description Protocol", RFC 4566, July 2006.


[4] Brandenburg, R., Stokking, H., Deventer, O., Boronat, F., Montagud, M., and K. Gross, "RTCP for inter-destination media synchronization", draft-ietf-avtcore-idms-02 (work in progress), October 2011.

[5] Mills, D., Martin, J., Burbank, J., and W. Kasch, "Network Time Protocol Version 4: Protocol and Algorithms Specification", RFC 5905, June 2010.

[6] Institute of Electrical and Electronics Engineers, "1588-2002 - IEEE Standard for a Precision Clock Synchronization Protocol for Networked Measurement and Control Systems", IEEE Std 1588-2002, 2002, <http://standards.ieee.org/findstds/standard/1588-2002.html>.

[7] Institute of Electrical and Electronics Engineers, "1588-2008 - IEEE Standard for a Precision Clock Synchronization Protocol for Networked Measurement and Control Systems", IEEE Std 1588-2008, 2008, <http://standards.ieee.org/findstds/standard/1588-2008.html>.

[8] "Timing and Synchronization for Time-Sensitive Applications in Bridged Local Area Networks", <http://standards.ieee.org/findstds/standard/802.1AS-2011.html>.

URIs

[9] <http://en.wikipedia.org/wiki/Genlock>

[10] <http://en.wikipedia.org/wiki/Word_clock>

[11] <http://www.ieee802.org/1/files/public/docs2007/ as-dolsen-time-accuracy-0407.pdf>

Appendix A. An Appendix




Aidan Williams Audinate Level 1, 458 Wattle St Ultimo, NSW 2007 Australia

Phone: +61 2 8090 1000 Fax: +61 2 8090 1001 Email: [email protected] URI: http://www.audinate.com/

Ray van Brandenburg TNO Brassersplein 2 Delft, The Netherlands

Phone: +31 88 86 63609 Fax: Email: [email protected] URI:

Kevin Gross AVA Networks

Phone: Fax: Email: URI:
