| Network Working Group | M. Westerlund |
| Internet-Draft | B. Burman |
| Intended status: Informational | Ericsson |
| Expires: May 3, 2018 | C. Perkins |
| University of Glasgow | |
| H. Alvestrand | |
| R. Even | |
| H. Zheng | |
| Huawei | |
| October 30, 2017 |
Guidelines for using the Multiplexing Features of RTP to Support Multiple Media Streams
draft-ietf-avtcore-multiplex-guidelines-05
The Real-time Transport Protocol (RTP) is a flexible protocol that can be used in a wide range of applications, networks, and system topologies. That flexibility makes for wide applicability, but can complicate the application design process. One particular design question that has received much attention is how to support multiple media streams in RTP. This memo discusses the available options and design trade-offs, and provides guidelines on how to use the multiplexing features of RTP to support multiple media streams.
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."
This Internet-Draft will expire on May 3, 2018.
Copyright (c) 2017 IETF Trust and the persons identified as the document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include 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.
The Real-time Transport Protocol (RTP) 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 was from the beginning designed for multiple participants in a communication session. It supports many paradigms of topologies and usages, as defined in [RFC7667]. 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.
There have been increased interest in more advanced usage of RTP, for example, multiple streams can occur when a single endpoint have multiple media sources, like multiple cameras or microphones that need to be sent simultaneously. Consequently, questions are raised regarding the most appropriate RTP usage. The limitations in some implementations, RTP/RTCP extensions, and signalling has also been exposed. The authors also hope that clarification on the usefulness of some functionalities in RTP will result in more complete implementations in the future.
The purpose of this document is to provide clear information about the possibilities of RTP when it comes to multiplexing. The RTP application designer needs to understand the implications that come from a particular usage of the RTP multiplexing points. The document will recommend against some usages as being unsuitable, in general or for particular purposes.
The document starts with some definitions and then goes into the existing RTP functionalities around multiplexing. Both the desired behaviour and the implications of a particular behaviour depend on which topologies are used, which requires some consideration. This is followed by a discussion of some choices in multiplexing behaviour and their impacts. Some archetypes of RTP usage are discussed. Finally, some recommendations and examples are provided.
The definitions in Section 3 of [RFC3550] are referenced normatively.
The taxonomy defined in [RFC7656] is referenced normatively.
The following terms and abbreviations are used in this document:
This document is focused on issues that affect RTP. Thus, issues that involve signalling protocols, such as whether SIP, Jingle or some other protocol is in use for session configuration, the particular syntaxes used to define RTP session properties, or the constraints imposed by particular choices in the signalling protocols, are mentioned only as examples in order to describe the RTP issues more precisely.
This document assumes the applications will use RTCP. While there are such applications that don't send RTCP, they do not conform to the RTP specification, and thus can be regarded as reusing the RTP packet format but not implementing the RTP protocol.
The reasons why an endpoint might choose to send multiple media streams are widespread. 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:
For each of these, it is necessary to decide if each additional media stream gets its own SSRC multiplexed within a RTP Session, or if it is necessary to use additional RTP sessions to group the media streams. The choice between these made due to one reason might not be the choice suitable for another reason. 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. As the discussion below will show, in reality we cannot choose a single one of the two solutions. To utilise 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.
This section describes the multiplexing points present in the RTP protocol that can be used to distinguish media streams and groups of media streams. Figure 1 outlines the process of demultiplexing incoming RTP streams:
|
| packets
+-- v
| +------------+
| | Socket |
| +------------+
| || ||
RTP | RTP/ || |+-----> SCTP ( ...and any other protocols)
Session | RTCP || +------> STUN (multiplexed using same port)
+-- ||
+-- ||
| (split by SSRC)
| || || ||
| || || ||
Media | +--+ +--+ +--+
Streams | |PB| |PB| |PB| Jitter buffer, process RTCP, FEC, etc.
| +--+ +--+ +--+
+-- | | |
(pick rendering context based on PT)
+-- | / |
| +---+ |
| / | |
Payload | +--+ +--+ +--+
Formats | |CR| |CR| |CR| Codecs and rendering
| +--+ +--+ +--+
+--
Figure 1: RTP Demultiplexing Process
An RTP Session is the highest semantic layer in the RTP protocol, and represents an association between a group of communicating endpoints. The set of participants that form an RTP session is defined as those that share a single synchronisation source space [RFC3550]. That is, if a group of participants are each aware of the synchronisation source identifiers belonging to the other participants, then those participants are in a single RTP session. A participant can become aware of a synchronisation source identifier by receiving an RTP packet containing it in the SSRC field or CSRC list, by receiving an RTCP packet mentioning it in an SSRC field, or through signalling (e.g., the SDP "a=ssrc:" attribute). Thus, the scope of an RTP session is determined by the participants' network interconnection topology, in combination with RTP and RTCP forwarding strategies deployed by the endpoints and any middleboxes, and by the signalling.
RTP does not contain a session identifier. Rather, it relies on the underlying transport layer to separate different sessions, and on the signalling to identify sessions in a manner that is meaningful to the application. The signalling layer might give sessions an explicit identifier, or their identification might be implicit based on the addresses and ports used. Accordingly, a single RTP Session can have multiple associated identifiers, explicit and implicit, belonging to different contexts. For example, when running RTP on top of UDP/IP, an RTP endpoint can identify and delimit an RTP Session from other RTP Sessions using the UDP source and destination IP addresses and UDP port numbers. Another example is when using SDP grouping framework [RFC5888] which uses an identifier per "m="-line; if there is a one-to-one mapping between "m="-lines and RTP sessions, that grouping framework identifier will identify an RTP Session. [I-D.ietf-mmusic-sdp-bundle-negotiation] extends the "m-"-line for bundled media, which adds complexity to demultiplexing media stream. Section 10.2 of [I-D.ietf-mmusic-sdp-bundle-negotiation] provides information about how RTP/RTCP streams are associated with SDP media description.
RTP sessions are globally unique, but their identity can only be determined by the communication context at an endpoint of the session, or by a middlebox that is aware of the session context. The relationship between RTP sessions depending on the underlying application, transport, and signalling protocol. The RTP protocol makes no normative statements about the relationship between different RTP sessions, however the applications that use more than one RTP session will have some higher layer understanding of the relationship between the sessions they create.
A synchronisation source (SSRC) identifies an RTP source or an RTP sink. Every endpoint will have at least one synchronisation source identifier, even if it does not send media (endpoints that are only RTP sinks still send RTCP, and use their synchronisation source identifier in the RTCP packets they send). An endpoint can have multiple synchronisation sources identifiers if it contains multiple RTP sources (i.e., if it sends multiple media streams). Endpoints that are both RTP sources and RTP sinks use the same synchronisation sources in both roles. At any given time, a RTP source has one and only one SSRC - although that can change over the lifetime of the RTP source or sink.
The synchronisation Source identifier is a 32-bit unsigned integer. It is present in every RTP and RTCP packet header, and in the payload of some RTCP packet types. It can also be present in SDP signalling. Unless pre-signalled using the SDP "a=ssrc:" attribute [RFC5576], the synchronisation source identifier is chosen at random. It is not dependent on the network address of the endpoint, and is intended to be unique within an RTP session. Synchronisation source identifier collisions can occur, and are handled as specified in [RFC3550] and [RFC5576], resulting in the synchronisation source identifier of the affecting RTP sources and/or sinks changing. An RTP source that changes its RTP Session identifier (e.g. source transport address) during a session has to choose a new SSRC identifier to avoid being interpreted as looped source.
Synchronisation source identifiers that belong to the same synchronisation context (i.e., that represent media streams that can be synchronised using information in RTCP SR packets) are indicated by use of identical CNAME chunks in corresponding RTCP SDES packets. SDP signalling can also be used to provide explicit grouping of synchronisation sources [RFC5576].
Note that RTP sequence number and RTP timestamp are scoped by the synchronisation source. Each RTP source will have a different synchronisation source, and the corresponding media stream will have a separate RTP sequence number and timestamp space.
An SSRC identifier is used by different type of sources as well as sinks:
Note that an endpoint that generates more than one media type, e.g. a conference participant sending both audio and video, need not (and commonly does 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 endpoint. The main property attributed to SSRCs associated with the same CNAME is that they are from a particular synchronisation context and can be synchronised at playback.
An RTP receiver receiving a previously unseen SSRC value will interpret it as a new source. It might 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 ought to 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.
The Contributing Source (CSRC) is not a separate identifier. Rather a synchronisation source identifier is listed as a CSRC in the RTP header of a packet generated by an RTP mixer if the corresponding SSRC was in the header of one of the packets that contributed to the mix.
It is not possible, in general, to extract media represented by an individual CSRC 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. The exception is the case when only a single CSRC is indicated as this represent forwarding of a media stream, possibly modified. The RTP header extension for Mixer-to-Client Audio Level Indication expands on the receivers information about a packet with a CSRC list. Due to these restrictions, CSRC will not be considered a fully qualified multiplexing point and will be disregarded in the rest of this document.
Each Media Stream utilises one or more RTP payload formats. An RTP payload format describes how the output of a particular media codec is framed and encoded into RTP packets. The payload format used is identified by the payload type field in the RTP data packet header. The combination therefore identifies a specific Media Stream encoding format. The format definition can be taken from [RFC3551] for statically allocated payload types, but ought to 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 signalling 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.
The payload type is scoped by sending endpoint within an RTP Session. All synchronisation sources sent from a single endpoint 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 RTP source's RTP timestamp time line, effectively time-multiplexing different Payload Types if any change occurs. The payload type used can change on a per-packet basis for an SSRC, for example a speech codec making use of generic comfort noise [RFC3389]. If there is a true need to send multiple Payload Types for the same SSRC that are valid for the same instant, 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 ought not exceed the transport MTU.
Other aspects of RTP payload format use are described in RTP Payload HowTo.
The payload type is not a multiplexing point at the RTP layer (see Appendix A for a detailed discussion of why using the payload type as an RTP multiplexing point does not work). The RTP payload type is, however, used to determine how to render a media stream, and so can be viewed as selecting a rendering context. The rendering context can be defined by the signalling, and the RTP payload type number is sometimes used to associate an RTP media stream with the signalling. This association is possible provided unique RTP payload type numbers are used in each context. For example, an RTP media stream can be associated with an SDP "m=" line by comparing the RTP payload type numbers used by the media stream with payload types signalled in the "a=rtpmap:" lines in the media sections of the SDP. If RTP media streams are being associated with signalling contexts based on the RTP payload type, then the assignment of RTP payload type numbers needs to be unique across signalling contexts; if the same RTP payload format configuration is used in multiple contexts, then a different RTP payload type number has to be assigned in each context to ensure uniqueness. If the RTP payload type number is not being used to associated RTP media streams with a signalling context, then the same RTP payload type number can be used to indicate the exact same RTP payload format configuration in multiple contexts. In case of bundled media, Section 10.2 of [I-D.ietf-mmusic-sdp-bundle-negotiation] provides more information on SDP signalling.
The impact of how RTP multiplexing is performed will in general vary with how the RTP Session participants are interconnected, described by RTP Topology.
Even the most basic use case, denoted Topo-Point-to-Point in [RFC7667], raises a number of considerations that are discussed in detail in following sections. They range over such aspects as:
For some Point to Multi-point topologies (e.g. Topo-ASM and Topo-SSM in [RFC7667]), multicast is used to interconnect the session participants. Special considerations (documented in Section 4.2.3) need to be made as multicast is a one to many distribution system.
Sometimes an RTP communication can end up in a situation when the peer it is communicating with is not compatible with the other peer for various reasons:
In many situations this is resolved by the inclusion of a translator between the two peers, as described by Topo-PtP-Translator in [RFC7667]. The translator's main purpose is to make the peer look to the other peer like something it is compatible with. There can also be other reasons than compatibility to insert a translator in the form of a middlebox or gateway, for example a need to monitor the media streams. If the stream transport characteristics are changed by the translator, appropriate media handling can require thorough understanding of the application logic, specifically any congestion control or media adaptation.
The point to point topology can contain one to many RTP sessions with one to many media sources per session, each having one or more RTP sources per media source.
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.
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 minimised, as described in the integrated layer processing design principle. 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:
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 might 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 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. As we saw in the discussion of RTP mixers, the RTP mixer has to embed application logic in order to handle streams anyway; the separation of streams according to stream type is just another piece of application logic, which might or might not be appropriate for a particular application. A type of application that can mix different media sources "blindly" is the audio only "telephone" bridge; most other type of application needs application-specific logic to perform the mix correctly.
The fifth argument discusses network aspects that we will discuss more below in Section 4.2. It also goes into aspects of implementation, like decomposed endpoints 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 to use different RTP sessions for media streams that don't share a media type. This document supports the first point; it is very valid. The later is one thing which needs to be further discussed, as imposing a single solution on all usages of RTP is inappropriate. Multiple Media Types in an RTP Session specification provides a detailed analysis of the potential issues in having multiple media types in the same RTP session. This document tries to provide an wider scoped consideration regarding the usage of RTP session and considers multiple media types in one RTP session as possible choice for the RTP application designer.
Using multiple SSRCs in an RTP session at one endpoint requires resolving some unclear aspects of the RTP specification. These could potentially lead to some interoperability issues as well as some potential significant inefficiencies. These are further discussed in "RTP Considerations for Endpoints Sending Multiple Media Streams". A application designer needs to consider these issues and the impact availability or lack of the optimization in the endpoints has on their application.
If an application will become affected by the issues described, using Multiple RTP sessions can mitigate these issues.
A common problem in a number of various RTP extensions has been how to bind related RTP sources and their media streams together. This issue is common to both using additional SSRCs and Multiple RTP sessions.
The solutions can be divided into some groups, RTP/RTCP based, Signalling based (SDP), grouping related RTP sessions, and grouping SSRCs within an RTP session. Most solutions are explicit, but some implicit methods have also been applied to the problem.
The SDP-based signalling solutions are:
This supports a lot of use cases. All these solutions have shortcomings 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.
Within RTP/RTCP based solutions when binding to an endpoint or synchronization context, i.e. the CNAME has not been sufficient and one way to bind related streams in multiple RTP sessions has been to use the same SSRC value across all the RTP sessions. RTP Retransmission is multiple RTP session mode, Generic FEC, as well as the RTP payload format for Scalable Video Coding in Multi Session Transmission (MST) mode uses this method. This method clearly works but might have some downside in RTP sessions with many participating SSRCs. The birthday paradox ensures that if you populate a single session with 9292 SSRCs at random, the chances are approximately 1% that at least one collision will occur. When a collision occur this will force one to change SSRC in all RTP sessions and thus resynchronizing all of them instead of only the single media stream having the collision. Therefore it is not recommended to use such method. Using [RFC7656] streams from the same media source should use the same RTP session.
It can be noted that Section 8.3 of the RTP Specification recommends using a single SSRC space across all RTP sessions for layered coding.
Another solution that has been applied to binding SSRCs has been an implicit method used by RTP Retransmission when doing retransmissions in the same RTP session as the source RTP media stream. This issues an RTP retransmission request, and then await a new SSRC carrying the RTP retransmission payload and where that SSRC is from the same CNAME. This limits a requestor to having only one outstanding request on any new source SSRCs per endpoint.
[I-D.ietf-mmusic-rid] provides an RTP/RTCP based mechanism capable of supporting explicit association within an RTP session.
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 losses 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 related flows as discussed above. Looking at the history of these schemes, there are schemes using multiple SSRCs and schemes using multiple RTP sessions, and some schemes that support both modes of operation.
Using multiple RTP sessions supports the case where some set of receivers might not be able to utilise 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 allows for flexibility. This is especially useful when receivers see very heterogeneous packet loss rates. Those receivers that are not seeing packet loss don't need to join the multicast group with the FEC data, and so avoid the overhead of receiving unnecessary FEC packets, for example.
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 have to be considered working with some legacy applications.
It is not uncommon that applications or services of similar usage, especially the ones intended for interactive communication, encounter a situation where one want to interconnect two or more of these applications.
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.
There are two fundamental approaches to gatewaying: RTP Translator interworking (RTP bridging), where the gateway acts as an RTP Translator, and the two applications are members of the same RTP session, and Gateway Interworking (with RTP termination), where there are independent RTP sessions running from each interconnected application to the gateway.
From an RTP perspective the RTP Translator approach could work if all the applications are using the same codecs with the same payload types, 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 might not always be true.
When one is gatewaying via an RTP Translator, a natural requirement is that the two applications being interconnected need to use the same approach to multiplexing. Furthermore, if one of the applications is capable of working in several modes (such as being able to use Additional SSRCs or Multiple RTP sessions at will), and the other one is not, successful interconnection depends on locking the more flexible application into the operating mode where interconnection can be successful, even if no participants using the less flexible application are present when the RTP sessions are being created.
When one terminates RTP sessions at the gateway, there are certain tasks that the gateway has to carry out:
For applications that uses any security mechanism, e.g. in the form of SRTP, then the gateway needs to be able to decrypt incoming packets and re-encrypt them in the other application's security context. This is necessary even if all that's needed is a simple remapping of SSRC numbers. If this is done, the gateway also needs to be a member of the security contexts of both sides, of course.
Other tasks a gateway might need to apply include transcoding (for incompatible codec types), rescaling (for incompatible video size requirements), suppression of content that is known not to be handled in the destination application, or the addition or removal of redundancy coding or scalability layers to fit the need of the destination domain.
From the above, we can see that the gateway needs to have an intimate knowledge of the application requirements; a gateway is by its nature application specific, not a commodity product.
This fact reveals 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 endpoints, 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 requires that each RTP session needs different master keys, as use of the same key in two RTP sessions for some ciphers can result in two-time pads that completely breaks the confidentiality of the packets.
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 endpoint 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 endpoint and the RTP mixer, meaning that the mixer functions as an RTP-terminating gateway.
When establishing RTP sessions that can contain endpoints 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:
This indicates that gateways attempting to interconnect to this class of devices has to make sure that only one media stream of each type gets delivered to the endpoint if it's expecting only one, and that the multiplexing format is what the device expects. It is highly unlikely that RTP translator-based interworking can be made to function successfully in such a context.
The multiplexing choice has impact on network level mechanisms that need to be considered by the implementer.
When it comes to Quality of Service mechanisms, they are either flow based or packet marking based. RSVP is an example of a flow based mechanism, while Diff-Serv is an example of a packet marking based one. For a packet 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. Additional SSRC 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.
It also needs to be noted that packet marking based QoS mechanisms can have limitations. A general observation is that different DSCP can be assigned to different packets within a flow as well as within an RTP Media Stream. However, care needs to be taken when considering which forwarding behaviours that are applied on path due to these DSCPs. In some cases the forwarding behaviour can result in packet reordering. For more discussion of this see [RFC7657].
More specific to the choice between using one or more RTP session can be the method for assigning marking to packets. If this is done using a network ingress function, it can have issues discriminating the different RTP media streams. The network API on the endpoint also needs to be capable of setting the marking on a per packet basis to reach the full functionality.
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 analyse and comment on the impact of requiring more underlying transport flows in the presence of NATs and Firewalls:
Additional SSRC keeps the additional media streams within one RTP Session and transport flow and does not introduce any additional NAT traversal complexities per media stream. This can be compared with normally one or two additional transport flows per RTP session when using multiple RTP sessions. Additional lower layer transport flows will be needed, unless an explicit de-multiplexing layer is added between RTP and the transport protocol. At time of writing no such mechanism was defined.
Multicast groups provides a powerful semantics for a number of real-time applications, especially the ones that desire broadcast-like behaviours with one endpoint 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.
In addition, the transport flow considerations in multicast are a bit different from unicast; NATs with port translation are not useful in the multicast environment, meaning that the entire port range of each multicast address is available for distinguishing between RTP sessions.
Thus it appears easiest and most straightforward to use multiple RTP sessions for sending different media flows used for adapting to network conditions. It is also common that streams that improve transport robustness are sent in their own multicast group to allow for interworking with legacy or to support different levels of protection.
Here are some common behaviours for RTP multicast:
All multicast configurations share a signalling requirement; 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 is to be noted that SDP offer/answer is not appropriate for ensuring this property. The signalling aspects of multicast are not explored further in this memo.
Security solutions for this type of group communications are also challenging. First of all the key-management and the security protocol needs to support group communication. Source authentication requires special solutions. For more discussion on this please review Options for Securing RTP Sessions.
When dealing with point-to-point, 2-member RTP sessions only, there are few security issues that are relevant to the choice of having one RTP session or multiple RTP sessions. However, there are a few aspects of multiparty sessions that might warrant consideration. For general information of possible methods of securing RTP, please review RTP Security Options.
When using SRTP the security context scope is important and can be a necessary differentiation in some applications. As SRTP's crypto suites (so far) are 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 issues.
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.
SRTP has no special functions for dealing with different sets of master keys for different SSRCs. The key-management functions have different capabilities to establish different set of keys, normally on a per endpoint basis. For example, DTLS-SRTP and Security Descriptions establish different keys for outgoing and incoming traffic from an endpoint. This key usage has to be written into the cryptographic context, possibly associated with different SSRCs.
Performing key-management for multi-party session can be a challenge. This section considers some of the issues.
Multi-party sessions, such as transport translator based sessions and multicast sessions, cannot use Security Description nor DTLS-SRTP without an extension as each endpoint provides its set of keys. In centralised conferences, the signalling counterpart is a conference server and the media plane unicast counterpart (to which DTLS messages would be sent) is the transport translator. Thus an extension like Encrypted Key Transport is needed or a MIKEY based solution that allows for keying all session participants with the same master key.
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 an RTP translator or mixer to rewrite an SSRC value in the RTP packet of an unencrypted session, the cost of doing it when using cryptographic security functions is higher. For example if using SRTP, the actual security context and exact crypto key are determined by the SSRC field value. If one changes it, the encryption and authentication tag needs to 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.
This section discusses some archetypes of how RTP multiplexing can be used in applications to achieve certain goals and a summary of their implications. For each archetype there is discussion of benefits and downsides.
In this archetype each endpoint 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 endpoints have one media stream each. If the application needs additional media flows between the endpoints, they will have to establish additional RTP sessions.
The Pros:
The Cons:
RTP applications that need to inter-work with legacy RTP applications, like most deployed VoIP and video conferencing solutions, 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, needs its own RTP session. In addition, the dynamic behaviour that can arise in multi-party applications can tax the signalling system and make timely media establishment more difficult.
In this archetype, each RTP session serves only a single media type. The RTP session can contain multiple media streams, either from a single endpoint or from multiple endpoints. This commonly creates a low number of RTP sessions, typically only one for audio and one for video, with a corresponding need for two listening ports when using RTP/RTCP multiplexing.
The Pros:
The Cons:
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 fate 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 prioritisation between media types. It handles multi-party session well, independently of multicast or centralised transport distribution, as additional sources can dynamically enter and leave the session.
In this archetype one goes one step further than in the above by using multiple RTP sessions also for a single media type, but still not as far as having a single SSRC per RTP session. 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 archetype are scalability over multicast, simulcast, need for extended QoS prioritisation of media streams due to their usage in the application, or the need for fine-grained signalling using today's tools.
The Pros:
The Cons:
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.
This archetype 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:
The Cons:
There are some clear relations between these archetypes. 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 needed.
The two other archetypes "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.
This section contains a number of recommendations for implementers or specification writers when it comes to handling multi-stream.
There are currently some issues that needs to be resolved before this document is ready to be published:
This document makes no request of IANA.
Note to RFC Editor: this section can be removed on publication as an RFC.
There is discussion of the security implications of choosing SSRC vs Multiple RTP session in Section 4.3.
| [RFC3550] | Schulzrinne, H., Casner, S., Frederick, R. and V. Jacobson, "RTP: A Transport Protocol for Real-Time Applications", STD 64, RFC 3550, DOI 10.17487/RFC3550, July 2003. |
| [RFC7656] | Lennox, J., Gross, K., Nandakumar, S., Salgueiro, G. and B. Burman, "A Taxonomy of Semantics and Mechanisms for Real-Time Transport Protocol (RTP) Sources", RFC 7656, DOI 10.17487/RFC7656, November 2015. |
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:
Signalling is not an architectural consideration for RTP itself, so this discussion has been moved to an appendix. However, it is hugely important for anyone building complete applications, so it is deserving of discussion.
The issues raised here need to be addressed in the WGs that deal with signalling; they cannot be addressed by tweaking, extending or profiling RTP.
There exist various signalling solutions for establishing RTP sessions. Many are SDP based, however SDP functionality is also dependent on the signalling protocols carrying the SDP. Where RTSP and SAP both use SDP in a declarative fashion, while SIP uses SDP with the additional definition of Offer/Answer. The impact on signalling and especially SDP needs to be considered as it can greatly affect how to deploy a certain multiplexing point choice.
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:
Some of these issues are clearly SDP's problem rather than RTP limitations. However, if the aim is to deploy an solution using additional SSRCs 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 additional SSRC only, a number of signalling extensions are needed to clarify that there are multiple sets of media streams with different properties and that they need in fact be kept different, since a single set will not satisfy the application's requirements.
For some parameters, such as resolution and framerate, a SSRC-linked mechanism has been proposed: [I-D.lennox-mmusic-sdp-source-selection].
SDP chose to use the m= line both to delineate an RTP session and to specify the top level of the MIME 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. This binding has to be loosened in order to use SDP to describe RTP sessions containing multiple MIME top level types.
There is an accepted WG item in the MMUSIC WG to define how multiple media lines describe a single underlying transport and thus it becomes possible in SDP to define one RTP session with media types having different MIME top level types.
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, an application might choose to take all incoming audio RTP streams, mix them and play them 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. The signalling used will have to identify the media streams affected by their RTP-level identifiers, which means that they have to be identified either by their session or by their SSRC + session.
In some applications, the receiver cannot utilise the media stream at all before it has received the signalling message describing the media stream and its usage. In other applications, there exists a default handling that is appropriate.
If all media streams in an RTP session are to be treated in the same way, identifying the session is enough. If SSRCs in a session are to be treated differently, signalling needs to identify both the session and the SSRC.
If this signalling affects how any RTP central node, like an RTP mixer or translator that selects, mixes or processes streams, treats the streams, the node will also need to receive the same signalling to know how to treat media streams with different usage in the right fashion.