<?xml version="1.0"encoding="US-ASCII"?>encoding="UTF-8"?> <!DOCTYPE rfc SYSTEM"rfc2629.dtd" [ <!ENTITY RFC2119 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2119.xml"> <!ENTITY RFC2198 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2198.xml"> <!ENTITY RFC2205 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2205.xml"> <!ENTITY RFC2474 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2474.xml"> <!ENTITY RFC4588 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.4588.xml"> <!ENTITY RFC5109 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.5109.xml"> <!ENTITY RFC3264 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.3264.xml"> <!ENTITY RFC2974 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.2974.xml"> <!ENTITY RFC3261 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.3261.xml"> <!ENTITY RFC3550 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.3550.xml"> <!ENTITY RFC3389 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.3389.xml"> <!ENTITY RFC3551 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.3551.xml"> <!ENTITY RFC3711 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.3711.xml"> <!ENTITY RFC3830 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.3830.xml"> <!ENTITY RFC4103 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.4103.xml"> <!ENTITY RFC4383 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.4383.xml"> <!ENTITY RFC4566 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.4566.xml"> <!ENTITY RFC4568 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.4568.xml"> <!ENTITY RFC4585 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.4585.xml"> <!ENTITY RFC5104 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.5104.xml"> <!ENTITY RFC5389 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.5389.xml"> <!ENTITY RFC5576 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.5576.xml"> <!ENTITY RFC5760 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.5760.xml"> <!ENTITY RFC5761 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.5761.xml"> <!ENTITY RFC5764 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.5764.xml"> <!ENTITY RFC5888 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.5888.xml"> <!ENTITY RFC6465 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.6465.xml"> <!ENTITY RFC7201 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.7201.xml"> <!ENTITY RFC7656 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.7656.xml"> <!ENTITY RFC7657 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.7657.xml"> <!ENTITY RFC7667 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.7667.xml"> <!ENTITY RFC7826 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.7826.xml"> <!ENTITY RFC7983 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.7983.xml"> <!ENTITY RFC8088 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.8088.xml"> <!ENTITY RFC8108 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.8108.xml"> <!ENTITY RFC8445 SYSTEM "http://xml.resource.org/public/rfc/bibxml/reference.RFC.8445.xml"> <!ENTITY I-D.ietf-mmusic-rid SYSTEM "https://xml2rfc.tools.ietf.org/public/rfc/bibxml3/reference.I-D.ietf-mmusic-rid.xml"> <!ENTITY I-D.ietf-mmusic-sdp-bundle-negotiation SYSTEM "https://xml2rfc.tools.ietf.org/public/rfc/bibxml3/reference.I-D.ietf-mmusic-sdp-bundle-negotiation.xml"> <!ENTITY I-D.ietf-avtcore-multi-media-rtp-session SYSTEM "https://xml2rfc.tools.ietf.org/public/rfc/bibxml3/reference.I-D.ietf-avtcore-multi-media-rtp-session.xml"> <!ENTITY I-D.ietf-perc-srtp-ekt-diet SYSTEM "https://xml2rfc.tools.ietf.org/public/rfc/bibxml3/reference.I-D.ietf-perc-srtp-ekt-diet.xml"> <!ENTITY I-D.ietf-avtext-rid SYSTEM "https://xml2rfc.tools.ietf.org/public/rfc/bibxml3/reference.I-D.ietf-avtext-rid.xml"> <!ENTITY I-D.ietf-perc-private-media-framework SYSTEM "https://xml2rfc.tools.ietf.org/public/rfc/bibxml3/reference.I-D.ietf-perc-private-media-framework.xml"> ]> <?rfc toc="yes"?> <?rfc tocompact="yes"?> <?rfc tocdepth="3"?> <?rfc tocindent="yes"?> <?rfc symrefs="yes"?> <?rfc sortrefs="yes"?> <?rfc comments="yes"?> <?rfc inline="yes"?> <?rfc compact="yes"?> <?rfc subcompact="no"?>"rfc2629-xhtml.ent"> <rfccategory="info"xmlns:xi="http://www.w3.org/2001/XInclude" docName="draft-ietf-avtcore-multiplex-guidelines-12" number="8872" ipr="trust200902"submissionType="IETF">submissionType="IETF" category="info" consensus="true" obsoletes="" updates="" xml:lang="en" tocInclude="true" tocDepth="3" symRefs="true" sortRefs="true" version="3"> <!-- xml2rfc v2v3 conversion 2.45.3 --> <front> <title abbrev="Guidelines for Multiplexing in RTP">Guidelines forusingUsing the Multiplexing Features of RTP to Support Multiple Media Streams</title> <seriesInfo name="RFC" value="8872"/> <author fullname="Magnus Westerlund" initials="M." surname="Westerlund"> <organization>Ericsson</organization> <address> <postal> <street>Torshamnsgatan 23</street><street>SE-164 80 Kista</street> <street>Sweden</street><code>164 80</code> <city>Kista</city> <country>Sweden</country> </postal><phone>+46 10 714 82 87</phone><email>magnus.westerlund@ericsson.com</email> </address> </author> <author fullname="Bo Burman" initials="B." surname="Burman"> <organization>Ericsson</organization> <address> <postal> <street>Gronlandsgatan 31</street><street>SE-164 60 Kista</street> <street>Sweden</street><code>164 60</code> <city>Kista</city> <country>Sweden</country> </postal> <email>bo.burman@ericsson.com</email> </address> </author> <author fullname="Colin Perkins" initials="C." surname="Perkins"> <organization>University of Glasgow</organization> <address> <postal><street>School<extaddr>School of ComputingScience</street> <street>Glasgow G12 8QQ</street> <street>United Kingdom</street>Science</extaddr> <city>Glasgow</city> <code>G12 8QQ</code> <country>United Kingdom</country> </postal> <email>csp@csperkins.org</email> </address> </author> <author fullname="Harald Tveit Alvestrand" initials="H." surname="Alvestrand"> <organization>Google</organization> <address> <postal> <street>Kungsbron 2</street><street>Stockholm 11122</street> <street>Sweden</street><city>Stockholm</city> <code>11122</code> <country>Sweden</country> </postal> <email>harald@alvestrand.no</email> </address> </author> <author fullname="Roni Even" initials="R." surname="Even"> <address> <email>ron.even.tlv@gmail.com</email> </address> </author> <dateday="16" month="June" year="2020"/>month="January" year="2021"/> <keyword>Simulcast</keyword> <abstract> <t>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 wideapplicability,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.</t> </abstract> </front> <middle> <sectionanchor="section-1" title="Introduction">anchor="sect-1" numbered="true" toc="default"> <name>Introduction</name> <t>The Real-time Transport Protocol (RTP) <xreftarget="RFC3550"/>target="RFC3550" format="default"/> 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. From the beginning, RTP wasfrom the beginningdesigned for multiple participants in a communication session. It supports many topology paradigms and usages, as defined in <xreftarget="RFC7667"/>.target="RFC7667" format="default"/>. RTP has several multiplexing points designed for differentpurposes. Thesepurposes; these points enable support of multiple RTP streams and switching between different encoding or packetizationoftechniques for the media. By using multiple RTP sessions, sets of RTP streams can be structured for efficient processing or identification. Thus, to meet an application's needs, an RTP application designer needs to understand howtobest to use the RTP session, the RTP stream identifier(SSRC),(synchronization source (SSRC)), and the RTP payloadtype to meet the application's needs.</t>type.</t> <t>Therehavehas been increased interest inmore advancedmore-advanced usage of RTP. For example, multiple RTP streams can be used when a single endpoint has multiple media sources (like multiple cameras or microphones)thatfrom which streams of media need to be sent simultaneously. Consequently, questions are raised regarding the most appropriate RTP usage. The limitations in some implementations, RTP/RTCP extensions, andsignallingsignaling have also been exposed. This document aims to clarify the usefulness of some functionalities in RTPwhichthat, hopefully, willhopefullyresult inmore completefuture implementationsin the future.</t>that are more complete.</t> <t>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 arising from a particular usage of the RTP multiplexing points.TheThis documentwill provideprovides some guidelines andrecommendrecommends against some usages as being unsuitable, in general or for particular purposes.</t><t>The<t>This document starts with some definitions and then goes intotheexisting RTP functionalities around multiplexing. Both the desiredbehaviourbehavior and the implications of a particularbehaviourbehavior depend on which topologies areused, whichused; therefore, this topic requires some consideration.This is followed by a discussion ofWe then discuss some choicesinregarding multiplexingbehaviourbehavior andtheir impacts.the impacts of those choices. Some designs of RTP usage are also discussed. Finally, some guidelines and examples are provided.</t> </section> <sectionanchor="section-2" title="Definitions"> <section anchor="section-2.1" title="Terminology">anchor="sect-2" numbered="true" toc="default"> <name>Definitions</name> <section anchor="sect-2.1" numbered="true" toc="default"> <name>Terminology</name> <t>The definitions inSection 3 of<xreftarget="RFC3550"/>target="RFC3550" sectionFormat="of" section="3"/> are referenced normatively.</t> <t>The taxonomy defined in <xreftarget="RFC7656"/>target="RFC7656" format="default"/> is referenced normatively.</t> <t>The following terms and abbreviations are used in this document:</t><t> <list hangIndent="3" style="hanging"> <t hangText="Multiparty:">A communication situation including<dl newline="true" spacing="normal"> <dt>Multi-party:</dt> <dd>Communication that includes multiple endpoints.<vspace blankLines="0"/>In this document,it"multi-party" will be used to refer tosituationsscenarios where more than two endpointscommunicate.</t> <t hangText="Multiplexing:">Thecommunicate.</dd> <dt>Multiplexing:</dt> <dd>An operationof takingthat takes multiple entities as input,<vspace blankLines="0"/>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.</t> <t hangText="RTP Receiver:">An Endpoint(demultiplexed) again.</dd> <dt>RTP Receiver:</dt> <dd>An endpoint orMiddleboxmiddlebox receiving RTP streams and RTCP messages. It uses at least one SSRC to send RTCP messages. An RTPReceiverreceiver may also be an RTPSender. </t> <t hangText="RTP Sender:">An Endpointsender.</dd> <dt>RTP Sender:</dt> <dd>An endpoint sending one or more RTPstreams,streams but also sending RTCPmessages. </t> <t hangText="RTPmessages.</dd> <dt>RTP SessionGroup:">OneGroup:</dt> <dd>One or more RTP sessions that are used together<vspace blankLines="0"/>to perform some function. Examplesareinclude multiple RTP sessions used to carry different layers of a layered encoding. In an RTP Session Group, CNAMEs are assumed to be valid across all RTPsessions,sessions and designatesynchronisationsynchronization contexts that can cross RTP sessions;i.e.i.e., SSRCs that map to a common CNAME can be assumed to have RTCP Sender Report (SR) timing information derived from a common clock such that they can besynchronisedsynchronized forplayout. </t> <t hangText="Signalling:">Theplayout.</dd> <dt>Signaling:</dt> <dd>The process of configuring endpoints to participate in<vspace blankLines="0"/>one or more RTPsessions.</t> </list> </t> <t>sessions.</dd> </dl> <aside><t> Note: The above definitions ofRTP Receiver"RTP receiver" andRTP Sender"RTP sender" are consistent with the usage in <xreftarget="RFC3550"/>. </t>target="RFC3550" format="default"/>. </t></aside> </section> <sectionanchor="section-2.2" title="Subjects Outanchor="sect-2.2" numbered="true" toc="default"> <name>Focus ofScope">This Document</name> <t>This document is focused on issues that affect RTP. Thus, issues that involvesignalling protocols,signaling protocols -- such as whether SIP <xreftarget="RFC3261"/>,target="RFC3261" format="default"/>, Jingle <xreftarget="JINGLE"/>target="JINGLE" format="default"/>, or some other protocol is in use for sessionconfiguration,configuration; the particular syntaxes used to define RTP sessionproperties,properties; or the constraints imposed by particular choices in thesignalling protocols,signaling protocols -- are mentioned only as examples in order to describe the RTP issues more precisely.</t> <t>This document assumes that the applications will use RTCP. While there are applications that don't send RTCP, they do not conform to the RTPspecification,specification and thus can be regarded as reusing the RTP packet format but not implementingthe RTP protocol.</t>RTP.</t> </section> </section> <sectionanchor="section-3" title="RTPanchor="sect-3" numbered="true" toc="default"> <name>RTP MultiplexingOverview">Overview</name> <sectionanchor="section-3.1" title="Reasonsanchor="sect-3.1" numbered="true" toc="default"> <name>Reasons for Multiplexing and Grouping RTPStreams">Streams</name> <t>There are several reasons why an endpoint might choose to send multiple media streams. In thebelow discussion,discussion below, please keep in mind that the reasons for having multiple RTP streams vary andincludeinclude, but are not limitedtoto, the following:</t><t> <list style="symbols"> <t>Multiple<ul spacing="normal"> <li>There might be multiple mediasources</t>sources.</li> <li> <t>Multiple RTP streams might be needed to represent one mediasourcesource, forinstance: <list style="symbols"> <t>Toexample: </t> <ul spacing="normal"> <li>To carry different layers ofana scalable encoding of a mediasource</t> <t>Alternativesource</li> <li>Alternative encodings during simulcast,for instanceusing different codecs for the same audiostream</t> <t>Alternativestream</li> <li>Alternative formats during simulcast,for instancemultiple resolutions of the same videostream</t> </list> </t> <t>Astream</li> </ul> </li> <li>A retransmission stream might repeat some parts of the content of another RTPstream</t> <t>Astream.</li> <li>A Forward Error Correction (FEC) stream might provide material that can be used to repair another RTPstream</t> </list> </t>stream.</li> </ul> <t>For each of these reasons, it is necessary to decideifwhether each additional RTP stream is sent within the same RTP session as the other RTPstreams,streams orifit is necessary to use additional RTP sessions to group the RTP streams.The choiceFor a combination of reasons, the suitable choice for onesituation,situation might not be thechoicesuitableinchoice for anothersituation or combination of reasons.situation. Theclearest understandingchoice isassociated witheasiest when multiplexing multiple media sources of the same media type. However, all reasons warrant discussion and clarificationonregarding how to deal with them. As the discussion below will show,in reality we cannot choosea singleone of SSRC or RTP session multiplexing solutions forsolution does not suit all purposes. Toutiliseutilize RTP well and as efficiently as possible, both are needed. The real issue isfinding the right guidance onknowing when to createadditionalmultiple RTP sessionsandversus whenadditionalto send multiple RTP streams inthe samea single RTPsession is the right choice.</t>session.</t> </section> <sectionanchor="section-3.2" title="RTPanchor="sect-3.2" numbered="true" toc="default"> <name>RTP MultiplexingPoints">Points</name> <t>This section describes the multiplexing points present intheRTPprotocolthat can be used to distinguish RTP streams and groups of RTP streams.Figure 1<xref target="ref-rtp-demultiplexing-process"/> outlines the process of demultiplexing incoming RTPstreamsstreams, startingalready at the socketwith one or more sockets representing the reception of one or more transport flows,e.g.e.g., based on the UDP destination port. It also demultiplexes RTP/RTCP from any other protocols, such asSTUNSession Traversal Utilities for NAT (STUN) <xreftarget="RFC5389"/>target="RFC5389" format="default"/> and DTLS-SRTP <xreftarget="RFC5764"/>target="RFC5764" format="default"/> on the same transport as described in <xreftarget="RFC7983"/>.target="RFC7983" format="default"/>. The Processing and Buffering (PB) stepof Figure 1in <xref target="ref-rtp-demultiplexing-process"/> terminatestheRTP/RTCPprotocoland prepares the RTP payload for input to the decoder.</t> <figureanchor="ref-rtp-demultiplexing-process" title="RTPanchor="ref-rtp-demultiplexing-process"> <name>RTP DemultiplexingProcess"> <artwork> <![CDATA[Process</name> <artwork name="" type="" align="left" alt=""><![CDATA[ | | | | | | packets +-- v v v | +------------+ | | Socket(s) | Transport Protocol Demultiplexing | +------------+ | || || RTP | RTP/ || |+-----> DTLS (SRTPKeying,keying, SCTP,etc)etc.) Session | RTCP || +------> STUN (multiplexed using same port) +-- || +-- || | ++(split by SSRC)-++---> Identify SSRC collision | || || || || | (associate withsignallingsignaling by MID/RID) | vv vv vv vv RTP | +--+ +--+ +--+ +--+ Jitter buffer, Streams | |PB| |PB| |PB| |PB| process RTCP, etc. | +--+ +--+ +--+ +--+ +-- | | | | (select decoder based onPT)payload type (PT)) +-- | / | / | +-----+ | / | / | |/ Payload | v v v Formats | +---+ +---+ +---+ | |Dec| |Dec| |Dec| Decoders | +---+ +---+ +---++-- ]]> </artwork>+--]]></artwork> </figure><t/><sectionanchor="section-3.2.1" title="RTP Session">anchor="sect-3.2.1" numbered="true" toc="default"> <name>RTP Session</name> <t>An RTP session is the highest semantic layer intheRTPprotocol,and represents an association between a group of communicating endpoints. RTP does not contain a session identifier, yet different RTP sessions must be possible to identify both across a set of different endpoints and from the perspective of a single endpoint.</t> <t>For RTP session separation across endpoints, the set of participants that form an RTP session is defined as those that share a singlesynchronisation sourceSSRC space <xreftarget="RFC3550"/>.target="RFC3550" format="default"/>. That is, if a group of participants are each aware of thesynchronisation sourceSSRC identifiers belonging to the other participants, then those participants are in a single RTP session. A participant can become aware ofa synchronisation sourcean SSRC identifier by receiving an RTP packet containingitthe identifier in the SSRC field orCSRCcontributing source (CSRC) list, by receiving an RTCP packetmentioninglisting it in an SSRC field, or throughsignallingsignaling (e.g., the Session Description Protocol (SDP) <xreftarget="RFC4566"/>target="RFC4566" format="default"/> "a=ssrc:" attribute <xreftarget="RFC5576"/>).target="RFC5576" format="default"/>). 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 thesignalling.</t>signaling.</t> <t>For RTP session separation within a singleendpointendpoint, RTP relies on the underlying transportlayer,layer andonthesignallingsignaling to identify RTP sessions in a manner that is meaningful to the application. A single endpoint can have one or more transport flows for the same RTP session, and a single RTP session can span multipletransport layertransport-layer flows even if all endpoints use a singletransport layertransport-layer flow per endpoint for that RTP session. Thesignallingsignaling layer might give RTP sessions an explicit identifier, or the 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 endpoint can identify and delimit an RTP session from other RTP sessions by their UDP source and destination IP addresses and their UDP port numbers. A single RTP session can be using multiple IP/UDP flows for receiving and/or sending RTP packets to other endpoints or middleboxes, even if the endpoint does not have multiple IP addresses. Using multiple IP addresses only makes it more likelyto requirethat multiple IP/UDPflows.flows will be required. Another example is SDP media descriptions (the "m=" line and thefollowingsubsequent associated lines) that signal the transport flow and RTP session configuration for the endpoint's part of the RTP session. The SDP grouping framework <xreftarget="RFC5888"/>target="RFC5888" format="default"/> allows labeling of the media descriptions to be used so that RTP Session Groups can be created. Through the use ofNegotiating<xref target="RFC8843">"Negotiating Media Multiplexing Using the Session Description Protocol(SDP) <xref target="I-D.ietf-mmusic-sdp-bundle-negotiation"/>,(SDP)"</xref>, multiple media descriptions become part of a common RTP session where each media description represents the RTP streams sent or received for a media source.</t><t>The RTP protocol<t>RTP makes no normative statements about the relationship between different RTPsessions, however thesessions; however, applications that use more than one RTP sessionwill have some higher layer understanding of the relationship betweenneed to understand how the different RTP sessions that theycreate.</t>create relate to one another.</t> </section> <sectionanchor="section-3.2.2" title="Synchronisationanchor="sect-3.2.2" numbered="true" toc="default"> <name>Synchronization Source(SSRC)"> <t>A synchronisation source (SSRC)(SSRC)</name> <t>An SSRC identifies a source of an RTP stream, or an RTP receiver when sending RTCP. Every endpoint has at least one SSRC identifier, even if it does not send RTP packets. RTP endpoints that are only RTP receivers still send RTCP and use their SSRC identifiers in the RTCP packets they send. An endpoint can have multiple SSRC identifiers if it sends multiple RTP streams. Endpoints thatarefunction as both RTP sender and RTP receiver use the same SSRC(s) in both roles.</t> <t>The SSRC is a 32-bit identifier. It is present in every RTP and RTCP packetheader,header and in the payload of some RTCP packet types. It can also be present in SDPsignalling.signaling. Unlesspre-signalled, e.g.presignaled, e.g., using the SDP "a=ssrc:" attribute <xreftarget="RFC5576"/>,target="RFC5576" format="default"/>, the SSRC is chosen at random. It is not dependent on the network address of theendpoint,endpoint and is intended to be unique within an RTP session. SSRC collisions canoccur,occur and are handled as specified in <xreftarget="RFC3550"/>target="RFC3550" format="default"/> and <xreftarget="RFC5576"/>,target="RFC5576" format="default"/>, resulting in the SSRC of the colliding RTP streams or receivers changing. An endpoint that changes its network transport address during a session has to choose a new SSRC identifier to avoid being interpreted as a looped source, unless a mechanism providing a virtual transport (such asICEInteractive Connectivity Establishment (ICE) <xreftarget="RFC8445"/>)target="RFC8445" format="default"/>) abstracts the changes.</t> <t>SSRC identifiers that belong to the samesynchronisationsynchronization context (i.e., that represent RTP streams that can besynchronisedsynchronized using information in RTCP SR packets) use identical CNAME chunks in corresponding RTCPSDESsource description (SDES) packets. SDPsignallingsignaling can also be used to provide explicit SSRC grouping <xreftarget="RFC5576"/>.</t>target="RFC5576" format="default"/>.</t> <t>In some cases, the same SSRC identifier value is used to relate streams in two different RTP sessions, such as in RTP retransmission <xreftarget="RFC4588"/>.target="RFC4588" format="default"/>. This is to beavoidedavoided, since there is no guarantee that SSRC values are unique across RTP sessions.ForIn the case of RTP retransmission <xreftarget="RFC4588"/> casetarget="RFC4588" format="default"/>, it is recommended to use explicit binding of the source RTP stream and the redundancy stream,e.g.e.g., using the RepairedRtpStreamId RTCP SDES item <xreftarget="I-D.ietf-avtext-rid"/>.target="RFC8852" format="default"/>. The RepairedRtpStreamId is a rather recent mechanism, so one cannot expect older applications to follow this recommendation. </t> <t>Note that the RTP sequence number and RTP timestamp are scoped by the SSRC and are thus specific per RTP stream.</t> <t>Different types of entities use an SSRC to identify themselves, as follows: </t><t> <list hangIndent="3" style="hanging"> <t hangText="A<ul spacing="normal"> <li>A real mediasource:">Usessource uses the SSRC to identify a "physical" mediasource.</t> <t hangText="Asource.</li> <li>A conceptual mediasource:">Usessource uses the SSRC to identify the result of applying some filtering function in a networknode,node -- forexampleexample, a filtering function in an RTP mixer that provides the most active speaker based on some criteria, or a mix representing a set of othersources.</t> <t hangText="Ansources.</li> <li>An RTPreceiver:">Usesreceiver uses the SSRC to identify itself as the source of its RTCPreports.</t> </list> </t>reports.</li> </ul> <t>An endpoint that generates more than one media type,e.g.e.g., a conference participant sending both audio and video, need not (and, indeed, should not) use the same SSRC value across RTP sessions. Using RTCP compound packets containing the CNAME SDES item is the designated methodto bindfor binding an SSRC to a CNAME, effectively cross-correlating SSRCs within and between RTPSessionssessions as coming from the same endpoint. The main property attributed to SSRCs associated with the same CNAME is that they are from a particularsynchronisationsynchronization context and can besynchronisedsynchronized at playback.</t> <t>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 its SSRC number due to an SSRC conflict.Use ofUsing theMIDmedia identification (MID) extension <xreftarget="I-D.ietf-mmusic-sdp-bundle-negotiation"/>target="RFC8843" format="default"/> helps to identify which media source the new SSRCrepresentsrepresents, anduse ofusing theRIDrestriction identifier (RID) extension <xreftarget="I-D.ietf-mmusic-rid"/>target="RFC8851" format="default"/> helps to identify what encoding or redundancy stream it represents, even though the SSRC changed. 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, making the new SSRC a new source.</t> </section> <sectionanchor="section-3.2.3" title="Contributinganchor="sect-3.2.3" numbered="true" toc="default"> <name>Contributing Source(CSRC)">(CSRC)</name> <t>TheContributing Source (CSRC)CSRC is not a separate identifier.RatherRather, an SSRC identifier is listed as a CSRC in the RTP header of a packet generated by an RTP mixer or videoMCU/switch,Multipoint Control Unit (MCU) / switch, if the corresponding SSRC was in the header of one of the packets that contributed to the output.</t> <t>It is not possible, in general, to extract media represented by an individualCSRCCSRC, since it is typically the result of a media merge(e.g.(e.g., mix) operation on the individual media streams corresponding to the CSRC identifiers. The exception is the casewhenwhere only a single CSRC isindicatedindicated, as thisrepresentrepresents the forwarding of an RTPstream, possiblystream that might have been modified. The RTP header extension (<xref target="RFC6465">"A Real-time Transport Protocol (RTP) Header Extension for Mixer-to-Client Audio LevelIndication <xref target="RFC6465"/>Indication"</xref>) expands on the receiver's information about a packet with a CSRC list. Due to these restrictions, a CSRC will not be considered a fully qualified multiplexing point and will be disregarded in the rest of this document.</t> </section> <sectionanchor="section-3.2.4" title="RTPanchor="sect-3.2.4" numbered="true" toc="default"> <name>RTP PayloadType">Type</name> <t>Each RTP streamutilisesutilizes 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 is identified by the payload type (PT) field in the RTP packet header. The combination of SSRC and PT therefore identifies a specific RTP stream in a specific encoding format. The format definition can be taken from <xreftarget="RFC3551"/>target="RFC3551" format="default"/> for statically allocated payloadtypes,types but ought to be explicitly defined insignalling,signaling, such as SDP,bothfor both static and dynamic payload types. The term "format" here includes those aspects described by out-of-bandsignallingsignaling 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 <xreftarget="RFC2198"/>.</t>target="RFC2198" format="default"/>.</t> <t>The RTP payload type is scoped by the sending endpoint within an RTP session. PT has the same meaning across all RTP streams in an RTP session. All SSRCs sent from a single endpoint share the same payload type definitions. The RTP payload type is designed such that only a single payload type is valid at anytimeinstant in time in the RTP stream's timestamptime line,timeline, effectively time-multiplexing different payload types if any change occurs. The payload type can change on a per-packet basis for anSSRC,SSRC -- forexampleexample, a speech codec making use of generic comfort noise <xreftarget="RFC3389"/>.target="RFC3389" format="default"/>. 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 <xreftarget="RFC2198"/>target="RFC2198" format="default"/> can be used. Several additionalconstraintsconstraints, other thanthe onesthose mentionedaboveabove, need to be met to enable thisuse,usage, one of which is that the combined payload sizes of the different payload types ought not exceed the transport MTU.</t> <t>Other aspects of using the RTP payload formatuseare described inHow<xref target="RFC8088">"How to Write an RTP PayloadFormat <xref target="RFC8088"/>.</t>Format"</xref>.</t> <t>The payload type is not a multiplexing point at the RTP layer (see <xreftarget="section-a"/>target="sect-a" format="default"/> 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 consume and decode an RTP stream. The RTP payload type number is sometimes used to associate an RTP stream with thesignalling,signaling, which in general requires that unique RTP payload type numbersarebe used in each context.Use ofUsing MID,e.g.e.g., when bundling "m=" sections <xreftarget="I-D.ietf-mmusic-sdp-bundle-negotiation"/>,target="RFC8843" format="default"/>, can replace the payload type assignalling associationa signaling association, and unique RTP payload types are then no longer required for that purpose.</t> </section> </section> <sectionanchor="section-3.3" title="Issuesanchor="sect-3.3" numbered="true" toc="default"> <name>Issues Related to RTPTopologies">Topologies</name> <t>The impact of how RTP multiplexing is performed will in general vary with how the RTP session participants are interconnected, as describedby RTP Topologyin <xreftarget="RFC7667"/>.</t>target="RFC7667">"RTP Topologies"</xref>.</t> <t>Even the most basic usecase, denoted Topo-Point-to-Pointcase -- "Topo-Point-to-Point" as described in <xreftarget="RFC7667"/>,target="RFC7667" format="default"/> -- raises a number ofconsiderations thatconsiderations, which are discussed in detail in the following sections. They range over such aspectsas:</t> <t> <list style="symbols"> <t>Doesas the following:</t> <ul spacing="normal"> <li>Does my communication peer support RTP as defined with multiple SSRCs per RTPsession?</t> <t>Dosession?</li> <li>Do I need network differentiation in the form of QoS (<xreftarget="section-4.2.1"/>)?</t> <t>Cantarget="sect-4.2.1" format="default"/>)?</li> <li>Can the application more easily process and handle the media streams if they are in different RTPsessions?</t> <t>Dosessions?</li> <li>Do I need to use additional RTP streams for RTP retransmission orFEC?</t> </list> </t>FEC?</li> </ul> <t>For somepoint to multi-pointpoint-to-multipoint topologies(e.g.(e.g., Topo-ASM and Topo-SSMin<xreftarget="RFC7667"/>),target="RFC7667" format="default"/>), multicast is used to interconnect the session participants. Special considerations (documented in <xreftarget="section-4.2.3"/>)target="sect-4.2.3" format="default"/>) are thenneededneeded, as multicast is a one-to-many distribution system.</t><t>Sometimes<t>Sometimes, an RTP communication session can end up in a situationwhenwhere the communicating peers are notcompatiblecompatible, for various reasons:</t><t> <list style="symbols"> <t>No<ul spacing="normal"> <li>No common media codec for a mediatypetype, thus requiringtranscoding.</t> <t>Differenttranscoding.</li> <li>Different support for multiple RTP streams and RTPsessions.</t> <t>Usagesessions.</li> <li>Usage of different media transportprotocols, i.e., RTP or other.</t> <t>Usage of different transport protocols,protocols (i.e., one peer uses RTP, but the other peer uses a different transport protocol).</li> <li>Usage of different transport protocols, e.g., UDP,DCCP,the Datagram Congestion Control Protocol (DCCP), orTCP.</t> <t>DifferentTCP.</li> <li>Different securitysolutions, e.g.,solutions (e.g., IPsec, TLS, DTLS, orSRTPthe Secure Real-time Transport Protocol (SRTP)) with different keyingmechanisms.</t> </list> </t> <t>In many situations this ismechanisms.</li> </ul> <t>These compatibility issues can often be resolved by the inclusion of a translator between the twopeers,peers -- the Topo-PtP-Translator, as describedby Topo-PtP-Translatorin <xreftarget="RFC7667"/>.target="RFC7667" format="default"/>. The translator's main purpose is to make the peers look compatible to each other. There can also beotherreasons other than compatibilityto insertfor inserting a translator in the form of a middlebox orgateway,gateway -- forexampleexample, a need to monitor the RTP streams. Beware that changing the stream transport characteristics in the translator can require a thorough understanding of aspects ranging from congestion control andmedia adaptationmedia-level adaptations to application-layer semantics.</t> <t>Within the uses enabled by the RTP standard, thepoint to pointpoint-to-point topology can contain one or more RTP sessions with one or more media sources per session, each having one or more RTP streams per media source.</t> </section> <sectionanchor="section-3.4" title="Issuesanchor="sect-3.4" numbered="true" toc="default"> <name>Issues Related to RTP andRTCP Protocol">RTCP</name> <t>Using multiple RTP streams is a well-supported feature of RTP. However, for most implementers or people writing RTP/RTCP applications or extensions attempting to apply multiple streams, it can be unclear when it is most appropriate to add an additional RTP stream in an existing RTP session and when it is better to use multiple RTP sessions. This section discusses the various considerationsneeded.</t>that need to be taken into account.</t> <sectionanchor="section-3.4.1" title="Theanchor="sect-3.4.1" numbered="true" toc="default"> <name>The RTPSpecification">Specification</name> <t>RFC 3550 contains some recommendations and abulletnumbered listwith 5(<xref target="RFC3550" sectionFormat="of" section="5.2"/>) of five argumentsforregarding different aspects of RTP multiplexing. Please reviewSection 5.2 of<xreftarget="RFC3550"/>.target="RFC3550" sectionFormat="of" section="5.2"/>. Five important aspects are quoted below.</t><t><list hangIndent="3" style="hanging"> <t hangText="1.">If,<ol spacing="normal" type="1"> <li><blockquote>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 changedencodings.</t></list> </t> <t>The firstencodings.</blockquote> <t>This argumentis toadvocates the use of differentSSRCSSRCs for each individual RTP stream,whichas this is fundamental to RTPoperation.</t> <t><list hangIndent="3" style="hanging"> <t hangText="2.">Anoperation.</t></li> <li><blockquote>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 packetloss.</t></list> </t> <t>The secondloss.</blockquote> <t>This argumentis advocatingadvocates against demultiplexing RTP streams within a session based only on their RTP payload typenumbers, whichnumbers; it stillstandsstands, as canbeenbe seen by the extensive list of issuesfounddiscussed in <xref target="sect-a"/>.</t></li> <!-- Note: "Section 6.4" is inAppendix A.</t> <t><list hangIndent="3" style="hanging"> <t hangText="3.">TheRFC 3550, so no xref --> <li><blockquote>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 typefield.</t></list> </t> <t>The thirdfield.</blockquote> <t>This argument is yet another argument against payload typemultiplexing.</t> <t><list hangIndent="3" style="hanging"> <t hangText="4.">Anmultiplexing.</t></li> <li><blockquote>An RTP mixer would not be able to combine interleaved streams of incompatible media into onestream.</t></list> </t> <t>The fourthstream.</blockquote> <t>This argumentisadvocates against multiplexing RTP packets that require different handling into the same session. In mostcasescases, the RTP mixer must embed application logic to handle streams; 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. One type of application that can mix different media sources blindly is the audio-only telephone bridge, although the ability to do that comes from the well-defined scenario that is aided by the use of a single media type, even though individual streams may use incompatible codec types; most other types of applications need application-specific logic to perform the mixcorrectly.</t> <t><list hangIndent="3" style="hanging"> <t hangText="5.">Carryingcorrectly.</t></li> <li><blockquote><t>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-processimplementations.</t></list></t> <t>The fifthimplementations.</t></blockquote> <t>This argument discusses network aspects that are described in <xreftarget="section-4.2"/>.target="sect-4.2" format="default"/>. It also goes into aspects of implementation, likeSplit Component Terminalsplit component terminals (seeSection 3.10 of<xreftarget="RFC7667"/>)target="RFC7667" sectionFormat="of" section="3.10"/>) -- endpoints where different processes orinter-connectedinterconnected devices handle different aspects of the wholemulti-media session.</t> <t>A summary ofmultimedia session.</t></li> </ol> <t>To summarize, RFC 3550's view on multiplexing is to use unique SSRCs for anything that is its own media/packetstream,stream andtouse different RTP sessions for media streams that don't share a media type. This document supports the first point; it is very valid. The latter needs further discussion, as imposing a single solution on all usages of RTP is inappropriate."Multiple Media<xref target="RFC8860">"Sending Multiple Types of Media inana Single RTPSession specification" <xref target="I-D.ietf-avtcore-multi-media-rtp-session"/>Session"</xref> updates RFC 3550 to allow multiple media types inaan RTPsession. It alsosession and provides a detailed analysis of the potential benefits and issuesinrelated to having multiple media types in the same RTP session. Thus,that document<xref target="RFC8860"/> provides a wider scope for an RTP session and considers multiple media types in one RTP session as a possible choice for the RTP application designer.</t> </section> <sectionanchor="section-3.4.2" title="Multipleanchor="sect-3.4.2" numbered="true" toc="default"> <name>Multiple SSRCs in aSession">Session</name> <t>Using multiple SSRCs at one endpoint in an RTP session requiresresolvingthat some unclear aspects of the RTPspecification.specification be resolved. These items could potentially lead to some interoperability issues as well as some potential significant inefficiencies, as further discussed in"RTP Considerations for Endpoints Sending"Sending MultipleMedia Streams"RTP Streams in a Single RTP Session" <xreftarget="RFC8108"/>.target="RFC8108" format="default"/>. An RTP application designer should consider these issues and the application's possibleapplicationimpactfromcaused by a lack of appropriate RTP handling or optimization in the peer endpoints.</t> <t>Using multiple RTP sessions can potentially mitigate application issues caused by multiple SSRCs in an RTP session.</t> </section> <sectionanchor="section-3.4.3" title="Bindinganchor="sect-3.4.3" numbered="true" toc="default"> <name>Binding RelatedSources">Sources</name> <t>A common problem in a number of various RTP extensions has been how to bind related RTP streams together. This issue is common to both using additional SSRCs and multiple RTP sessions.</t> <t>The solutions can be divided into a few groups:</t><t> <list style="symbols"> <t>RTP/RTCP based</t> <t>Signalling<ul spacing="normal"> <li>RTP/RTCP based</li> <li>Signaling based,e.g. SDP</t> <t>Groupinge.g., SDP</li> <li>Grouping related RTPsessions</t> <t>Groupingsessions</li> <li>Grouping SSRCs within an RTPsession</t> </list> </t>session</li> </ul> <t>Most solutions are explicit, but some implicit methods have also been applied to the problem.</t> <t>The SDP-basedsignallingsignaling solutions are:</t><t> <list hangIndent="3" style="hanging"> <t hangText="SDP Media Description Grouping:">The<dl newline="true" spacing="normal"> <dt>SDP media description grouping:</dt> <dd>The SDPGrouping Frameworkgrouping framework <xreftarget="RFC5888"/> <vspace blankLines="0"/>target="RFC5888" format="default"/> uses various semantics to group any number of media descriptions.ThisSDP media description grouping has primarily beengroupingused to group RTP sessions, but in combination with <xreftarget="I-D.ietf-mmusic-sdp-bundle-negotiation"/>target="RFC8843" format="default"/>, it can also group multiple media descriptions within a single RTPsession.</t> <t hangText="SDP Media Multiplexing:">Negotiatingsession.</dd> <dt>SDP media multiplexing:</dt> <dd><xref target="RFC8843">"Negotiating Media Multiplexing Using the Session Description Protocol(SDP) <xref target="I-D.ietf-mmusic-sdp-bundle-negotiation"/> <vspace blankLines="0"/>(SDP)"</xref> uses information taken from both SDP and RTCPinformationto associate RTP streams to SDP media descriptions. This allows both SDP and RTCP to group RTP streams belonging to an SDP mediadescription,description andtogroup multiple SDP media descriptions into a single RTPsession.</t> <t hangText="SDPsession.</dd> <dt>SDP SSRCgrouping:">Source-Specificgrouping:</dt> <dd><xref target="RFC5576">"Source-Specific Media Attributes inSDP <xref target="RFC5576"/> <vspace blankLines="0"/>the Session Description Protocol (SDP)"</xref> includes a solution for grouping SSRCs in the same wayasthat theGroupinggrouping framework groupsMedia Descriptions.</t> </list> </t>media descriptions.</dd> </dl> <t>The above grouping constructs support many use cases. Those solutions have shortcomings in cases where the session's dynamic properties are such that it is difficult or a drain on resources to keep the list of related SSRCs up to date.</t><t>An<t>One RTP/RTCP-based grouping solution is to use the RTCP SDES CNAME to bind related RTP streams to an endpoint ortoa synchronization context. For applications with a single RTP stream per type (media,sourcesource, or redundancy stream), the CNAME is sufficient for thatpurpose independentlypurpose, independent of whether one or more RTP sessions are used. However, some applications choose not to use a CNAME because of perceived complexity or a desire not to implement RTCP and instead use the same SSRC value to bind related RTP streams across multiple RTP sessions. RTPRetransmissionretransmission <xreftarget="RFC4588"/> intarget="RFC4588" format="default"/>, when configured to use multiple RTPsession modesessions, andGenericgeneric FEC <xreftarget="RFC5109"/>target="RFC5109" format="default"/> both use the CNAME method to relate the RTP streams, which may work but might have some downsides in RTP sessions with many participating SSRCs. It is not recommended to use identical SSRC values across RTP sessions to relate RTP streams;Whenwhen an SSRC collision occurs, this will force a change of that SSRC in all RTP sessions and will thus resynchronize all ofthemthe streams instead of only the single media streamhavingexperiencing the collision.</t> <t>Another methodtofor implicitlybindbinding SSRCs is used by RTPRetransmissionretransmission <xreftarget="RFC4588"/>target="RFC4588" format="default"/> when using the same RTP session as the source RTP stream for retransmissions.TheA receiver that is missing a packet issues an RTP retransmissionrequest,request and then awaits a new SSRC carrying the RTP retransmissionpayload andpayload, where that SSRC is from the same CNAME. This limits a requester to having only one outstanding retransmission request on any newsourceSSRCs per endpoint.</t><t>RTP<t><xref target="RFC8851">"RTP Payload FormatRestrictions <xref target="I-D.ietf-mmusic-rid"/>Restrictions"</xref> provides anRTP/RTCP basedRTP/RTCP-based mechanism to unambiguously identify the RTP streams within an RTP session and restrict the streams' payload format parameters in a codec-agnostic way beyond what is provided with the regular payload types. The mapping is done by specifying an "a=rid" value in the SDP offer/answersignallingsignaling and having the corresponding RtpStreamId value as an SDES item and an RTP headerextension.extension <xref target="RFC8852"/>. The RID solution also includes a solution for binding redundancy RTP streams to their original source RTP streams, given that those streams use RIDidentifiers.</t>identifiers. The redundancy stream uses the RepairedRtpStreamId SDES item and RTP header extension to declare the RtpStreamId value of the source stream to create the binding.</t> <t>Experience hasfoundshown that an explicit binding between the RTP streams, agnostic of SSRC values, behaves well. That way, solutions using multiple RTP streams in a single RTP session and in multiple RTP sessions will use the same type of binding.</t> </section> <sectionanchor="section-3.4.4" title="Forwardanchor="sect-3.4.4" numbered="true" toc="default"> <name>Forward ErrorCorrection">Correction</name> <t>There exist a number ofForward Error Correction (FEC) basedFEC-based schemesfor howdesigned to mitigate packet loss in the original streams. Most of the FEC schemes protect a single source flow.TheThis protection is achieved by transmitting a certain amount of redundant information that is encoded such that it can repair one or more instances of packetlossesloss over the set of packets the redundant information protects. This sequence of redundant information needs to be transmitted as its own mediastream, orstream or, in some cases, instead of the original media stream. Thus, many of these schemes create a need for binding relatedflowsflows, 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.</t> <t>Using multiple RTP sessions supports the case where some set of receivers might not be able toutiliseutilize the FEC information. By placing it in a separate RTP session and if separating RTP sessionsonat the transport level, FEC can easily be ignoredalready onat the transport level, without considering anyRTP layerRTP-layer information.</t> <t>In usages involving multicast,having thesending FEC informationon its ownin a separate multicast group allows for similar flexibility. This is especially useful when receivers see heterogeneous packet loss rates. A receiver can decide, based on measurement of experienced packet loss rates, whether to join a multicast group withthesuitable FEC data repair capabilities.</t> </section> </section> </section> <sectionanchor="section-4" title="Considerationsanchor="sect-4" numbered="true" toc="default"> <name>Considerations for RTPMultiplexing"> <section anchor="section-4.1" title="Interworking Considerations">Multiplexing</name> <section anchor="sect-4.1" numbered="true" toc="default"> <name>Interworking Considerations</name> <t>There are several different kinds of interworking, and this section discussestwo;two: interworking directly between differentapplications,applications and the interworking of applications through an RTPTranslator.translator. The discussion includes the implications of potentially different RTP multiplexing point choices and limitations that have to be considered when working with some legacy applications.</t> <sectionanchor="section-4.1.1" title="Application Interworking">anchor="sect-4.1.1" numbered="true" toc="default"> <name>Application Interworking</name> <t>It is not uncommon that applications or services of similar but not identical usage, especiallythe onesthose intended for interactive communication, encounter a situation where onewantwants to interconnect two or more of these applications.</t> <t>In these cases, one ends up in a situation where one might use a gateway to interconnect applications. This gateway must then either change the multiplexing structure or adhere to the respective limitations in each application.</t> <t>There are two fundamental approaches to building a gateway: using RTPTranslatortranslator interworking (RTP bridging), where the gateway acts as an RTPTranslatortranslator with the two interconnected applications being members of the same RTP session; or usingGateway Interworkinggateway interworking (<xref target="sect-4.1.3"/>) with RTP termination, where there are independent RTP sessions between each interconnected application and the gateway.</t> <t>For interworking to be feasible, any security solution in use needs to be compatible and capable of exchanging keys with either the peer or the gateway under theusedtrustmodel.model being used. Secondly, the applications need to use media streams in a way that makes sense in both applications. </t> </section> <sectionanchor="section-4.1.2" title="RTPanchor="sect-4.1.2" numbered="true" toc="default"> <name>RTP TranslatorInterworking">Interworking</name> <t>From an RTP perspective, the RTPTranslatortranslator approach could work if all the applications are using the same codecs with the same payload types, have made the same multiplexing choices, and have the same capabilitiesinregarding the number of simultaneous RTP streams combined with the same set of RTP/RTCP extensions being supported. Unfortunately, this might not always be true.</t> <t>When a gateway is implemented via an RTPTranslator,translator, an important consideration is if the two applications being interconnected need to use the same approach to multiplexing. If one side is using RTP session multiplexing and the other is using SSRC multiplexing with BUNDLE <xreftarget="I-D.ietf-mmusic-sdp-bundle-negotiation"/>,target="RFC8843" format="default"/>, it may be possible for the RTP translator to map the RTP streams between both sides using some method,e.g.e.g., based on the number and order of SDP "m=" lines from each side. There are also challengeswithrelated to SSRC collisionhandlinghandling, since, unless SSRC translation is applied on the RTP translator, there may be a collision on the SSRC multiplexing side that the RTP session multiplexing side will not be aware of. Furthermore, if one of the applications is capable of working in several modes (such as being able to use additional RTP streams in one RTP session or multiple RTP sessions atwill),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 none of the participants are using the less flexible application when the RTP sessions are being created.</t> </section> <sectionanchor="section-4.1.3" title="Gateway Interworking">anchor="sect-4.1.3" numbered="true" toc="default"> <name>Gateway Interworking</name> <t>When one terminates RTP sessions at the gateway, there are certain tasks that the gateway has to carry out:</t><t> <list style="symbols"> <t>Generating<ul spacing="normal"> <li>Generating appropriate RTCP reports for all RTP streams (possibly based on incoming RTCPreports),reports) originating from SSRCs controlled by thegateway.</t> <t>Handlinggateway.</li> <li>Handling SSRC collision resolution in each application's RTPsessions.</t> <t>Signalling,sessions.</li> <li>Signaling, choosing, and policing appropriatebit-ratesbitrates for eachsession.</t> </list> </t>session.</li> </ul> <t>For applications that use any security mechanism, e.g., in the form of SRTP, the gateway needs to be able to decrypt and verify source integrity of the incomingpackets,packets and then re-encrypt, integrity protect, and sign the packets as the peer 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 bothsides,sides and thus a trusted entity.</t> <t>The gateway might also need to apply transcoding (for incompatible codec types), media-level adaptations that cannot be solved through media negotiation (such as 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 needs of the destination domain.</t> <t>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 applicationspecific,specific and not a commodity product.</t> <t>These gateways might therefore potentially block application evolution by blocking RTP and RTCP extensions that the applications have been extended with but that are unknown to the gateway.</t> <t>If one uses a securitymechanism,mechanism like SRTP, the gateway and the necessary trust in it by the peersispose an additional risk tothecommunication security. The gateway alsoincurincurs additional complexities in the form of the decrypt-encrypt cycles needed for each forwarded packet. SRTP, due to its keying structure, also requires that each RTP sessionneedsneed different master keys, as the use of the same key in two RTP sessionscancan, for someciphersciphers, result in a reuse of a one-time pad that completely breaks the confidentiality of the packets.</t> </section> <sectionanchor="section-4.1.4" title="Multiple SSRC Legacy Considerations">anchor="sect-4.1.4" numbered="true" toc="default"> <name>Legacy Considerations for Multiple SSRCs</name> <t>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 or video). Even in conferencing applications, especially voice-only, the conference focus or bridgehas provided a single streamprovides to each participant a single stream containing a mix of the other participants. 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.</t> <t>Applications and systems that aren't updated to handle multiple streams following these recommendations can have issues with participating in RTP sessions containing multiple SSRCs within a single session, such as:</t><t> <list style="numbers"> <t>Need<ol spacing="normal" type="1"> <li>The need to handle more than one stream simultaneously rather than replacing analready existingalready-existing stream with a newone.</t> <t>Beone.</li> <li>Being capable of decoding multiple streamssimultaneously.</t> <t>Besimultaneously.</li> <li>Being capable of rendering multiple streamssimultaneously.</t> </list> </t>simultaneously.</li> </ol> <t>This indicates that gateways attempting to interconnect to this class of devices have to make sure that only one RTP stream of each media type gets delivered to the endpoint if it's expecting onlyone,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.</t> </section> </section> <sectionanchor="section-4.2" title="Network Considerations">anchor="sect-4.2" numbered="true" toc="default"> <name>Network Considerations</name> <t>The RTP implementer needs to consider that the RTP multiplexing choice also impactsnetwork levelnetwork-level mechanisms.</t> <sectionanchor="section-4.2.1" title="Quality of Service"> <t>Qualityanchor="sect-4.2.1" numbered="true" toc="default"> <name>Quality ofServiceService</name> <t>QoS mechanisms are either flow based or packet marking based. RSVP <xreftarget="RFC2205"/>target="RFC2205" format="default"/> is an example of aflow basedflow-based mechanism, whileDiff-ServDiffserv <xreftarget="RFC2474"/>target="RFC2474" format="default"/> is an example of apacket marking based one.</t>packet-marking-based mechanism.</t> <t>For aflow basedflow-based scheme, additionalSSRCSSRCs will receive the same QoS as all other RTP streams being part of the same 5-tuple (protocol, source address, destination address, source port, destination port), which is the most common selector forflow basedflow-based QoS.</t> <t>For apacket marking basedpacket-marking-based scheme, the method of multiplexing will not affect the possibilityto useof using QoS. Different Differentiated Services Code Points(DSCP)(DSCPs) can be assigned to different packets within a transport flow(5-Tuple)(5-tuple) as well as within an RTP stream, assuming the usage of UDP or other transportprotocolprotocols that do not have issues with packet reordering within the transport flow (5-tuple). To avoidpacket reordingpacket-reordering issues, packets belonging to the same RTP flow shouldlimits itslimit their use ofDSCPDSCPs tothosepackets whose correspondingPer HopPer-Hop Behavior (PHB)thatdo not enable reordering. If the transport protocol being used assumesin orderin&nbhy;order delivery ofpacket, such aspackets (e.g., TCP andSCTP,the Stream Control Transmission Protocol (SCTP)), then a single DSCP should be used. For more discussionofon this topic, see <xreftarget="RFC7657"/>.</t>target="RFC7657" format="default"/>.</t> <t>The method for assigning marking to packets can impact what number of RTP sessions to choose. If this marking is done using a network ingress function, it can have issues discriminating the different RTP streams. The network API on the endpoint also needs to be capable of setting the marking on a per-packet basis to reachthefull functionality.</t> </section> <sectionanchor="section-4.2.2" title="NATanchor="sect-4.2.2" numbered="true" toc="default"> <name>NAT and FirewallTraversal">Traversal</name> <t>In today'snetworksnetworks, there exist a large number of middleboxes.The onesThose that normally have the most impact on RTP are Network Address Translators(NAT)(NATs) and Firewalls(FW).</t> <t>Below(FWs).</t> <t>Below, weanalyseanalyze and comment on the impact of requiring more underlying transport flows in the presence of NATs andFirewalls:</t> <t> <list hangIndent="3" style="hanging"> <t hangText="End-PointFWs:</t> <dl newline="true" spacing="normal"> <dt>Endpoint PortConsumption:">AConsumption:</dt> <dd>A given IP address only has 65536<vspace blankLines="0"/>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 a large number of simultaneous streams. However, if the application uses ICE to authenticate STUN requests, a server can serve multiple endpoints from the same localport,port and use the whole 5-tuple (source and destination address, source and destination port, protocol) as the identifier of flows after having securely bound them to the remote endpoint address using the STUN request. In theory, the minimum number of media server ports neededareis the maximum number of simultaneous RTP sessions a single endpoint can use. In practice,implementationimplementations will probably benefit from using more server ports to simplify implementation or avoid performancebottlenecks.</t> <t hangText="NAT State:">Ifbottlenecks.</dd> <dt>NAT State:</dt> <dd>If an endpoint sits behind a NAT, each flow it generates<vspace blankLines="0"/>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 SmallOffice/HomeOffice&wj;/Home Office (SOHO) NATs,memory or processing are usuallythe most limitedresources.resource is memory or processing. Forlarge scalelarge-scale NATs serving many internal endpoints, available external ports are likely the scarce resource. Port limitationsisare primarily a problem for largercentralisedcentralized NATs whereendpoint independentendpoint-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, as a comparison, a real-time video conference session with audio and video likely uses less than 10 UDP flows, compared to certain web applications that can use 100+ TCP flows to various servers from a single browserinstance.</t> <t hangText="NAT Traversal Extra Delay:">Performinginstance.</dd> <dt>Extra Delay Added by NAT Traversal:</dt> <dd>Performing the NAT/FW traversal takes a<vspace blankLines="0"/>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.Thebest casebest-case scenario for additional NAT/FW traversal time after finding the first valid candidate pair following the specified ICE procedures is 1.5*RTT + Ta*(Additional_Flows-1), where Ta is the pacing timer. That assumes a message in one direction, immediately followed by acheck back. The reason itreturn message in the opposite direction to confirm reachability. It isn't more,is thatbecause ICE first finds one candidate pair thatworksworks, prior to attempting to establish multiple flows. Thus, there is no extra time until one has found a working candidate pair. Based on that working pair, the extra time is needed toin parallelestablishthe,the additional flows (two or three, in mostcases 2-3, additional flows.cases) in parallel. However, packet loss causes extradelays,delays of at least 500ms, which is thems (the minimal retransmission timer forICE.</t> <t hangText="NATICE).</dd> <dt>NAT Traversal FailureRate:">DueRate:</dt> <dd>Due to the need to establish more than a<vspace blankLines="0"/>single flow through the NAT, there is some risk that establishing the first flowsucceedswill succeed butthatone or more of the additional flows will fail. The riskthatof thishappenshappening is hard toquantify,quantify butought toshould be fairlylowlow, as one flow from the same interfaces has just been successfully established.ThusThus, only such rare eventssuchas NAT resource overload,orselecting particular port numbers that arefilteredfiltered, etc., ought to be reasons forfailure.</t> <t hangText="Deepfailure.</dd> <dt>Deep Packet Inspection and MultipleStreams:">FirewallsStreams:</dt> <dd>FWs differ in how<vspace blankLines="0"/>deeply they inspect packets.Due to all previous issues with firewallPrevious experience using FWs and SessionBoarderBorder Gateways(SBG)(SBGs) with RTPtransport media e.g. in Voice over IP (VoIP) systems,shows that thereexistsis a significant risk thatdeeply inspecting firewallsthe FWs and SBGs willhave similar legacy issues with multiple SSRCs as somereject RTPstack implementations.</t> </list> </t>sessions that use multiple SSRCs.</dd> </dl> <t>Using additional RTP streams in the same RTP session and transport flow does not introduce any additional NAT traversal complexities per RTP stream. This can be compared withnormally(normally) one or two additional transport flows per RTP session when using multiple RTP sessions. Additionallower layerlower-layer transport flows will be needed, unless an explicitde-multiplexingdemultiplexing layer is added between RTP and the transport protocol. At the time ofwritingthis writing, no such mechanism was defined.</t> </section> <sectionanchor="section-4.2.3" title="Multicast">anchor="sect-4.2.3" numbered="true" toc="default"> <name>Multicast</name> <t>Multicast groupsprovidesprovide a powerful tool for a number of real-time applications, especiallythe onesthose that desire broadcast-likebehavioursbehaviors with one endpoint transmitting to a large number of receivers, like in IPTV.There is also theAn RTP/RTCP extension to better supportSource SpecificSource-Specific Multicast (SSM) <xreftarget="RFC5760"/>.target="RFC5760" format="default"/> is also available. Many-to-many communication, which RTP <xreftarget="RFC3550"/>target="RFC3550" format="default"/> was originally built to support, has several limitations in common with multicast.</t> <t>One limitation is that, for any group,sender side adaptationsender-side adaptations with the intent to suit all receivers would have to adapt to the most limited receiver experiencing the worst conditions among the group participants, which imposes degradation for all participants. For broadcast-type applications with a large number of receivers, this is not acceptable. Instead, various receiver-based solutions are employed to ensure that the receivers achieve the 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 layeronin a different multicast group, one RTP session per multicast group is used.</t> <t>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.</t> <t>Thus, when using broadcast applications 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 improving transport robustness are sent in their own multicast group to allow for interworking with legacy applications or to support different levels of protection.</t> <t>Many-to-many applications have differentneedsneeds, and the most appropriate multiplexing choice will depend on how the actual application is realized. Multicast applications that are capable of usingsender sidesender-side congestion control can avoid the use of multiple multicast sessions and RTP sessions that result from the use ofreceiver sidereceiver-side congestion control.</t> <t>The properties of a broadcast application using RTPmulticast:</t> <t> <list style="numbers"> <t>Usesmulticast are as follows:</t> <ol spacing="normal" type="1"> <li>The application uses a group of RTPsessions,sessions -- not just one. Each endpoint will need to be a member of a number of RTP sessions in order to performwell.</t> <t>Withinwell.</li> <li>Within each RTP session, the number of RTP receivers is likely to be much larger than the number of RTPsenders.</t> <t>The applications need signallingsenders.</li> <li>The application needs signaling functions to identify the relationships between RTPsessions.</t> <t>The applications need signallingsessions.</li> <li>The application needs signaling or RTP/RTCP functions to identify the relationships between SSRCs in different RTP sessions whenneeds beyondmore complex relations than those that can be expressed by the CNAMEexist.</t> </list> </t>exist.</li> </ol> <t>Both broadcast and many-to-many multicast applications share asignallingsignaling requirement; all of the participants need the same RTP and payload type configuration. Otherwise, Acouldcould, forexampleexample, be using payload type 97 as the video codec H.264 while B thinks it is MPEG-2. SDP offer/answer <xreftarget="RFC3264"/>target="RFC3264" format="default"/> is not appropriate for ensuring this property in a broadcast/multicast context. Thesignallingsignaling aspects of broadcast/multicast are not explored further in this memo.</t> <t>Security solutions for this type of group communication are also challenging. First, the key-management mechanism and the security protocol need to support group communication. Second, source authentication requires special solutions. For more discussion on this topic, please reviewOptions<xref target="RFC7201">"Options for Securing RTPSessions <xref target="RFC7201"/>.</t>Sessions"</xref>.</t> </section> </section> <sectionanchor="section-4.3" title="Securityanchor="sect-4.3" numbered="true" toc="default"> <name>Security andKey Management Considerations">Key-Management Considerations</name> <t>When dealing withpoint-to-point, 2-memberpoint-to-point two-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 ofmultipartymulti-party sessions that might warrant consideration. For general informationofregarding possible methods of securing RTP, please reviewRTP Security Options<xref target="RFC7201"/>.</t> <sectionanchor="section-4.3.1" title="Securityanchor="sect-4.3.1" numbered="true" toc="default"> <name>Security ContextScope">Scope</name> <t>When using SRTP <xreftarget="RFC3711"/>,target="RFC3711" format="default"/>, the security context scope is important and can be a necessary differentiation in some applications. As SRTP's crypto suites are (so far) built around symmetric keys, the receiver will need to have the same key as the sender.This results in thatAs a result, no one in a multi-party session can be certain that a received packetreallywas really sent by the claimed sender and not by another party having access to the key. The single SRTP algorithm not having thisproperyproperty isthe TESLATimed Efficient Stream Loss-Tolerant Authentication (TESLA) source authentication <xreftarget="RFC4383"/>.target="RFC4383" format="default"/>. However, TESLA adds delay to achieve source authentication. In most cases, symmetric ciphers provide sufficient securitypropertiesproperties, butcreate issuesin a fewcases.</t>cases they can create issues.</t> <t>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 RTP streams. This requires that everyonere-keysrekey without disclosing the new keys to the excluded party.</t> <t>A second case is whenusingsecurity is used as an enforcing mechanism for stream access differentiation between different receivers.TakeTake, forexampleexample, a scalable layer or ahigh qualityhigh-quality simulcast version that only users paying a premium are allowed to access. The mechanism preventing a receiver from getting thehigh qualityhigh-quality stream can be based on the stream being encrypted with a key thatuserusers can't access without paying a premium, using the key-management mechanism to limit access to the key.</t><t>SRTP <xref target="RFC3711"/> as<t>As specified in <xref target="RFC3711" format="default"/>, SRTP usesper SSRCuniquekeys, howeverkeys per SSRC; however, the original assumption was asingle sessionsingle-session master key from whichSSRC specificSSRC-specific RTP and RTCP keyswherewere derived. However, that assumption was proven incorrect, as the application usage and the developedkey-mamangementkey-management mechanisms have chosen many different methods for ensuringSSRCuniquekeys.keys per SSRC. The key-management functions have differentcapabilitiesabilities to establish different sets of keys, normally on a per-endpoint basis. For example, DTLS-SRTP <xreftarget="RFC5764"/>target="RFC5764" format="default"/> and Security Descriptions <xreftarget="RFC4568"/>target="RFC4568" format="default"/> 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. Thus, limitations doexistexist, depending on the chosen key-management method and due to the integration of particular implementations of the key-management method and SRTP.</t> </section> <sectionanchor="section-4.3.2" title="Keyanchor="sect-4.3.2" numbered="true" toc="default"> <name>Key Management for Multi-partySessions">Sessions</name> <t>The capabilities of the key-management method combined with the RTP multiplexing choicesaffectsaffect the resulting security properties, control over the secured media, and whohavehas access to it.</t> <t>Multi-party sessions contain at least one RTP stream from each active participant. Depending on the multi-party topology <xreftarget="RFC7667"/>,target="RFC7667" format="default"/>, each participant can both send and receive multiple RTP streams. Transport translator-based sessions (Topo-Trn-Translator) and multicast sessions(Topo-ASM),(Topo-ASM) canneitheruse neither SecurityDescriptionDescriptions <xreftarget="RFC4568"/>target="RFC4568" format="default"/> nor DTLS-SRTP <xreftarget="RFC5764"/>target="RFC5764" format="default"/> without anextension asextension, because each endpoint provides its own set of keys. Incentralisedcentralized conferences, thesignallingsignaling counterpart is a conference server, and the transport translator is themedia planemedia-plane unicast counterpart (to which DTLS messages would be sent). Thus, an extension like Encrypted Key Transport <xreftarget="I-D.ietf-perc-srtp-ekt-diet"/>target="RFC8870" format="default"/> or aMIKEY <xref target="RFC3830"/> basedsolution based on Multimedia Internet KEYing (MIKEY) <xref target="RFC3830" format="default"/> that allows for keying all session participants with the same master key is needed.</t><t>Privacy Enchanced<t>Privacy-Enhanced RTP Conferencing (PERC) also enables a different trust model with semi-trustedmedia switchingmedia-switching RTP middleboxes <xreftarget="I-D.ietf-perc-private-media-framework"/>.</t>target="RFC8871" format="default"/>.</t> </section> <sectionanchor="section-4.3.3" title="Complexity Implications"> <t>The usage of security functionsanchor="sect-4.3.3" numbered="true" toc="default"> <name>Complexity Implications</name> <t>There cansurface complexity implications frombe complex interactions between the choice of multiplexing andtopology.topology and the security functions. This becomes especially evident in RTP topologies having any type of middlebox that processes or modifies RTP/RTCP packets. Whilethere is very smallthe overheadforof an RTP translator or mixerto rewriterewriting an SSRC value in the RTP packet of an unencryptedsession,session is low, the cost is higher when using cryptographic security functions. For example, if using SRTP <xreftarget="RFC3711"/>,target="RFC3711" format="default"/>, the actual security context and exact crypto key are determined by the SSRC field value. If one changesSSRC,the SSRC value, the encryption and authentication must use 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.</t> </section> </section> </section> <sectionanchor="section-5" title="RTPanchor="sect-5" numbered="true" toc="default"> <name>RTP Multiplexing DesignChoices">Choices</name> <t>This section discusses how some RTP multiplexing design choices can be used in applications to achieve certaingoals,goals anda summary ofsummarizes the implications of such choices.For each design there is discussion ofThe benefits anddownsides.</t>downsides of each design are also discussed.</t> <sectionanchor="section-5.1" title="Multipleanchor="sect-5.1" numbered="true" toc="default"> <name>Multiple Media Types in OneSession">Session</name> <t>This design uses a single RTP session for multiple different media types, like audio and video, and possibly also transport robustness mechanisms like FEC or retransmission. An endpoint can send zero,oneone, ormoremultiple media sources per media type, resulting in a number of RTP streams of various media types for both source and redundancy streams.</t><t>The Advantages:</t> <t> <list style="numbers"><t>Advantages:</t> <ol spacing="normal" type="1"> <li> <t>Only a single RTP session is used, whichimplies:<list style="symbols"> <t>Minimalimplies:</t> <ul spacing="normal"> <li>Minimal need to keep NAT/FWstate.</t> <t>Minimal NAT/FW-traversal cost.</t> <t>Fate-sharingstate.</li> <li>Minimal NAT/FW traversal cost.</li> <li>Fate-sharing for all mediaflows.</t> <t>Minimalflows.</li> <li>Minimal overhead for security associationestablishment.</t> </list> </t> <t>Dynamicestablishment.</li> </ul> </li> <li>Dynamic allocation of RTP streams can be handled almost entirely at the RTP level.How localizedThe extent to which this allocation can be kepttoat the RTP level depends on the application's needs for an explicit indication ofthestream usage and in how timely a fashion that information can besignalled.</t> </list> </t> <t>The Disadvantages:</t> <t> <list style="letters"> <t>Itsignaled.</li> </ol> <t>Disadvantages:</t> <ol spacing="normal" type="1"> <li>It is less suitable for interworking with other applications that use individual RTP sessions per media type or multiple sessions for a single media type, due to the risk of SSRCcollisioncollisions and thus a potential need for SSRCtranslation.</t> <t>Negotiationtranslation.</li> <li>Negotiation of individual bandwidths for the different media types is currently only possible in SDP when using RID <xreftarget="I-D.ietf-mmusic-rid"/>.</t> <t>Ittarget="RFC8851" format="default"/>.</li> <li>It is not suitable forSplit Component Terminalsplit component terminals (seeSection 3.10 of<xreftarget="RFC7667"/>).</t> <t>Flow-basedtarget="RFC7667" sectionFormat="of" section="3.10"/>).</li> <li>Flow-based QoS cannot be used to provide separate treatment of RTP streams compared to others in the single RTPsession.</t> <t>Ifsession.</li> <li>If there is significant asymmetry between the RTP streams' RTCP reporting needs, there are some challengesinrelated to configuration and usage to avoid wasting RTCP reporting on the RTP stream that does not needthatsuch frequentreporting.</t> <t>Itreporting.</li> <li>It is not suitable for applications where some receivers like to receive only a subset of the RTP streams, especially if multicast or a transport translator is beingused.</t> <t>There isused.</li> <li>There are some additionalconcern withconcerns regarding legacy implementations that do not support the RTP specification fully when it comes to handling multipleSSRCSSRCs per endpoint, as multiple simultaneous media types are sent as separateSSRCSSRCs in the same RTPsession.</t> <t>Ifsession.</li> <li>If the applications need finer control over which session participants are included in different sets of security associations, most key-management mechanisms will have difficulties establishing such asession.</t> </list> </t>session.</li> </ol> </section> <sectionanchor="section-5.2" title="Multipleanchor="sect-5.2" numbered="true" toc="default"> <name>Multiple SSRCs of the Same MediaType">Type</name> <t>In this design, each RTP session serves only a single media type. The RTP session can contain multiple RTP streams,eitherfrom either a single endpoint orfrommultiple 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 <xreftarget="RFC5761"/>.</t> <t>The Advantages</t> <t> <list style="numbers"> <t>Ittarget="RFC5761" format="default"/>.</t> <t>Advantages:</t> <ol spacing="normal" type="1"> <li>It works well withSplit Component Terminalsplit component terminals (seeSection 3.10 of<xreftarget="RFC7667"/>)target="RFC7667" sectionFormat="of" section="3.10"/>) where the split is per mediatype.</t> <t>Ittype.</li> <li>It enables flow-based QoS with differentprioritisationprioritization levels between mediatypes.</t> <t>Fortypes.</li> <li>For applications with dynamic usage of RTPstreams, i.e.streams (i.e., streams are frequently added andremoved,removed), having much of the state associated with the RTP session rather than per individual SSRC can avoid the need for in-sessionsignallingsignaling of meta-information about each SSRC. Inthesimplecasescases, this allows forunsignalledunsignaled RTP streams wheresession levelsession-level information and an RTCP SDES item(e.g.(e.g., CNAME) aresuffient.sufficient. In the more complex cases where more source-specific metadata needs to besignalledsignaled, the SSRC can be associated with an intermediate identifier,e.g.e.g., the MID conveyed as an SDES item as defined inSection 15 of<xreftarget="I-D.ietf-mmusic-sdp-bundle-negotiation"/>.</t> <t>There is lowtarget="RFC8843" sectionFormat="of" section="15"/>.</li> <li>The overheadforof security associationestablishment.</t> </list> </t> <t>The Disadvantages</t> <t> <list style="letters"> <t>There are aestablishment is low.</li> </ol> <t>Disadvantages:</t> <ol spacing="normal" type="1"> <li> <t>A slightly higher number of RTP sessionsneededare needed, compared toMultiple Media Typesmultiple media types in oneSession <xref target="section-5.1"/>.session (<xref target="sect-5.1" format="default"/>). Thisimplies: <list style="symbols"> <t>Moreimplies the following: </t> <ul spacing="normal"> <li>More NAT/FW state isneeded.</t> <t>Thereneeded.</li> <li>The cost of NAT/FW traversal is increasedNAT/FW-traversal costin terms of both processing anddelay.</t> </list> </t> <t>Theredelay.</li> </ul> </li> <li>There is some potential for concernwithregarding legacy implementations that don't support the RTP specification fully when it comes to handling multipleSSRCSSRCs perendpoint.</t> <t>Itendpoint.</li> <li>It is not possible to control securityassociationassociations for sets of RTP streams within the same media type with today's key-management mechanisms, unless these are split into different RTP sessions (<xreftarget="section-5.3"/>).</t> </list> </t>target="sect-5.3" format="default"/>).</li> </ol> <t>For RTP applications where all RTP streams of the same media type share the same usage, this structure provides efficiency gains in the amount of network state used and provides morefate sharingfate-sharing with other media flows of the same type. At the same time, itisstillmaintainingmaintains almost all functionalities for the negotiation signaling of properties per individual mediatype,type and also enablesflow basedflow-based QoSprioritisationprioritization between media types. It handles multi-party sessions well, independently of multicast orcentralisedcentralized transport distribution, as additional sources can dynamically enter and leave the session.</t> </section> <sectionanchor="section-5.3" title="Multipleanchor="sect-5.3" numbered="true" toc="default"> <name>Multiple Sessions for One MediaType">Type</name> <t>This design goes one step further thanabove (<xref target="section-5.2"/>)the design discussed in <xref target="sect-5.2" format="default"/> by also using multiple RTP sessionsalsofor a single media type. The main reason for going in this direction is that the RTP application needs separation of the RTP streamsdueaccording to their usage, suchas e.g.as, for example, scalability over multicast, simulcast, the need for extended QoSprioritisation,prioritization, or the need for fine-grainedsignallingsignaling using RTP session-focusedsignallingsignaling tools.</t><t>The Advantages:</t> <t> <list style="numbers"> <t>This<t>Advantages:</t> <ol spacing="normal" type="1"> <li>This design is more suitable for multicast usage where receivers can individually select which RTP sessions they want to participate in, assuming that each RTP session has its own multicastgroup.</t> <t>Thegroup.</li> <li>When multiple different usages exist, the application can indicate its usage of the RTP streamsonat the RTP sessionlevel, when multiple different usages exist.</t> <t>Therelevel.</li> <li>There is less need for SSRC-specific explicitsignallingsignaling for each media stream and thus a reduced need for explicit and timelysignallingsignaling when RTP streams are added orremoved.</t> <t>Itremoved.</li> <li>It enables detailed QoSprioritisationprioritization for flow-basedmechanisms.</t> <t>Itmechanisms.</li> <li>It works well withSplit Component Terminalsplit component terminals (seeSection 3.10 of<xreftarget="RFC7667"/>).</t> <t>Thetarget="RFC7667" sectionFormat="of" section="3.10"/>).</li> <li>The scope for who is included in a security association can be structured around the different RTP sessions, thus enabling such functionality with existingkey-management.</t> </list> </t> <t>The Disadvantages:</t> <t> <list style="letters"> <t>Therekey-management mechanisms.</li> </ol> <t>Disadvantages:</t> <ol spacing="normal" type="1"> <li>There is an increased amount of session configuration state compared toMultiplemultiple SSRCs of theSame Media Type,same media type (<xref target="sect-5.2"/>), due to the increased amount of RTPsessions.</t> <t>Forsessions.</li> <li>For RTP streams that are part of scalability,simulcastsimulcast, or transport robustness, a methodto bindfor binding sources across multiple RTP sessions isneeded.</t> <t>Thereneeded.</li> <li>There is some potential for concernwithregarding legacy implementations that don't support the RTP specification fully when it comes to handling multipleSSRCSSRCs perendpoint.</t> <t>There is higherendpoint.</li> <li>The overheadforof security associationestablishment,establishment is higher, due to the increased number of RTPsessions.</t> <t>Ifsessions.</li> <li>If the applications needmore fine-grainedfiner controlthan per RTP sessionover which participantsthatin a given RTP session are included in different sets of security associations, most of today's key-management mechanisms will have difficulties establishing such asession.</t> </list> </t>session.</li> </ol> <t>Formore complexmore-complex RTP applications that have several different usages for RTP streams of the same mediatype,type orusesthat use scalability or simulcast, this solution can enable thosefunctionsfunctions, at the cost of increased overhead associated with the additional sessions. This type of structure is suitable formore advancedmore-advanced applications as well as multicast-based applications requiring differentiation to different participants.</t> </section> <sectionanchor="section-5.4" title="Singleanchor="sect-5.4" numbered="true" toc="default"> <name>Single SSRC perEndpoint">Endpoint</name> <t>In thisdesigndesign, each endpoint in a point-to-point session has only a singleSSRC, thusSSRC; thus, the RTP session contains only twoSSRCs,SSRCs -- one local and one remote. This session can be usedboth unidirectional, i.e. only a singleeither unidirectionally (i.e., one SSRC sends an RTPstream,stream that is received by the other SSRC) orbi-directional, i.e.bidirectionally (i.e., the two SSRCs bothendpoints have onesend an RTP stream and receive the RTP streameach.sent by the other endpoint). If the application needs additional media flows between the endpoints, it will have to establish additional RTP sessions.</t><t>The Advantages:</t> <t> <list style="numbers"> <t>This<t>Advantages:</t> <ol spacing="normal" type="1"> <li>This design has greatlegacy interoperabilitypotential for interoperability with legacy applications, as it will not tax any RTP stackimplementations.</t> <t>The signalling has good possibilitiesimplementations.</li> <li>The signaling system makes it possible to negotiate and describe the exact formats and bitrates for each RTP stream, especially using today's tools inSDP.</t> <t>ItSDP.</li> <li>It is possible to control securityassociationassociations per RTP stream with currentkey-management,key-management functions, since each RTP stream is directly related to an RTPsession,session and the most commonly used keying mechanismsoperatesoperate on a per-sessionbasis.</t> </list> </t> <t>The Disadvantages:</t> <t> <list style="letters"> <t>There is a linear growth of thebasis.</li> </ol> <t>Disadvantages:</t> <ol spacing="normal" type="1"> <li>The amount of NAT/FW state grows linearly with the number of RTPstreams.</t> <t>There is increasedstreams.</li> <li>NAT/FW traversal increases delay and resourceconsumption from NAT/FW-traversal.</t> <t>Thereconsumption.</li> <li>There are likelylarger signallingmore signaling message andsignallingsignaling processing requirements due to the increased amount of session-relatedinformation.</t> <t>Thereinformation.</li> <li>There is higher potential for a single RTP stream to fail during transport between the endpoints, due to the need for a separateNAT/FW-NAT/FW traversal for every RTPstreamstream, since there is only one stream persession.</t> <t>Thesession.</li> <li>The amount of explicit state for relating RTP streams grows, depending on how the application relates RTPstreams.</t> <t>The portstreams.</li> <li>Port consumption might become a problem forcentralisedcentralized services, where the central node's port or 5-tuple filter consumption grows rapidly with the number ofsessions.</t> <t>Forsessions.</li> <li>For applications wheretheRTP stream usage is highly dynamic,i.e. enteringi.e., entities frequently enter andleaving,leave sessions, the amount ofsignallingsignaling can become high. Issues can also arise from the need for timely establishment of additional RTPsessions.</t> <t>If,sessions.</li> <li>If, against therecommendation,recommendation in <xref target="RFC3550"/>, the same SSRC value is reused in multiple RTP sessions rather than being randomly chosen, interworking with applications that use a different multiplexing structure will require SSRCtranslation.</t> </list> </t>translation.</li> </ol> <t>RTP applications with a strong need to interwork with legacy RTP applications 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 non-mixed multi-party sessions, as any given RTP stream from each participant, although having the same usage in the application, needs its own RTP session. In addition, the dynamicbehaviourbehavior that can arise in multi-party applications can tax thesignallingsignaling system and make timely media establishment more difficult.</t> </section> <sectionanchor="section-5.5" title="Summary">anchor="sect-5.5" numbered="true" toc="default"> <name>Summary</name> <t>Both the"Single"single SSRC perEndpoint"endpoint" (<xref target="sect-5.4"/>) andthe "Multiple Media Types"multiple media types inOne Session" areone session" (<xref target="sect-5.1"/>) casesthatrequire full explicitsignallingsignaling of the media streamrelations.relationships. However, they operate on two differentlevelslevels, where the first primarily enablessession level binding,session-level binding and the second needsSSRC levelSSRC-level binding. From another perspective, the two solutions are the twoextreme pointsextremes when it comes to the number of RTP sessions needed.</t> <t>The two otherdesigns, "Multipledesigns -- multiple SSRCs of theSame Media Type"same media type (<xref target="sect-5.2"/>) and"Multiple Sessionsmultiple sessions forOne Media Type",one media type (<xref target="sect-5.3"/>) -- are two examples that primarilyallowsallow for some implicit mapping of the role or usage of the RTP streams based on which RTP session they appear in.It thusThus, they potentiallyallowsallow for lesssignalling andsignaling and, inparticular reducesparticular, reduce the need for real-timesignallingsignaling in sessions with a dynamically changing number of RTP streams. They also represent pointsin-betweenbetween the first two designs when it comes to the amount of RTP sessions established,i.e. representingi.e., they represent an attempt to balance the amount of RTP sessions with the functionality the communication session provides at bothonthe network level andon signallingthe signaling level.</t> </section> </section> <sectionanchor="section-6" title="Guidelines">anchor="sect-6" numbered="true" toc="default"> <name>Guidelines</name> <t>This section contains a number of multi-stream guidelines for implementers, system designers,orand specification writers.</t><t> <list hangIndent="3" style="hanging"> <t hangText="Do<dl newline="true" spacing="normal"> <dt>Do not require the use of the same SSRC value across RTPsessions:"> <vspace blankLines="0"/>sessions:</dt> <dd> As discussed in <xreftarget="section-3.4.3"/>target="sect-3.4.3" format="default"/>, thereexist drawbacks inare downsides to using the same SSRC in multiple RTP sessions as a mechanism to bind related RTP streams together. It is instead recommended to use a mechanism to explicitly signal therelation, eitherrelationship, inRTP/RTCPeither RTP&wj;/RTCP orinthesignallingsignaling mechanism used to establish the RTPsession(s).</t> <t hangText="Usesession(s).</dd> <dt>Use additional RTP streams for additional mediasources:">Insources:</dt> <dd>In the cases where an RTP endpoint needs to transmit additional RTP streams of the same media type in the application, with the same processing requirements at the network and RTP layers, it is suggested to send them in the same RTP session. Forexampleexample, in the case of a telepresence room where there are threecameras,cameras and each camera captures2two persons sitting at the table,sendingwe suggest that each cameraassend its own RTP stream within a single RTPsession is suggested.</t> <t hangText="Usesession.</dd> <dt>Use additional RTP sessions for streams with differentrequirements:">requirements:</dt> <dd> When RTP streams have different processing requirements from the network or the RTP layer at the endpoints, it is suggested that the different types of streamsarebe put in different RTP sessions. This includes the case where different participants want different subsets of the set of RTPstreams.</t> <t hangText="Whenstreams.</dd> <dt>Use grouping when using multiple RTPsessions, use grouping:">sessions:</dt> <dd> When using multiple RTP session solutions, it is suggested to explicitly group the involved RTP sessions when needed using asignalling mechanism,signaling mechanism -- forexample Theexample, see <xref target="RFC5888">"The Session Description Protocol (SDP) GroupingFramework <xref target="RFC5888"/>,Framework"</xref> -- using some appropriate groupingsemantics.</t> <t hangText="RTP/RTCP Extensions Support Multiplesemantics.</dd> <dt>Ensure that RTP/RTCP extensions support multiple RTPStreamsstreams asWellwell asMultiplemultiple RTPSessions:">Whensessions:</dt> <dd>When defining an RTP or RTCP extension, the creator needs to consider if this extension is applicabletofor use with additional SSRCs and multiple RTP sessions. Any extension intended to be generic must support both. Extensions that are not as generally applicable will have to considerifwhether interoperability is better served by defining a single solution or providing bothoptions.</t> <t hangText="Extensionsoptions.</dd> <dt>Provide adequate extensions forTransport Support:">Whentransport support:</dt> <dd>When defining new RTP/RTCP extensions intended for transport support, like the retransmission or FEC mechanisms, they must include support for both multiple RTP streams in the same RTP session and multiple RTP sessions, such that application developers can choose freely from the set of mechanisms without concerning themselves with which of the multiplexing choices a particular solutionsupports.</t> </list> </t>supports.</dd> </dl> </section> <sectionanchor="section-8" title="IANA Considerations">anchor="sect-8" numbered="true" toc="default"> <name>IANA Considerations</name> <t>This documentmakeshas norequest of IANA.</t> <t>Note to RFC Editor: this section can be removed on publication as an RFC.</t>IANA actions.</t> </section> <sectionanchor="section-9" title="Security Considerations">anchor="sect-9" numbered="true" toc="default"> <name>Security Considerations</name> <t>The security considerationsofdiscussed in the RTP specification <xreftarget="RFC3550"/>,target="RFC3550" format="default"/>; any applicable RTP profile <xreftarget="RFC3551"/>,<xref target="RFC4585"/>,<xref target="RFC3711"/>,target="RFC3551" format="default"/> <xref target="RFC4585" format="default"/> <xref target="RFC3711" format="default"/>; and the extensions for sending multiple media types in a single RTP session <xreftarget="I-D.ietf-avtcore-multi-media-rtp-session"/>,target="RFC8860" format="default"/>, RID <xreftarget="I-D.ietf-mmusic-rid"/>,target="RFC8851" format="default"/>, BUNDLE <xreftarget="I-D.ietf-mmusic-sdp-bundle-negotiation"/>,target="RFC8843" format="default"/>, <xreftarget="RFC5760"/>,target="RFC5760" format="default"/>, and <xreftarget="RFC5761"/>,target="RFC5761" format="default"/> apply if selected and thus need to be considered in the evaluation.</t><t>There is discussion of<t><xref target="sect-4.3" format="default"/> discusses the security implications of choosing multipleSSRC vsSSRCs vs. multiple RTPsessions in <xref target="section-4.3"/>.</t>sessions.</t> </section><section title="Contributors"> <t>Hui Zheng (Marvin) contributed to WG draft versions -04 and -05</middle> <back> <references> <name>References</name> <references> <name>Normative References</name> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.3550.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.3551.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.3711.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.4585.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.5576.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.5760.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.5761.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.7656.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.7667.xml"/> <!-- draft-ietf-avtcore-multi-media-rtp-session (RFC 8860) --> <reference anchor="RFC8860" target="https://www.rfc-editor.org/info/rfc8860"> <front> <title>Sending Multiple Types of Media in a Single RTP Session</title> <author initials="M." surname="Westerlund" fullname="Magnus Westerlund"> <organization/> </author> <author initials="C." surname="Perkins" fullname="Colin Perkins"> <organization/> </author> <author initials="J." surname="Lennox" fullname="Jonathan Lennox"> <organization/> </author> <date month="January" year="2021"/> </front> <seriesInfo name="RFC" value="8860"/> <seriesInfo name="DOI" value="10.17487/RFC8860"/> </reference> <!-- draft-ietf-mmusic-rid (RFC 8851) --> <reference anchor='RFC8851' target="https://www.rfc-editor.org/info/rfc8851"> <front> <title>RTP Payload Format Restrictions</title> <author initials='A.B.' surname='Roach' fullname='Adam Roach' role="editor"> <organization /> </author> <date month='January' year='2021' /> </front> <seriesInfo name="RFC" value="8851"/> <seriesInfo name="DOI" value="10.17487/RFC8851"/> </reference> <!-- draft-ietf-mmusic-sdp-bundle-negotiation (RFC 8843) --> <reference anchor="RFC8843" target="https://www.rfc-editor.org/info/rfc8843"> <front> <title>Negotiating Media Multiplexing Using thedocument. </t> </section> <section title="Acknowledgments"> <t>The Authors like to acknowledge and thank Cullen Jennings, Dale R Worley, Huang Yihong (Rachel), Benjamin Kaduk, Mirja Kuehlewind, and Vijay GurbaniSession Description Protocol (SDP)</title> <author initials="C" surname="Holmberg" fullname="Christer Holmberg"> <organization/> </author> <author initials="H" surname="Alvestrand" fullname="Harald Alvestrand"> <organization/> </author> <author initials="C" surname="Jennings" fullname="Cullen Jennings"> <organization/> </author> <date month="January" year="2021"/> </front> <seriesInfo name="RFC" value="8843"/> <seriesInfo name="DOI" value="10.17487/RFC8843"/> </reference> <!-- draft-ietf-avtext-rid (RFC 8852) --> <reference anchor='RFC8852' target="https://www.rfc-editor.org/info/rfc8852"> <front> <title>RTP Stream Identifier Source Description (SDES)</title> <author initials='A.B.' surname='Roach' fullname='Adam Roach'> <organization /> </author> <author initials='S' surname='Nandakumar' fullname='Suhas Nandakumar'> <organization /> </author> <author initials='P' surname='Thatcher' fullname='Peter Thatcher'> <organization /> </author> <date month='January' year='2021' /> </front> <seriesInfo name="RFC" value="8852"/> <seriesInfo name="DOI" value="10.17487/RFC8852"/> </reference> <!-- draft-ietf-perc-srtp-ekt-diet (RFC 8870) --> <reference anchor="RFC8870" target="https://www.rfc-editor.org/info/rfc8870"> <front> <title>Encrypted Key Transport forreviewDTLS andcomments. </t> </section> </middle> <back> <references title="Normative References"> &RFC3550; &RFC3551; &RFC3711; &RFC4585; &RFC5576; &RFC5760; &RFC5761; &RFC7656; &RFC7667; &I-D.ietf-avtcore-multi-media-rtp-session; &I-D.ietf-mmusic-rid; &I-D.ietf-mmusic-sdp-bundle-negotiation; &I-D.ietf-perc-srtp-ekt-diet;Secure RTP</title> <author initials="C" surname="Jennings" fullname="Cullen Jennings"> <organization>company</organization> </author> <author initials="J" surname="Mattsson" fullname="John Mattsson"> <organization>company</organization> </author> <author initials="D" surname="McGrew" fullname="David A. McGrew"> <organization>company</organization> </author> <author initials="D" surname="Wing" fullname="Dan Wing"> <organization>company</organization> </author> <author initials="F" surname="Andreasen" fullname="Flemming Andreasen"> <organization>company</organization> </author> <date month="January" year="2021"/> </front> <seriesInfo name="RFC" value="8870"/> <seriesInfo name="DOI" value="10.17487/RFC8870"/> </reference> </references><references title="Informative References"> &RFC2198; &RFC2205; &RFC2474; &RFC2974; &RFC3261; &RFC3264; &RFC3389; &RFC3830; &RFC4103; &RFC4383; &RFC4566; &RFC4568; &RFC4588; &RFC5104; &RFC5109; &RFC5389; &RFC5764; &RFC5888; &RFC6465; &RFC7201; &RFC7657; &RFC7826; &RFC7983; &RFC8088; &RFC8108; &RFC8445; &I-D.ietf-avtext-rid; &I-D.ietf-perc-private-media-framework;<references> <name>Informative References</name> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.2198.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.2205.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.2474.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.2974.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.3261.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.3264.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.3389.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.3830.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.4103.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.4383.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.4566.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.4568.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.4588.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.5104.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.5109.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.5389.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.5764.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.5888.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.6465.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.7201.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.7657.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.7826.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.7983.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.8088.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.8108.xml"/> <xi:include href="https://xml2rfc.ietf.org/public/rfc/bibxml/reference.RFC.8445.xml"/> <!-- draft-ietf-perc-private-media-framework (RFC 8871) --> <reference anchor='RFC8871' target="https://www.rfc-editor.org/info/rfc8871"> <front> <title>A Solution Framework for Private Media in Privacy-Enhanced RTP Conferencing (PERC)</title> <author initials='P' surname='Jones' fullname='Paul Jones'> <organization /> </author> <author initials='D' surname='Benham' fullname='David Benham'> <organization /> </author> <author initials='C' surname='Groves' fullname='Christian Groves'> <organization /> </author> <date month='January' year='2021'/> </front> <seriesInfo name="RFC" value="8871"/> <seriesInfo name="DOI" value="10.17487/RFC8871"/> </reference> <referenceanchor="JINGLE">anchor="JINGLE" target="https://xmpp.org/extensions/xep-0166.html"> <front> <title>XEP-0166: Jingle</title> <author initials="S." surname="Ludwig"> </author> <author initials="J." surname="Beda"> </author> <author initials="P." surname="Saint-Andre"> </author> <author initials="R." surname="McQueen"> </author> <author initials="S." surname="Egan"> </author> <author initials="J." surname="Hildebrand"> </author> <date month="September" year="2018"/> </front><seriesInfo name="XMPP.org" value="https://xmpp.org/extensions/xep-0166.html"/></reference> </references> </references> <sectionanchor="section-a" title="Dismissinganchor="sect-a" numbered="true" toc="default"> <name>Dismissing Payload TypeMultiplexing">Multiplexing</name> <t>This section documents a number of reasons why using the payload type as a multiplexing point is unsuitable for most issues related to multiple RTP streams. Attempting to usePayloadpayload type multiplexing beyond its defined usage haswell knownwell-known negative effects onRTPRTP, as discussed below. To use the payload type as the single discriminator for multiple streams implies that all the different RTP streams are being sent with the same SSRC, thus using the same timestamp and sequence number space.This hasThe manyeffects:</t> <t> <list style="numbers"> <t>Putting constraintseffects of using payload type multiplexing are as follows:</t> <ol spacing="normal" type="1"> <li>Constraints are placed on the RTP timestamp rate for the multiplexed media. For example, RTP streams that use different RTP timestamp rates cannot be combined, as the timestamp values need to be consistent across all multiplexed media frames.ThusThus, streams are forced to use the same RTP timestamp rate. When this is not possible, payload type multiplexing cannot beused.</t> <t>Manyused.</li> <li>Many RTP payload formats can fragment a media object over multiple RTP 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 canrelatively simplebesolveddone relatively easily on the sender side by ensuring that the fragments of each RTP stream are sent insequence.</t> <t>Somesequence.</li> <li>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 a repair mechanism within a single media context. Thetext/ T140text&wj;/t140 payload format <xreftarget="RFC4103"/>target="RFC4103" format="default"/> is an example of such a format. These formats will need a sequence numbering abstraction function between RTP and the individual RTP stream before being used with payload typemultiplexing.</t> <t>Sendingmultiplexing.</li> <li>Sending multiple media streams in the same sequence number space makes it impossible to determine which media stream lost a packet.This as the payload type that is used for demultiplex the media stream is not received. Thus, causingSuch a scenario causes difficulties, since the receiverdifficulties in determiningcannot determine to which streamtoit should applypacket losspacket-loss concealment or other stream-specificloss mitigation mechanisms.</t> <t>Ifloss-mitigation mechanisms.</li> <li>If RTPRetransmissionretransmission <xreftarget="RFC4588"/>target="RFC4588" format="default"/> is used andthere is a loss,packet loss occurs, it is possible to ask for the missing packet(s) by SSRC and sequencenumber,number -- not by payload type. If only some of the payload type multiplexed streams are of interest, there is no wayof tellingto tell which missingpacket(s)packet or packets belong to theinteresting stream(s)stream or streams of interest, and all lost packets need to be requested, wastingbandwidth.</t> <t>Thebandwidth.</li> <li>The current RTCP feedback mechanisms are built around providing feedback on RTP streams based on stream ID (SSRC), packet (sequencenumbers)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 RTP payload type is difficult without extending existing RTCPreporting.</t> <t>Thereporting.</li> <li>The current RTCP media control messages<xref target="RFC5104"/>specification <xref target="RFC5104" format="default"/> is oriented around controlling particular media flows,i.e.i.e., requests are done by addressing a particular SSRC. Such mechanisms would need to be redefined to support payload typemultiplexing.</t> <t>Themultiplexing.</li> <li>The number of payload typesareis 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 <xreftarget="RFC5761"/>target="RFC5761" format="default"/> where a number of payload types are blocked due to the overlap between RTP andRTCP.</t> <t>AtRTCP.</li> <li>At times, there is a need to group multiplexedstreams and thisstreams. This is currently possible for RTP sessions andfor SSRC,SSRCs, but there is no defined way to group payloadtypes.</t> <t>Ittypes.</li> <li>It is currently not possible to signal bandwidth requirements per RTP stream when using payload typemultiplexing.</t> <t>Mostmultiplexing.</li> <li>Most existing SDPmedia levelmedia-level attributes cannot be applied on aper payload type levelper-payload-type basis and would requirere-definitionredefinition in thatcontext.</t> <t>Acontext.</li> <li>A legacy endpoint that does not understand the indication that different RTP payload types are different RTP streams might be slightly confused by the large amount of possibly overlapping or identically defined RTP payloadtypes.</t> </list> </t>types.</li> </ol> </section> <sectionanchor="section-b" title="Signalling Considerations"> <t>Signallinganchor="sect-b" numbered="true" toc="default"> <name>Signaling Considerations</name> <t>Signaling is not an architectural consideration for RTP itself, so this discussion has been moved to an appendix. However, it is extremely important for anyone building complete applications, so it is deserving of discussion.</t> <t>We documentsalientsome issues here that need to be addressedby the WGs that usewhen using some form of signaling to establish RTP sessions. These issues cannotsimplybe addressed by simply tweaking, extending, or profilingRTP, butRTP; rather, they require a dedicated andindepthin-depth look at the signaling primitives that set up the RTP sessions.</t> <t>There exist varioussignallingsignaling solutions for establishing RTP sessions. Many are based on SDP <xreftarget="RFC4566"/> based, howevertarget="RFC4566" format="default"/>; however, SDP functionality is also dependent on thesignallingsignaling protocols carrying the SDP.RTSPThe Real-Time Streaming Protocol (RTSP) <xreftarget="RFC7826"/>target="RFC7826" format="default"/> andSAPthe Session Announcement Protocol (SAP) <xreftarget="RFC2974"/>target="RFC2974" format="default"/> both use SDP in a declarative fashion, while SIP <xreftarget="RFC3261"/>target="RFC3261" format="default"/> uses SDP with the additional definition ofOffer/Answeroffer/answer <xreftarget="RFC3264"/>.target="RFC3264" format="default"/>. The impact onsignallingsignaling, and especiallySDPon SDP, needs to beconsideredconsidered, as it can greatly affect how to deploy a certain multiplexing point choice.</t> <sectionanchor="section-b.1" title="Session Oriented Properties">anchor="sect-b.1" numbered="true" toc="default"> <name>Session-Oriented Properties</name> <t>One aspect oftheexistingsignallingsignaling protocols is thatit isthey are focused on RTPsessions, orsessions or, in the case of SDP, the concept of mediadescription concept. There are adescriptions. A number of thingsthataresignalled onsignaled at the media descriptionlevellevel, but those are not necessarily strictly bound to an RTP session and could be of interestto signal specificallyfor signaling, especially for a particular RTP stream (SSRC) within the session. The following properties have been identified as being potentially usefulto signalfor signaling, and not onlyonat the RTP session level:</t><t> <list style="symbols"> <t>Bitrate/Bandwidth exist<ul spacing="normal"> <li>Bitrate and/or bandwidth can be specified today onlyatas an aggregate limit, or as a common "any RTP stream" limit, unless either codec-specific bandwidth limiting or RTCPsignallingsignaling usingTMMBRTemporary Maximum Media Stream Bit Rate Request (TMMBR) messages <xreftarget="RFC5104"/>target="RFC5104" format="default"/> isused.</t> <t>Whichused. </li> <li>Which SSRCthatwill use which RTP payload type (this information will be visiblefromin the first mediapacket,packet but is sometimes useful toknowhave before the packetarrival).</t> </list> </t>arrives).</li> </ul> <t>Some of these issues are clearly SDP's problem rather than RTP limitations. However, if the aim is to deployana solutionusing additional SSRCsthat uses several SSRCs and contains several sets of RTP streams with different properties (encoding/packetizationparameter, bit-rate,parameters, bitrate, etc.), putting each set in a different RTP session would directly enable negotiation of the parameters for each set. If insisting on additionalSSRCSSRCs only, a number ofsignallingsignaling extensions are needed to clarify that there are multiple sets of RTP streams with different properties and that theyneedin fact need to be kept different, since a single set will not satisfy the application's requirements.</t> <t>For some parameters, such as RTP payload type,resolutionresolution, andframerate, aframe rate, an SSRC-linked mechanism has been proposed in <xreftarget="I-D.ietf-mmusic-rid"/></t>target="RFC8851" format="default"/>.</t> </section> <sectionanchor="section-b.2" title="SDPanchor="sect-b.2" numbered="true" toc="default"> <name>SDP Prevents Multiple MediaTypes">Types</name> <t>SDPchose to useuses them="m=" linebothto both delineate an RTP session andtospecify thetop level of the MIMEtop-level mediatype;type: audio, video, text, image, application. This media type is used as the top-level media type for identifying the actual payload format and is bound to a particular payload type using thertpmap"a=rtpmap:" attribute. This binding has to be loosened in order to use SDP to describe RTP sessions containing multipleMIME top leveltop-level media types.</t> <t><xreftarget="I-D.ietf-mmusic-sdp-bundle-negotiation"/>target="RFC8843" format="default"/> describes how to let multiple SDP media descriptions use a single underlying transport in SDP, which allowsto definethe definition of one RTP session withmedia types havingdifferentMIME top leveltop-level media types.</t> </section> <sectionanchor="section-b.3" title="Signallinganchor="sect-b.3" numbered="true" toc="default"> <name>Signaling RTP StreamUsage">Usage</name> <t>RTP streams being transported in RTP havesomea particular usage in an RTP application.ThisIn many applications to date, this usage of the RTP stream isin many applications so farimplicitlysignalled.signaled. For example, an application might choose to take all incoming audio RTP streams, mixthemthem, and play them out. However, inmore advancedmore-advanced applications that use multiple RTPstreamsstreams, there will be more than a single usage or purpose among the set of RTP streams being sent or received. RTP applications will need to somehow signal thisusage somehow.usage. Thesignallingsignaling that is used will have to identify the RTP streams affected by theirRTP- levelRTP-level identifiers, which means that they have to be identifiedeitherby either their session orbytheir SSRC + session.</t> <t>In some applications, the receiver cannotutiliseutilize the RTP stream at all before it has received thesignallingsignaling message describing the RTP stream and its usage. In other applications, there exists a default handling method that is appropriate.</t> <t>If all RTP 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,signallingsignaling needs to identify both the session and the SSRC.</t> <t>If thissignallingsignaling affects how any RTP central node, like an RTP mixer or translator that selects,mixesmixes, or processes streams, treats the streams, the node will also need to receive the samesignallingsignaling to know how to treat RTP streams with differentusageusages in the right fashion.</t> </section> </section> <section numbered="false" toc="default"> <name>Acknowledgments</name> <t>The authors would like to acknowledge and thank <contact fullname="Cullen Jennings"/>, <contact fullname="Dale R. Worley"/>, <contact fullname="Huang Yihong (Rachel)"/>, <contact fullname="Benjamin Kaduk"/>, <contact fullname="Mirja Kühlewind"/>, and <contact fullname="Vijay Gurbani"/> for review and comments.</t> </section> <section numbered="false" toc="default"> <name>Contributors</name> <t><contact fullname="Hui Zheng (Marvin)"/> contributed to WG draft versions -04 and -05 of the document. </t> </section> </back> </rfc>