One document matched: draft-ietf-avtcore-multi-media-rtp-session-02.xml


<?xml version="1.0" encoding="US-ASCII"?>
<!DOCTYPE rfc SYSTEM "rfc2629.dtd">
<?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"?>
<rfc category="std" docName="draft-ietf-avtcore-multi-media-rtp-session-02"
     ipr="trust200902" updates="3550, 3551">
  <front>
    <title abbrev="Multiple Media Types in an RTP Session">Multiple Media
    Types in an RTP Session</title>

    <author fullname="Magnus Westerlund" initials="M." surname="Westerlund">
      <organization>Ericsson</organization>

      <address>
        <postal>
          <street>Farogatan 6</street>

          <city>SE-164 80 Kista</city>

          <country>Sweden</country>
        </postal>

        <phone>+46 10 714 82 87</phone>

        <email>magnus.westerlund@ericsson.com</email>
      </address>
    </author>

    <author fullname="Colin Perkins" initials="C. " surname="Perkins">
      <organization>University of Glasgow</organization>

      <address>
        <postal>
          <street>School of Computing Science</street>

          <city>Glasgow</city>

          <code>G12 8QQ</code>

          <country>United Kingdom</country>
        </postal>

        <email>csp@csperkins.org</email>
      </address>
    </author>

    <author fullname="Jonathan Lennox" initials="J." surname="Lennox">
      <organization abbrev="Vidyo">Vidyo, Inc.</organization>

      <address>
        <postal>
          <street>433 Hackensack Avenue</street>

          <street>Seventh Floor</street>

          <city>Hackensack</city>

          <region>NJ</region>

          <code>07601</code>

          <country>US</country>
        </postal>

        <email>jonathan@vidyo.com</email>
      </address>
    </author>

    <date day="25" month="February" year="2013"/>

    <workgroup>AVTCORE WG</workgroup>

    <abstract>
      <t>This document specifies how an RTP session can contain media streams
      with media from multiple media types such as audio, video, and text.
      This has been restricted by the RTP Specification, and thus this
      document updates RFC 3550 and RFC 3551 to enable this behaviour for
      applications that satisfy the applicability for using multiple media
      types in a single RTP session.</t>
    </abstract>
  </front>

  <middle>
    <section title="Introduction">
      <t>When the <xref target="RFC3550">Real-time Transport Protocol
      (RTP)</xref> was designed, close to 20 years ago, IP networks were very
      different compared to the ones in 2013 when this is written. The almost
      ubiquitous deployment of Network Address Translators (NAT) and Firewalls
      has increased the cost and likely-hood of communication failure when
      using many different transport flows. Thus there exists a pressure to
      reduce the number of concurrent transport flows.</t>

      <t><xref target="RFC3550">RTP</xref> recommends against sending several
      different types of media, for example audio and video, in a single RTP
      session. The <xref target="RFC3551">RTP profile for Audio and Video
      Conferences with Minimal Control (RTP/AVP)</xref> mandates a similar
      restriction. The motivation for these limitations is partly to allow
      lower layer Quality of Service (QoS) mechanisms to be used, and partly
      due to limitations of the RTCP timing rules that assumes all media in a
      session to have similar bandwidth. The <xref target="RFC4566">Session
      Description Protocol (SDP)</xref>, as one of the dominant signalling
      method for establishing RTP session, has enforced this rule, simply by
      not allowing multiple media types for a given receiver destination or
      set of ICE candidates, which is the most common method to determine
      which RTP session the packets are intended for.</t>

      <t>The fact that these limitations have been in place for so long a
      time, in addition to RFC 3550 being written without fully considering
      multiple media types in an RTP session, does result in a number of
      considerations being needed when allowing this behaviour. This document
      provides such considerations regarding applicability as well as
      functionality, including normative specification of behaviour.</t>

      <t>First, some basic definitions are provided. This is followed by a
      background that discusses the motivation in more detail. A overview of
      the solution of how to provide multiple media types in one RTP session
      is then presented. Next is the formal applicability this specification
      have followed by the normative specification. This is followed by a
      discussion how some RTP/RTCP Extensions is expected to function in the case of
      multiple media types in one RTP session. A specification of the
      requirements on signalling from this specification and a look how this
      is realized in SDP using <xref
      target="I-D.ietf-mmusic-sdp-bundle-negotiation">Bundle</xref>. The
      document ends with the security considerations.</t>
    </section>

    <section title="Definitions">
      <t>The following terms are used with supplied definitions:<list
          style="hanging">
          <t hangText="Endpoint:">A single entity sending or receiving RTP
          packets. It can be decomposed into several functional blocks, but
          as long as it behaves as a single RTP stack entity it is
          classified as a single endpoint.</t>

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

          <t hangText="Media Type:">Audio, video, text or application whose
          form and meaning are defined by a specific real-time
          application.</t>

          <t hangText="QoS:">Quality of Service, i.e. network mechanisms
          that intended to ensure that the packets within a flow or with a
          specific marking are transported with certain properties.</t>

          <t hangText="RTP Session:">As defined by <xref target="RFC3550"/>,
          the endpoints belonging to the same RTP Session are those that
          share a single SSRC space. That is, those endpoints can see an
          SSRC identifier transmitted by any one of the other endpoints. An
          endpoint can receive an SSRC either as SSRC or as CSRC in RTP and
          RTCP packets. Thus, the RTP Session scope is decided by the
          endpoints' network interconnection topology, in combination with
          RTP and RTCP forwarding strategies deployed by endpoints and any
          interconnecting middle nodes.</t>
        </list></t>

      <t>The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
      "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
      document are to be interpreted as described in <xref
      target="RFC2119"/>.</t>

    </section>

    <section title="Motivation ">
      <t>This section discusses in more detail the main motivations why
      allowing multiple media types in the same RTP session is suitable.</t>

      <section title="NAT and Firewalls">
        <t>The existence of NATs and Firewalls at almost all Internet access
        has had implications on protocols like RTP that were designed to use
        multiple transport flows. First of all, the NAT/FW traversal solution
        needs to ensure that all these transport flows are
        established. This has three consequences:<list style="numbers">
            <t>Increased delay to perform the transport flow establishment</t>

            <t>The more transport flows, the more state and the more resource
            consumption in the NAT and Firewalls. When the resource
            consumption in NAT/FWs reaches their limits, unexpected behaviours
            usually occur.</t>

            <t>More transport flows means a higher risk that some transport
            flow fails to be established, thus preventing the application to
            communicate.</t>
          </list></t>

        <t>Using fewer transport flows reduces the risk of communication
        failure, improved establishment behaviour and less load on NAT and
        Firewalls.</t>
      </section>

      <section title="No Transport Level QoS">
        <t>Many RTP-using applications don't utilize any network level Quality
        of Service functions. Nor do they expect or desire any separation in
        network treatment of its media packets, independent of whether they
        are audio, video or text. When an application has no such desire, it
        doesn't need to provide a transport flow structure that simplifies
        flow based QoS.</t>
      </section>

      <section title="Architectural Equality">
        <t>For applications that don't require different lower-layer QoS for
        different media types, and that have no special requirements for RTP
        extensions or RTCP reporting, the requirement to separate different
        media into different RTP sessions might seem unnecessary. Provided the
        application accepts that all media flows will get similar RTCP
        reporting, using the same RTP session for several types of media at
        once appears a reasonable choice. The architecture ought to be agnostic
        about the type of media being carried in an RTP session to the extent
        possible given the constraints of the protocol.</t>
      </section>
    </section>

    <section title="Overview of Solution">
      <t>The goal of the solution is to enable each RTP session to contain
      more than just one media type. This includes having multiple RTP
      sessions containing a given media type, for example having three
      sessions containing both video and audio.</t>

      <t>The solution is quite straightforward. The first step is to override
      the SHOULD and SHOULD NOT language of the <xref target="RFC3550">RTP
      specification</xref>. Similar change is needed to a sentence in Section
      6 of <xref target="RFC3551"/> that states that "different media types
      SHALL NOT be interleaved or multiplexed within a single RTP Session".
      This is resolved by appropriate exception clauses given that this
      specification and its applicability is followed.</t>

      <t>Within an RTP session where multiple media types have been configured
      for use, an SSRC can only send one type of media during its lifetime
      (i.e., it can switch between different audio codecs, since those are
      both the same type of media, but cannot switch between audio and video).
      Different SSRCs MUST be used for the different media sources, the same
      way multiple media sources of the same media type already have to do.
      The payload type will inform a receiver which media type the SSRC is
      being used for. Thus the payload type MUST be unique across all of the
      payload configurations independent of media type that is used in the
      RTP session.</t>

      <t>Some few extra considerations within the RTP sessions also needs to
        be considered. RTCP bandwidth and regular reporting suppression (RTP/AVPF
        and RTP/SAVPF) SHOULD be configured to reduce the impact for bit-rate
      variations between streams and media types. It is also clarified how
      timeout calculations are to be done to avoid any issues. Certain payload
      types like FEC also need additional rules.</t>

      <t>The final important part of the solution to this is to use signalling
      and ensure that agreement on using multiple media types in an RTP
      session exists, and how that then is configured. This memo describes
      some existing requirements, while an external reference defines how this
      is accomplished in SDP.</t>
    </section>

    <section title="Applicability">
      <t>This specification has limited applicability, and anyone intending to
      use it needs to ensure that their application and usage meets the below
      criteria.</t>

      <section title="Usage of the RTP session">
        <t>Before choosing to use this specification, an application
        implementer needs to ensure that they don't have a need for different
        RTP sessions between the media types for some reason. The main rule is
        that if one expects to have equal treatment of all media packets, then
        this specification might be suitable. The equal treatment include
        anything from network level up to RTCP reporting and feedback. The
        document <xref
        target="I-D.westerlund-avtcore-multiplex-architecture">Guidelines for
        using the Multiplexing Features of RTP</xref> gives more detailed
        guidance on aspects to consider when choosing how to use RTP and
        specifically sessions. RTP-using applications that need or would
        prefer multiple RTP sessions, but do not require the functionalities
        or behaviours that multiple transport flows give, can consider using
        <xref target="I-D.westerlund-avtcore-transport-multiplexing">Multiple
        RTP Sessions on a Single Lower-Layer Transport</xref>. It needs to be
        noted that some difference in treatment is still possible to achieve,
        for example marking based QoS, or RTCP feedback traffic for only some
        media streams. </t>

        <t>The second important consideration is the resulting behaviour when
        media flows to be sent within a single RTP session does not have
        similar bandwidth. There are limitations in the RTCP timing rules, and
        this implies a common RTCP reporting interval across all participants
        in a session. If an RTP session contains flows with very different
        bandwidths, for example low-rate audio coupled with high-quality
        video, this can result in either excessive or insufficient RTCP for
        some flows, depending how the RTCP session bandwidth, and hence
        reporting interval, is configured. This is discussed further in <xref
        target="sec.rtcp"/>.</t>
      </section>

      <section title="Signalled Support">
        <t>Usage of this specification is not compatible with anyone following
        RFC 3550 and intending to have different RTP sessions for each media
        type. Therefore there needs to be mutual agreement to use multiple media
        types in one RTP session by all participants within that RTP session.
        This agreement has to be determined using signalling in most cases.</t>

        <t>This requirement can be a problem for signalling solutions that
        can't negotiate with all participants. For declarative signalling
        solutions, mandating that the session is using multiple media types in
        one RTP session can be a way of attempting to ensure that all
        participants in the RTP session follow the requirement. However, for
        signalling solutions that lack methods for enforcing that a receiver
        supports a specific feature, this can still cause issues.</t>
      </section>

      <section title="Homogeneous Multi-party">
        <t>In multiparty communication scenarios it is important to separate
        two different cases. One case is where the RTP session contains
        multiple participants in a common RTP session. This occurs for example
        in Any Source Multicast (ASM) and Transport Translator topologies as
        defined in <xref target="RFC5117">RTP Topologies</xref>. It can also
        occur in some implementations of RTP mixers that share the same
        SSRC/CSRC space across all participants. The second case is when the
        RTP session is terminated in a middlebox and the other participants
        sources are projected or switched into each RTP session and rewritten
        on RTP header level including SSRC mappings.</t>

        <t>For the first case, with a common RTP session or at least shared
        SSRC/CSRC values, all participants in multiparty communication are
        REQUIRED to support multiple media types in an RTP session. An
        participant using two or more RTP sessions towards a multiparty
        session can't be collapsed into a single session with multiple media
        types. The reason is that in case of multiple RTP sessions, the same
        SSRC value can be use in both RTP sessions without any issues, but
        when collapsed to a single session there is an SSRC collision. In
        addition some collisions can't be represented in the multiple separate
        RTP sessions. For example, in a session with audio and video, an SSRC
        value used for video will not show up in the Audio RTP session at the
        participant using multiple RTP sessions, and thus not trigger any
        collision handling. Thus any application using this type of RTP
        session structure MUST have a homogeneous support for multiple media
        types in one RTP session, or be forced to insert a translator node
        between that participant and the rest of the RTP session.</t>

        <t>For the second case of separate RTP sessions for each multiparty
        participant and a central node it is possible to have a mix of single
        RTP session users and multiple RTP session users as long as one is
        willing to remap the SSRCs used by a participant with multiple RTP
        sessions into non-used values in the single RTP session SSRC space for
        each of the participants using a single RTP session with multiple
        media types. It can be noted that this type of implementation has
        to understand all types of RTP/RTCP extension being used in
        the RTP sessions to correctly be able to translate them between the
        RTP sessions. It can also negatively impact the possibility for loop
        detection, as SSRC/CSRC can't be used to detect the loops, instead
        some other media stream identity name space that is common across all
        interconnect parts are needed.</t>
      </section>

      <section title="Reduced number of Payload Types">
        <t>An RTP session with multiple media types in it have only a single
        7-bit Payload Type range for all its payload types. Within the 128
        available values, only 96 or less if <xref
        target="RFC5761">"Multiplexing RTP Data and Control Packets on a
        Single Port"</xref> is used, all the different RTP payload
        configurations for all the media types need to fit in the available 
        space. For most applications
        this will not be a real problem, but the limitation exists and could
        be encountered.</t>
      </section>

      <section title="Stream Differentiation">
        <t>If network level differentiation of the media streams of different
        media types are desired using this specification can cause severe
        limitations. All media streams in an RTP session, independent of the
        media type, will be sent over the same underlying transport flow. Any
        flow-based Quality of Service (QoS) mechanism will be unable to
        provide differentiated treatment between different media types, e.g.
        to prioritize audio over video. If differentiated treatment is desired
        using flow-based QoS, separate RTP sessions over different underlying
        transport flows needs to be used. </t>

        <t>Any marking-based QoS scheme like DiffServ is not affected unless a
        network ingress marks based on flows, in which case the same
        considerations as for flow based QoS applies.</t>
      </section>

      <section title="Non-compatible Extensions">
        <t>There exist some RTP and RTCP extensions that rely on the existence
        of multiple RTP sessions. If the goal of using an RTP session with
        multiple media types is to have only a single RTP session, then these
        extensions can't be used. If one has no need to have different RTP
        sessions for the media types but is willing to have multiple RTP
        sessions, one for the main media transmission and one for the
        extension, they can be used. It is to be noted that this assumes that
        it is possible to get the extension working when the related RTP
        session contains multiple media types.</t>

        <t>Identified RTP/RTCP extensions that require multiple RTP Sessions
        are:<list style="hanging">
            <t hangText="RTP Retransmission:"><xref target="RFC4588">RTP
            Retransmission</xref> has a session multiplexed mode. It also has
            a SSRC multiplexed mode that can be used instead. So use the mode
            that is suitable for the RTP application.</t>

            <t hangText="XOR-Based FEC:">The <xref target="RFC5109">RTP
            Payload Format for Generic Forward Error Correction</xref> and its
            predecessor <xref target="RFC2733"/> requires a separate RTP
            session unless the FEC data is carried in <xref
            target="RFC2198">RTP Payload for Redundant Audio Data</xref>.
            However, using the Generic FEC with the Redundancy payload has
            another set of restrictions, see <xref
            target="sec-generic-fec"/>.</t>

            <t hangText="">Note that the <xref
            target="RFC5576">Source-Specific Media Attributes</xref>
            specification defines an SDP syntax (the "FEC" semantic of the
            "ssrc-group" attribute) to signal FEC relationships between
            multiple media streams within a single RTP session. However, this
            can't be used as the FEC repair packets need to have the
            same SSRC value as the source packets being protected. <xref
            target="RFC5576"/> does not normatively update and resolve that
            restriction. There is ongoing work on an ULP extension to allow it
            be use FEC streams within the same RTP Session as the source
            stream <xref target="I-D.lennox-payload-ulp-ssrc-mux"/>. </t>
          </list></t>

        <t/>
      </section>
    </section>

    <section title="RTP Session Specification">
      <t>This section defines what needs to be done or avoided to make an RTP
      session with multiple media types function without issues.</t>

      <section title="RTP Session">
        <t>Section 5.2 of <xref target="RFC3550">"RTP: A Transport Protocol
        for Real-Time Applications"</xref> states:<list style="empty">
            <t>For example, in a teleconference composed of audio and video
            media encoded separately, each medium SHOULD be carried in a
            separate RTP session with its own destination transport
            address.</t>

            <t>Separate audio and video streams SHOULD NOT be carried in a
            single RTP session and demultiplexed based on the payload type or
            SSRC fields.</t>
          </list></t>

        <t>This specification changes both of these sentences. The first
        sentence is changed to:<list style="empty">
            <t>For example, in a teleconference composed of audio and video
            media encoded separately, each medium SHOULD be carried in a
            separate RTP session with its own destination transport address,
            unless specification [RFCXXXX] is followed and the application
            meets the applicability constraints.</t>
          </list></t>

        <t>The second sentence is changed to:<list style="empty">
            <t>Separate audio and video streams SHOULD NOT be carried in a
            single RTP session and demultiplexed based on the payload type or
            SSRC fields, unless multiplexed based on both SSRC and payload
            type and usage meets what Multiple Media Types in an RTP Session
            [RFCXXXX] specifies.</t>
          </list></t>

        <t>Second paragraph of Section 6 in <xref target="RFC3551">RTP Profile
        for Audio and Video Conferences with Minimal Control</xref> says:</t>

        <t><list style="empty">
            <t>The payload types currently defined in this profile are
            assigned to exactly one of three categories or media types: audio
            only, video only and those combining audio and video. The media
            types are marked in Tables 4 and 5 as "A", "V" and "AV",
            respectively. Payload types of different media types SHALL NOT be
            interleaved or multiplexed within a single RTP session, but
            multiple RTP sessions MAY be used in parallel to send multiple
            media types. An RTP source MAY change payload types within the
            same media type during a session. See the section "Multiplexing
            RTP Sessions" of RFC 3550 for additional explanation.</t>
          </list>This specifications purpose is to violate that existing SHALL
        NOT under certain conditions. Thus also this sentence has to be changed
        to allow for multiple media type's payload types in the same session.
        The above sentence is changed to:<list style="empty">
            <t>Payload types of different media types SHALL NOT be interleaved
            or multiplexed within a single RTP session unless as specified
            and under the restriction in Multiple Media Types in an RTP
            Session [RFCXXXX]. Multiple RTP sessions MAY be used in parallel
            to send multiple media types.</t>
          </list></t>

        <t>RFC-Editor Note: Please replace RFCXXXX with the RFC number of this
        specification when assigned.</t>

        <t>We can now go on and discuss the five bullets that are motivating
        the previous in Section 5.2 of the <xref target="RFC3550">RTP
        Specification</xref>. They are repeated here for the reader's
        convenience:<list style="numbers">
            <t>If, say, two audio streams shared the same RTP session and the
            same SSRC value, and one were to change encodings and thus acquire
            a different RTP payload type, there would be no general way of
            identifying which stream had changed encodings.</t>

            <t>An SSRC is defined to identify a single timing and sequence
            number space. Interleaving multiple payload types would require
            different timing spaces if the media clock rates differ and would
            require different sequence number spaces to tell which payload
            type suffered packet loss.</t>

            <t>The RTCP sender and receiver reports (see Section 6.4 of RFC
            3550) can only describe one timing and sequence number space per
            SSRC and do not carry a payload type field.</t>

            <t>An RTP mixer would not be able to combine interleaved streams
            of incompatible media into one stream.</t>

            <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-process implementations.</t>
          </list></t>

        <t>Bullets 1 to 3 are all related to that each media source has to use
        one or more unique SSRCs to avoid these issues as mandated <xref
        target="sec-source-restrcitctions">below</xref>. Bullet 4 can be
        served by two arguments, first of all each SSRC will be associated
        with a specific media type, communicated through the RTP payload type,
        allowing a middlebox to do media type specific operations. The second
        argument is that in many contexts blind combining without additional
        contexts are anyway not suitable. Regarding bullet 5 this is a
        understood and explicitly stated applicability limitations for the
        method described in this document.</t>
      </section>

      <section anchor="sec-source-restrcitctions"
               title="Sender Source Restrictions">
        <t>A SSRC in the RTP session MUST only send one media type (audio,
        video, text etc.) during the SSRC's lifetime. The main motivation is
        that a given SSRC has its own RTP timestamp and sequence number
        spaces. The same way that you can't send two streams of encoded audio
        on the same SSRC, you can't send one audio and one video encoding on
        the same SSRC. Each media encoding when made into an RTP stream needs
        to have the sole control over the sequence number and timestamp space.
        If not, one would not be able to detect packet loss for that
        particular stream. Nor can one easily determine which clock rate a
        particular SSRCs timestamp will increase with. For additional
        arguments why RTP payload type based multiplexing of multiple media
        streams doesn't work see Appendix A in <xref
        target="I-D.westerlund-avtcore-multiplex-architecture"/>.</t>
      </section>

      <section title="Payload Type Applicability">
        <t>Most Payload Types have a native media type, like an audio codec is
        natural belonging to the audio media type. However, there exist a
        number of RTP payload types that don't have a native media type. For
        example, transport robustness mechanisms like <xref
        target="RFC4588">RTP Retransmission</xref> and <xref
        target="RFC5109">Generic FEC</xref> inherit their media type from what
        they protect. RTP Retransmission is explicitly bound to the payload
        type it is protecting, and thus will inherit it. However Generic FEC
        is a excellent example of an RTP payload type that has no natural
        media type. The media type for what it protects is not relevant as it
        is the recovered RTP packets that have a particular media type, and
        thus Generic FEC is best categorized as an application media type.</t>

        <t>The above discussion is relevant to what limitations exist for RTP
        payload type usage within an RTP session that has multiple media
        types. In fact <xref target="sec-generic-fec">this document</xref>
        suggest that for usage of Generic FEC (XOR-based) as defined in RFC
        5109 can actually use a single media type when used with independent
        RTP sessions for source and repair data. <list style="hanging">
            <t>Note a particular SSRC carrying Generic FEC will clearly only
            protect a specific SSRC and thus that instance is bound to the
            SSRC's media type. For this specific case, it is possible to have
            one be applicable to both. However, in cases when the signalling
            is setup to enable fall back to using separate RTP sessions, then
            using a different media type, e.g. application, than the media
            being protected can create issues.</t>
          </list></t>
      </section>

      <section anchor="sec.rtcp" title="RTCP ">
        <t>An RTP session has a single set of parameters that configure the
        session bandwidth, the RTCP sender and receiver fractions (e.g., via
        the SDP "b=RR:" and "b=RS: lines), and the parameters of the <xref
        target="RFC4585">RTP/AVPF profile</xref> (e.g., trr-int) if that
        profile (or its <xref target="RFC5124"> secure extension,
        RTP/SAVPF</xref>) is used. As a consequence, the RTCP reporting
        interval will be the same for every SSRC in an RTP session. This
        uniform RTCP reporting interval can result in RTCP reports being sent
        more often than is considered desirable for a particular media type.
        For example, if an audio flow is multiplexed with a high quality video
        flow where the session bandwidth is configured to match the video
        bandwidth, this can result in the RTCP packets having a greater
        bandwidth allocation than the audio data rate. If the reduced minimum
        RTCP interval described in Section 6.2 of <xref target="RFC3550"/> is
        used in the session, which might be appropriate for video where rapid
        feedback is wanted, the audio sources could be expected to send RTCP
        packets more often than they send audio data packets. This is most
        likely undesirable, and while the mismatch can be reduced through
        careful tuning of the RTCP parameters, particularly trr_int in
        RTP/AVPF sessions, it is inherent in the design of the RTCP timing
        rules, and affects all RTP sessions containing flows with mismatched
        bandwidth.</t>

        <t>Having multiple media types in one RTP session also results in more
        SSRCs being present in this RTP session. This increasing the amount of
        cross reporting between the SSRCs. From an RTCP perspective, two RTP
        sessions with half the number of SSRCs in each will be slightly more
        efficient. If someone needs either the higher efficiency due to the
        lesser number of SSRCs or the fact that one can't tailor RTCP usage
        per media type, they need to use independent RTP sessions.</t>

        <t>When it comes to handling multiple SSRCs in an RTP session there is
        a clarification under discussion in <xref
        target="I-D.lennox-avtcore-rtp-multi-stream">Real-Time Transport
        Protocol (RTP) Considerations for Multi-Stream Endpoints</xref>. When
        it comes to configuring RTCP the need for regular periodic reporting
        needs to be weighted against any feedback or control messages being
        sent. The applications using RTP/AVPF or RTP/SAVPF are RECOMMENDED to consider
        setting trr-int parameter to a value suitable for the applications
        needs, thus potentially reducing the need for regular reporting and
        thus releasing more bandwidth for use for feedback or control.</t>

        <t>Another aspect of an RTP session with multiple media types is that
        the used RTCP packets, RTCP Feedback Messages, or RTCP XR metrics used
        might not be applicable to all media types. Instead all RTP/RTCP
        endpoints need to correlate the media type of the SSRC being
        referenced in an messages/packet and only use those that apply to that
        particular SSRC and its media type. Signalling solutions might have
        shortcomings when it comes to indicate that a particular set of RTCP
        reports or feedback messages only apply to a particular media type
        within an RTP session.</t>

        <section title="Timing out SSRCs">
          <t>All used SSRCs in the RTP session MUST use the same timeout
            behaviour to avoid premature timeouts. This will depend on the
            RTP profile and its configuration.  The RTP specification
            provides several options that can influence the values used
            when calculating the time-interval, to avoid such issues when
            using this specification we make clarification on the
            calculations.
        </t>

        <t> For RTP/AVP, RTP/SAVP, RTP/AVPF, and RTP/SAVPF with
          T_rr_interval = 0 the timeout interval SHALL be calculated using
          a multiplier of 5, i.e. the timeout interval becomes 5*Td. The
          Td calculation SHALL be done using a Tmin value of 5 seconds,
          not the reduced minimal interval even if used to calculate RTCP
          packet transmission intervals.  If using either the RTP/AVPF or
          RTP/SAVPF profiles with T_rr_interval != 0 then the calculation
          as specified in Section 3.5.4 of RFC 4585 SHALL be used with a
          multiplier of 5, i.e.  Tmin in the Td calculation is the
          T_rr_interval.</t>
          
        <t>
          Note: If endpoints implementing the RTP/AVP and RTP/AVPF profiles
          (or their secure variants) are combined in a single RTP session,
          and the RTP/AVPF endpoints use a non-zero T_rr_interval that is
          significantly lower than 5 seconds, then there is a risk that
          the RTP/AVP endpoints will prematurely timeout the RTP/AVPF
          endpoints due to their different RTCP timeout intervals. Since
          an RTP session can only use a single RTP profile, this issue
          ought never occur. If such mixed RTP profiles are used, however,
          the RTP/AVPF session MUST NOT use a non-zero T_rr_interval that
          is smaller than 5 seconds. </t>

        <t>
          (tbd: it has been suggested that a minimum non-zero T_rr_interval
          of 4 seconds is more appropriate, due to the nature of the timing 
          rules).
        </t>

        </section>

        <section title="Tuning RTCP transmissions">
          <t>This sub-section discusses what tuning can be done to reduce
          downsides of the shared RTCP packet intervals.</t>

        <t>When using the RTP/AVP or RTP/SAVP profile the tuning one can do is very
          limited. The controls one has are very limited to the RTCP
          bandwidth values and if one scales the minimum RTCP interval
          according to the bandwidth. As the scheduling algorithm includes
          both random factors and reconsideration, one can't simply calculate
          the expected average transmission interval using formula for Td. But
          it does indicate the important factors affecting the transmission
          interval, namely the RTCP bandwidth available for the role (Active
          Sender or Participant), the average RTCP packet size and the number
          of SSRCs classified in the relevant role. Note, that if the ratio of
          senders to total number of session participants are larger than the
          ratio of RTCP bandwidth for senders in relation to the total RTCP
          bandwidth, then senders and receivers are treated together.</t>

          <t>Lets start with some basic observations:<list style="letters">
              <t>Unless scaled minimum RTCP interval is used, then Td prior to
              randomization and reconsideration can never be less than 5
              seconds (assuming default Tmin of 5 seconds).</t>

              <t>If scaled minimum RTCP interval is used Td can become as low
              as 360 divided by RTP Session bandwidth in kilobits. In SDP the
              RTP session bandwidth is signalled using b=AS. A RTP Session
              bandwidth of 72 kbps results in Tmin being 5 seconds. A RTP
              session bandwidth of 360 kbps of course gives a Tmin of 1
              second, and to achieve a Tmin equal to once every frame for a 25
              Hz video stream requires an RTP session bandwidth of 9 Mbps!
              (The use of the RTP/AVPF or RTP/SAVPF profile allows smaller
              Tmin, and hence more frequent RTCP report, as discussed below).</t>

              <t>Lets calculate the number (n) of SSRCs in the RTP session
              that 5% of the session bandwidth can support to yield a Td value
              equal to Tmin with minimal scaling. For this calculation we have
              to make two assumptions. The first is that we will consider most
              or all SSRC being senders resulting in everyone sharing the
              available bandwidth. Secondly we will select an average
              RTCP packet size. This packet will consist of an SR, containing
              (n-1) report blocks up to 31 report blocks, a SDES item with at
              least a CNAME (17 bytes value) in it. Such a basic packet will
              be 800 bytes for n>=32. With these parameters, and as the
              bandwidth goes up the time interval is proportionally decreased
              (due to minimal scaling), thus all the example bandwidths 72
              kbps, 360 kbps and 9 Mbps all support 9 SSRCs. </t>

              <t>The actual transmission interval for a Td value is
              [0.5*Td/1.21828,1.5*Td/1.21828], which means that for Td = 5
              seconds, the interval is actually [2.052,6.156] and the
              distribution is not uniform, it is an exponential increasing
              one. The probability for sending at time X, given it is within
              the interval, is probability of picking X in the interval times
              the probability to randomly picking a number that is <=X
              within the interval with an uniform probability distribution.
              This results in that the majority of the probability mass is
              above the Td value.</t>
            </list></t>

            <t>To conclude, with RTP/AVP and RTP/SAVP the key limitation for small
          unicast sessions are going to be the Tmin value. Thus the RTP
          session bandwidth configured in RTCP has to be sufficient high to
          reach the reporting goals the application has following the rules
          for scaled minimal RTCP interval. </t>

        <t>When using RTP/AVPF or RTP/SAVPF we get a quite powerful additional tool,
          the setting of the T_rr_interval which has several effects on the
          RTCP reporting. First of all as Tmin is set to 0 after the initial
          transmission and regular reporting interval is instead affected of
          the regular bandwidth based calculation and the T_rr_interval. This
          has the affect that we are no longer restricted by the minimal
          interval or even the scaling rule for the minimal rule. Instead the
          RTCP bandwidth and the T_rr_interval is the governing factors. Now
          it also becomes important to separate between the applications need
          for regular reports and RTCP feedback packet types. In both regular
          RTCP mode, as in Early RTCP Mode, the usage of the T_rr_Interval
          prevents regular RTCP packets, i.e. packets without any Feedback
          packets to be sent more often than T_rr_interval. This value is a
          hard as no regular RTCP packet can be sent less than T_rr_interval
          after the previous regular packet packet. </t>

          <t>So for applications that has a use for feedback packets for some
          media streams, for example video packets but don't want to frequent
          regular reporting for audio could configure the T_rr_interval to a
          value so that the regular reporting for both audio and video is at a
          level that is considered acceptable for the audio. Then use feedback
          packets, which will include RTCP SR/RR packets, unless <xref
          target="RFC5506">reduced-size RTCP feedback packets</xref> are used,
          and can include other report information in addition to the feedback
          packet that needs to be sent. That way the available RTCP bandwidth
          can be focused for use, which provides the most utility for the
          application. </t>

          <t>Using T_rr_interval still requires one to determine suitable
          values for the RTCP bandwidth value, in fact it might make it even
          more important, as one is more likely to affect the RTCP behaviour
          and performance, than when using RTP/AVP, as their is fewer limitations
          affecting the RTCP transmission. </t>

          <t>When using T_rr_interval, i.e. having it be non zero, there are
          configurations that have to be avoided. If the resulting Td value is
          smaller but close to T_rr_interval then the interval in which the
          actual regular RTCP packet transmission falls into becomes very
          large, from 0.5 times T_rr_interval up to 2.73 times the
          T_rr_interval. Therefore for configuration where one intends to
          have Td smaller than T_rr_interval, then Td is RECOMMENDED to be
          targeted at values less than 1/4th of T_rr_interval which results in
          that the range becomes [0.5*T_rr_interval,
          1.81*T_rr_interval].</t>

        <t>With RTP/AVPF using T_rr_interval of 0 or with another low value,
          which will be significantly lower than Td still has its utility and
          different behaviour compared to RTP/AVP. This avoids the Tmin limitations
          of RTP/AVP, thus allowing more frequent regular RTCP reporting. In fact
          this will result that the RTCP traffic becomes as high as the
          configured values.</t>

          <t>
            (tbd: a future version of this memo will include examples of how
            to choose RTCP parameters for common scenarios)
          </t>

          <t>There exist no method within the specification for using
          different regular RTCP reporting interval depending on media type or
          individual media stream. </t>
        </section>
      </section>
    </section>

    <section title="Extension Considerations">
      <t>This section discusses the impact on some RTP/RTCP extensions due to
      usage of multiple media types in on RTP session. Only extensions where
      something worth noting has been included.</t>

      <section title="RTP Retransmission">
        <t>SSRC-multiplexed <xref target="RFC4588">RTP retransmission</xref>
        is actually very straightforward. Each retransmission RTP payload type
        is explicitly connected to an associated payload type. If
        retransmission is only to be used with a subset of all payload types,
        this is not a problem, as it will be evident from the retransmission
        payload types which payload types that have retransmission enabled for
        them.</t>

        <t>Session-multiplexed RTP retransmission is also possible to use
        where an retransmission session contains the retransmissions of the
        associated payload types in the source RTP session. The only
        difference to previously is that the source RTP session is one which
        contains multiple media types. Thus it is even more likely that only a
        subset of the source RTP session's payload types and SSRCs are
        actually retransmitted.</t>

        <t>Open Issue: When using SDP to signal retransmission for one RTP
        session with multiple media types and one RTP session for the
        retransmission data will cause a situation where one will have
        multiple m= lines grouped using FID and the ones belonging to
        respective RTP session being grouped using BUNDLE. This usage might
        contradict both the <xref target="RFC5888">FID semantics</xref> and an
        assumption in the <xref target="RFC4588">RTP retransmission
        specification</xref>.</t>
      </section>

      <section anchor="sec-generic-fec" title="Generic FEC">
        <t>The <xref target="RFC5109">RTP Payload Format for Generic Forward
        Error Correction</xref>, and also its predecessor <xref
        target="RFC2733"/>, requires some considerations, and they are
        different depending on what type of configuration of usage one
        has.</t>

        <t>Independent RTP Sessions, i.e. where source and repair data are
        sent in different RTP sessions. As this mode of configuration requires
        different RTP session, there has to be at least one RTP session for
        source data, this session can be one using multiple media types. The
        repair session only needs one RTP Payload type indicating repair data,
        i.e. x/ulpfec or x/parityfec depending if RFC 5109 or RFC 2733 is
        used. The media type in this session is not relevant and can in theory
        be any of the defined ones. It is RECOMMENDED that one uses
        "Application".</t>

        <t>In stream, using <xref target="RFC2198">RTP Payload for Redundant
        Audio Data</xref> combining repair and source data in the same
        packets. This is possible to use within a single RTP session. However,
        the usage and configuration of the payload types can create an issue.
        First of all it might be necessary to have one payload type per media
        type for the FEC repair data payload format, i.e. one for audio/ulpfec
        and one for text/ulpfec if audio and text are combined in an RTP
        session. Secondly each combination of source payload and its FEC
        repair data has to be an explicit configured payload type. This has
        potential for making the limitation of RTP payload types available
        into a real issue.</t>
      </section>
    </section>

    <section title="Signalling">
      <t>The Signalling requirements</t>

      <t>Establishing an RTP session with multiple media types requires
      signalling. This signalling needs to fulfil the following
      requirements:<list style="numbers">
          <t>Ensure that any participant in the RTP session is aware that this
          is an RTP session with multiple media types.</t>

          <t>Ensure that the payload types in use in the RTP session are using
          unique values, with no overlap between the media types.</t>

          <t>Configure the RTP session level parameters, such as RTCP RR and
          RS bandwidth, AVPF trr-int, underlying transport, the RTCP
          extensions in use, and security parameters, commonly for the RTP
          session.</t>

          <t>RTP and RTCP functions that can be bound to a particular media
          type SHOULD be reused when possible also for other media types,
          instead of having to be configured for multiple code-points. Note:
          In some cases one will not have a choice but to use multiple
          configurations.</t>
        </list></t>

      <t/>

      <section title="SDP-Based Signalling">
        <t>The signalling of multiple media types in one RTP session in SDP is
        specified in <xref
        target="I-D.ietf-mmusic-sdp-bundle-negotiation">"Multiplexing
        Negotiation Using Session Description Protocol (SDP) Port
        Numbers"</xref>.</t>
      </section>
    </section>

    <section anchor="IANA" title="IANA Considerations">
      <t>This document makes no request of IANA.</t>

      <t>Note to RFC Editor: this section is to be removed on publication as an
      RFC.</t>
    </section>

    <section anchor="Security" title="Security Considerations">
      <t>Having an RTP session with multiple media types doesn't change the
      methods for securing a particular RTP session. One possible difference
      is that the different media have often had different security
      requirements. When combining multiple media types in one session, their
      security requirements also have to be combined by selecting the most
      demanding for each property. Thus having multiple media types can result
      in increased overhead for security for some media types to ensure that
      all requirements are meet.</t>

      <t>Otherwise, the recommendations for how to configure and RTP session
      do not add any additional requirements compared to normal RTP, except
      for the need to be able to ensure that the participants are aware that
      it is a multiple media type session. If not that is ensured it can cause
      issues in the RTP session for both the unaware and the aware one.
      Similar issues can also be produced in an normal RTP session by creating
      configurations for different end-points that doesn't match each
      other.</t>
    </section>

    <section anchor="Acknowledgements" title="Acknowledgements">
      <t>The authors would like to thank Christer Holmberg, Gunnar
      Hellström, and Charles Eckel for the feedback on the document.</t>
    </section>
  </middle>

  <back>
    <references title="Normative References">
      <?rfc include="reference.RFC.2119"?>

      <?rfc include='reference.RFC.3550'?>

      <?rfc include='reference.RFC.3551'?>

      <?rfc include='reference.I-D.ietf-mmusic-sdp-bundle-negotiation'?>

      <?rfc include='reference.I-D.lennox-avtcore-rtp-multi-stream'?>
    </references>

    <references title="Informative References">
      <?rfc include='reference.I-D.westerlund-avtcore-multiplex-architecture'?>

      <?rfc include='reference.I-D.westerlund-avtcore-transport-multiplexing'?>

      <?rfc include='reference.I-D.lennox-payload-ulp-ssrc-mux'?>

      <?rfc include='reference.RFC.2198'?>

      <?rfc include='reference.RFC.2733'?>

      <?rfc include='reference.RFC.4566'?>

      <?rfc include='reference.RFC.4585'?>

      <?rfc include='reference.RFC.4588'?>

      <?rfc include='reference.RFC.5109'?>

      <?rfc include='reference.RFC.5117'?>

      <?rfc include='reference.RFC.5124'?>

      <?rfc include='reference.RFC.5506'?>

      <?rfc include='reference.RFC.5576'?>

      <?rfc include='reference.RFC.5761'?>

      <?rfc include='reference.RFC.5888'?>
    </references>
  </back>
</rfc>

PAFTECH AB 2003-20262026-04-23 20:41:48