Mapping RTP streams to CLUE media captures
draft-even-clue-rtp-mapping-03

CLUE WG                                                          R. Even
Internet-Draft                                       Huawei Technologies
Intended status: Standards Track                               J. Lennox
Expires: January 17, 2013                                          Vidyo
                                                           July 16, 2012

               Mapping RTP streams to CLUE media captures
                   draft-even-clue-rtp-mapping-03.txt

Abstract

   This document describes mechanisms and recommended practice for
   mapping RTP media streams defined in SDP to CLUE media captures.

Status of this Memo

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

   Internet-Drafts are working documents of the Internet Engineering
   Task Force (IETF).  Note that other groups may also distribute
   working documents as Internet-Drafts.  The list of current Internet-
   Drafts is at http://datatracker.ietf.org/drafts/current/.

   Internet-Drafts are draft documents valid for a maximum of six months
   and may be updated, replaced, or obsoleted by other documents at any
   time.  It is inappropriate to use Internet-Drafts as reference
   material or to cite them other than as "work in progress."

   This Internet-Draft will expire on January 17, 2013.

Copyright Notice

   Copyright (c) 2012 IETF Trust and the persons identified as the
   document authors.  All rights reserved.

   This document is subject to BCP 78 and the IETF Trust's Legal
   Provisions Relating to IETF Documents
   (http://trustee.ietf.org/license-info) in effect on the date of
   publication of this document.  Please review these documents
   carefully, as they describe your rights and restrictions with respect
   to this document.  Code Components extracted from this document must
   include Simplified BSD License text as described in Section 4.e of
   the Trust Legal Provisions and are provided without warranty as
   described in the Simplified BSD License.

Even & Lennox           Expires January 17, 2013                [Page 1]
Internet-Draft             RTP mapping to CLUE                 July 2012

Table of Contents

   1.  Introduction . . . . . . . . . . . . . . . . . . . . . . . . .  3
   2.  Terminology  . . . . . . . . . . . . . . . . . . . . . . . . .  3
   3.  RTP topologies for CLUE  . . . . . . . . . . . . . . . . . . .  3
   4.  Mapping CLUE Media Captures to RTP streams . . . . . . . . . .  5
     4.1.  Static Mapping . . . . . . . . . . . . . . . . . . . . . .  6
     4.2.  Dynamic mapping  . . . . . . . . . . . . . . . . . . . . .  7
     4.3.  Recommendations  . . . . . . . . . . . . . . . . . . . . .  7
   5.  Application to CLUE Media Requirements . . . . . . . . . . . .  7
   6.  Examples . . . . . . . . . . . . . . . . . . . . . . . . . . .  9
   7.  Acknowledgements . . . . . . . . . . . . . . . . . . . . . . .  9
   8.  IANA Considerations  . . . . . . . . . . . . . . . . . . . . .  9
   9.  Security Considerations  . . . . . . . . . . . . . . . . . . . 10
   10. References . . . . . . . . . . . . . . . . . . . . . . . . . . 10
     10.1. Normative References . . . . . . . . . . . . . . . . . . . 10
     10.2. Informative References . . . . . . . . . . . . . . . . . . 10
   Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 11

Even & Lennox           Expires January 17, 2013                [Page 2]
Internet-Draft             RTP mapping to CLUE                 July 2012

1.  Introduction

   Telepresence systems can send and receive multiple media streams.
   The CLUE framework [I-D.ietf-clue-framework] defines media captures
   as a source of Media, such as from one or more Capture Devices.  A
   Media Capture (MC) may be the source of one or more Media streams.  A
   Media Capture may also be constructed from other Media streams.  A
   middle box can express Media Captures that it constructs from Media
   streams it receives.

   SIP offer answer [RFC3264] uses SDP [RFC4566] to describe the
   RTP[RFC3550] media streams.  Each RTP stream has a payload type
   number and SSRC.  The content of the RTP stream is created by the
   encoder in the endpoint.  This may be an original content from a
   camera or a content created by an intermediary device like an MCU.

   This document makes recommendations, for this telepresence
   architecture, about how RTP and RTCP streams should be encoded and
   transmitted, and how their relation to CLUE Media Captures should be
   communicated.  The proposed solution supports multiple RTP topologies

2.  Terminology

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

3.  RTP topologies for CLUE

   The typical RTP topologies used by telepresence systems specify
   different behaviors for RTP and RTCP distribution.  The relevant
   topologies include point-to-point, as well as media mixers, media-
   switching mixers, and source-projection mixers.

   In the point-to-point topology, one peer communicates directly with a
   single peer over unicast.  There can be one or more RTP sessions, and
   each RTP session can carry multiple RTP streams identified by their
   SSRC.  All SSRCs will be recognized by the peers based on the
   information in the RTCP SDES report that will include the CNAME and
   SSRC of the sent RTP streams.  There are different use point to point
   ise cases as specified in CLUE use case
   [I-D.ietf-clue-telepresence-use-cases].  There may be a difference
   between the symmetric and asymmetric use cases.  While in the
   symmetric use case the typical mapping will be from a Media capture
   device to a render device (e.g. camera to monitor) in the asymmetric

Even & Lennox           Expires January 17, 2013                [Page 3]
Internet-Draft             RTP mapping to CLUE                 July 2012

   case the render device may receive different capture information (RTP
   stream from a different camera) if it has fewer rendering devices
   (monitors).  In some cases, a CLUE session which, at a high-level, is
   point-to-point may nonetheless have RTP which is best described by
   one of the mixer topologies below.  For example, a CLUE endpoint can
   produce composited or switched captures for use by a receiving system
   with fewer displays than the sender has cameras.

   In the Media Mixer topology, the peers communicate only with the
   mixer.  The mixer provides mixed or composited media streams, using
   its own SSRC for the sent streams.  There are two cases here.  In the
   first case the mixer may have separate RTP sessions with each peer
   (similar to the point to point topology) terminating the RTCP
   sessions on the mixer; this is known as Topo-RTCP-Terminating MCU in
   [RFC5117].  In the second case, the mixer can use a conference-wide
   RTP session similar to RFC 5117's Topo-mixer or Topo-Video-switching.
   The major difference is that for the second case, the mixer uses
   conference-wide RTP sessions, and distributes the RTCP reports to all
   the RTP session participants, enabling them to learn all the CNAMEs
   and SSRCs of the participants and know the contributing source or
   sources (CSRCs) of the original streams from the RTP header.  In the
   first case, the Mixer terminates the RTCP and the participants cannot
   know all the available sources based on the RTCP information.  The
   conference roster information including conference participants,
   endpoints, media and media-id (SSRC) can be available using the
   conference event package [RFC4575] element.

   In the Media-Switching Mixer topology, the peer to mixer
   communication is unicast with mixer RTCP feedback.  It is
   conceptually similar to a compositing mixer as described in the
   previous paragraph, except that rather than compositing or mixing
   multiple sources, the mixer provides one or more conceptual sources
   selecting one source at a time from the original sources.  The Mixer
   creates a conference-wide RTP session by sharing remote SSRC values
   as CSRCs to all conference participants.

   In the Source-Projection Mixer topology, the peer to mixer
   communication is unicast with RTCP mixer feedback.  Every potential
   sender in the conference has a source which is "projected" by the
   mixer into every other session in the conference; thus, every
   original source is maintained with an independent RTP identity to
   every receiver, maintaining separate decoding state and its original
   RTCP SDES information.  However, RTCP is terminated at the mixer,
   which might also perform reliability, repair, rate adaptation, or
   transcoding on the stream.  Senders' SSRCs may be renumbered by the
   mixer.  The sender may turn the projected sources on and off at any
   time, depending on which sources it thinks are most relevant for the
   receiver; this is the primary reason why this topology must act as an

Even & Lennox           Expires January 17, 2013                [Page 4]
Internet-Draft             RTP mapping to CLUE                 July 2012

   RTP mixer rather than as a translator, as otherwise these disabled
   sources would appear to have enormous packet loss.  Source switching
   is accomplished through this process of enabling and disabling
   projected sources, with the higher-level semantic assignment of
   reason for the RTP streams assigned externally.

   The above topologies demonstrate two major RTP/RTCP behaviors:

   1.  The mixer may either use the source SSRC when forwarding RTP
       packets, or use its own created SSRC.  Still the mixer will
       distribute all RTCP information to all participants creating
       conference-wide RTP session/s.  This allows the participants to
       learn the available RTP sources in each RTP session.  The
       original source information will be the SSRC or in the CSRC
       depending on the topology.  The point to point case behaves like
       this.

   2.  The mixer terminates the RTCP from the source, creating separate
       RTP sessions with the peers.  In this case the participants will
       not receive the source SSRC in the CSRC.  Since this is usually a
       mixer topology, the source information is available from the SIP
       conference event package [RFC4575] Subscribing to the conference
       event package allows each participant to know the SSRCs of all
       sources in the conference.

4.  Mapping CLUE Media Captures to RTP streams

   The different topologies described in Section 3 support different
   SSRC distribution models and RTP stream multiplexing points.

   Most video conferencing systems today can separate multiple RTP
   sources by placing them into separate RTP sessions using, the SDP
   description.  For example, main and slides video sources are
   separated into separate RTP sessions based on the content attribute
   [RFC4796].  This solution works straightforwardly if the multiplexing
   point is at the UDP transport level, where each RTP stream uses a
   separate RTP session.  This will also be true for mapping the RTP
   streams to Media Captures if each media capture uses a separate RTP
   session, and the consumer can identify it based on the receiving RTP
   port.  In this case, SDP only needs to label the RTP session with an
   identifier that identifies the media capture in the CLUE description.
   In this case, it does not change the mapping even if the RTP session
   is switched using same or different SSRC.  (The multiplexing is not
   at the SSRC level).

   Even though Session multiplexing is supported by CLUE, for scaling
   reasons, CLUE recommends using SSRC multiplexing in a single or

Even & Lennox           Expires January 17, 2013                [Page 5]
Internet-Draft             RTP mapping to CLUE                 July 2012

   multiple sessions.  So we need to look at how to map RTP streams to
   Media Capture IDs when SSRC multiplexing is used.

   When looking at SSRC multiplexing we can see that in various
   topologies, the SSRC behavior may be different:

   1.  The SSRCs are static (assigned by the MCU/Mixer), and there is an
       SSRC for each media capture encoding defined in the CLUE
       protocol.  Source information may be conveyed using CSRC, or, in
       the case of topo-RTCP-Terminating MCU, is not conveyed.

   2.  The SSRCs are dynamic, representing the original source and are
       relayed by the Mixer/MCU to the participants.

   In the above two cases the MCU/Mixer creates its own advertisement,
   with a virtual room capture scene.

   Another case we can envision is that the MCU / Mixer relays all the
   capture scenes from all advertisements to all consumers.  This means
   that the advertisement will include multiple capture scenes, each
   representing a separate TP room with its own coordinate system.  A
   general tools for distributing roster information is by using an
   event package, for example by extending the conference event package.

4.1.  Static Mapping

   Static mapping is widely used in current MCU implementations.  It is
   also common for a point to point symmetric use case when both
   endpoints have the same capabilities.  For capture encodings with
   static SSRCs, it is most straightforward to indicate this mapping
   outside the media stream, in the CLUE or SDP signaling.  An SDP
   source attribute [RFC5576] could be defined to associate CLUE capture
   IDs with SSRCs in SDP.  Each SSRC will have a captureID value that
   will be specified also in the CLUE media capture as an attribute.
   The provider advertisement could, if it wished, use the same SSRC for
   media capture encodings that are mutually exclusive.  (This would be
   natural, for example, if two advertised captures are implemented as
   different configurations of the same physical camera, zoomed in or
   out.)

   Note: there may be more than one RTP session for a media capture like
   in simulcast.  We still need to figure out how to describe it in SDP
   and CLUE.

   Another method for static mapping may be to use the provider
   advertisement could to indicate the intended SSRC directly.  The
   advnatge of using the SDP SSRC attribute is that RFC5576 [RFC5576]
   the issue of SSRC collision and provide guideline how to address

Even & Lennox           Expires January 17, 2013                [Page 6]
Internet-Draft             RTP mapping to CLUE                 July 2012

   them.

4.2.  Dynamic mapping

   Dynamic mapping using RTP header extension is described in
   draft-lennox-clue-rtp-usage [I-D.lennox-clue-rtp-usage]section 10.2.
   The draft does not specify what is the capture id value.  As
   specified for the static case there should be a capture id attribute
   in the CLUE media capture information to enable this mapping.

4.3.  Recommendations

   The recommendation is that endpoints MUST support both the static
   declaration of capture encoding SSRCs, and the RTP header extension
   method of sharing capture IDs, with the extension in every media
   packet.  For low bandwidth situations, this may be considered
   excessive overhead; in which case endpoints MAY support the combined
   approach from [I-D.lennox-clue-rtp-usage].  The SDP offer MAY specify
   the SSRC mapping to media capture.  In the case of static mapping
   topologies there will be no need to use the header extensions in the
   media, since the SSRC for the RTP stream will remain the same during
   the call unless a collision is detected and handled according to
   RFC5576 [RFC5576].  If the used topology uses dynamic mapping then
   the RTP header extension will be used to indicate the RTP stream
   switch for the media capture.  In this case the SDP description may
   be used to negotiate the initial SSRC but this will be left for the
   implementation.  Note that if the SSRC is defined explicitly in the
   SDP the SSRC collision should be handled as in RFC5576.

5.   Application to CLUE Media Requirements

   [I-D.lennox-clue-rtp-usage] offers a number of requirements that are
   believed to be necessary for a CLUE RTP mapping.  The solutions
   described in this document are believed to meet that requirement,
   though some of them are only possible for some of the topologies.
   (Since the requirements are generally of the form "it must be
   possible for a sender to do something", this is adequate; a sender
   which wishes to perform that action needs to choose a topology which
   allows the behavior it wants.

   In this section we address only those requirements where the
   topologies or the association mechanisms treat the requirements
   differently.

   Media-4: It must be possible for an original source to move among
   switched captures (i.e. at one time be sent for one switched capture,
   and at a later time be sent for another one).

Even & Lennox           Expires January 17, 2013                [Page 7]
Internet-Draft             RTP mapping to CLUE                 July 2012

   This applies naturally for static sources with a Switched Mixer.  For
   dynamic sources with a Source-Projecting Mixer, this just requires
   the capture tag in the header extension element to be updated
   appropriately.

   Media-6: Whenever a given source is transmitted for a switched
   capture, it must be immediately possible for a receiver to determine
   the switched capture it corresponds to, and thus that any previous
   source is no longer being mapped to that switched capture.

   For a Switched Mixer, this applies naturally.  For a Source-
   Projecting mixer, this is done based on the header extension.

   Media-7: It must be possible for a receiver to identify the original
   source that is currently being mapped to a switched capture, and
   correlate it with out-of-band (non-Clue) information such as rosters.

   For a Switched Mixer, this is done based on the CSRC, if the mixer is
   providing CSRCs; if for a Source-Projecting Mixer, this is done based
   on the SSRC.

   Media-8: It must be possible for a source to move among switched
   captures without requiring a refresh of decoder state (e.g., for
   video, a fresh I-frame), when this is unnecessary.  However, it must
   also be possible for a receiver to indicate when a refresh of decoder
   state is in fact necessary.

   This can be done by a Source-Projecting Mixer, but not by a Switching
   Mixer.  The last requirement can be accomplished through an FIR
   message [RFC5104], though potentially a faster mechanism (not
   requiring a round-trip time from the receiver) would be preferable.

   Media-9: If a given source is being sent on the same transport flow
   to satisfy more than one capture (e.g. if it corresponds to more than
   one switched capture at once, or to a static capture as well as a
   switched capture), it should be possible for a sender to send only
   one copy of the source.

   For a Source-Projecting Mixer, this can be accomplished by sending
   multiple dynamic capture IDs for the same source; this can also be
   done for an environment with a hybrid of mixer topologies and static
   and dynamic captures, described below in Section 6.  It is not
   possible for static captures from a Switched Mixer.

   Media-12: If multiple sources from a single synchronization context
   are being sent simultaneously, it must be possible for a receiver to
   associate and synchronize them properly, even for sources that are
   are mapped to switched captures.

Even & Lennox           Expires January 17, 2013                [Page 8]
Internet-Draft             RTP mapping to CLUE                 July 2012

   For a Mixed or Switched Mixer topology, receivers will see only a
   single synchronization context (CNAME), corresponding to the mixer.
   For a Source-Projecting Mixer, separate projecting sources keep
   separate synchronization contexts based on their original CNAMEs,
   thus allowing independent synchronization of sources from independent
   rooms without needing global synchronization.  In hybrid cases,
   however (e.g. if audio is mixed), all sources which need to be
   synchronized with the mixed audio must get the same CNAME (and thus a
   mixer-provided timebase) as the mixed audio.

6.  Examples

   It is possible for a CLUE device to send multiple instances of the
   topologies in Section 3 simultaneously.  For example, an MCU which
   uses a traditional audio bridge with switched video would be a Mixer
   topology for audio, but a Switched Mixer or a Source-Projecting Mixer
   for video.  In the latter case, the audio could be sent as a static
   source, whereas the video could be dynamic.

   More notably, it is possible for an endpoint to send the same sources
   both for static and dynamic captures.  Consider the example in
   Section 11.1 of [I-D.ietf-clue-framework], where an endpoint can
   provide both three cameras (VC0, VC1, and VC2) for left, center, and
   right views, as well as a switched view (VC3) of the loudest panel.

   It is possible for a consumer to request both the (VC0 - VC2) set and
   VC3.  It is worth noting that the content of VC3 is, at all times,
   exactly the content of one of VC0, VC1, or VC2.  Thus, if the sender
   uses the Source-Selection Mixer topology for VC3, the consumer that
   receives these three sources would not need to send any additional
   media traffic over just sending (VC0 - VC2).

   In this case, the advertiser could describe VC0, VC1, and VC2 in its
   initial advertisement or SDP with static SSRCs, whereas VC3 would
   need to be dynamic.  The role of VC3 would move among VC0, VC1, or
   VC2, indicated by the RTP header extension on those streams' RTP
   packets.

7.  Acknowledgements

   place holder

8.  IANA Considerations

   TBD

Even & Lennox           Expires January 17, 2013                [Page 9]
Internet-Draft             RTP mapping to CLUE                 July 2012

9.  Security Considerations

   TBD.

10.  References

10.1.  Normative References

   [I-D.ietf-clue-framework]
              Romanow, A., Duckworth, M., Pepperell, A., and B. Baldino,
              "Framework for Telepresence Multi-Streams",
              draft-ietf-clue-framework-05 (work in progress),
              February 2012.

   [I-D.lennox-clue-rtp-usage]
              Lennox, J., Witty, P., and A. Romanow, "Real-Time
              Transport Protocol (RTP) Usage for Telepresence Sessions",
              draft-lennox-clue-rtp-usage-04 (work in progress),
              June 2012.

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

10.2.  Informative References

   [I-D.ietf-clue-telepresence-use-cases]
              Romanow, A., Botzko, S., Duckworth, M., Even, R., and I.
              Communications, "Use Cases for Telepresence Multi-
              streams", draft-ietf-clue-telepresence-use-cases-02 (work
              in progress), January 2012.

   [RFC3264]  Rosenberg, J. and H. Schulzrinne, "An Offer/Answer Model
              with Session Description Protocol (SDP)", RFC 3264,
              June 2002.

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

   [RFC4566]  Handley, M., Jacobson, V., and C. Perkins, "SDP: Session
              Description Protocol", RFC 4566, July 2006.

   [RFC4575]  Rosenberg, J., Schulzrinne, H., and O. Levin, "A Session
              Initiation Protocol (SIP) Event Package for Conference
              State", RFC 4575, August 2006.

   [RFC4796]  Hautakorpi, J. and G. Camarillo, "The Session Description

Even & Lennox           Expires January 17, 2013               [Page 10]
Internet-Draft             RTP mapping to CLUE                 July 2012

              Protocol (SDP) Content Attribute", RFC 4796,
              February 2007.

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

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

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

   [RFC5888]  Camarillo, G. and H. Schulzrinne, "The Session Description
              Protocol (SDP) Grouping Framework", RFC 5888, June 2010.

   [RFC6184]  Wang, Y., Even, R., Kristensen, T., and R. Jesup, "RTP
              Payload Format for H.264 Video", RFC 6184, May 2011.

   [RFC6236]  Johansson, I. and K. Jung, "Negotiation of Generic Image
              Attributes in the Session Description Protocol (SDP)",
              RFC 6236, May 2011.

Authors' Addresses

   Roni Even
   Huawei Technologies
   Tel Aviv,
   Israel

   Email: even.roni@huawei.com

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

   Email: jonathan@vidyo.com

Even & Lennox           Expires January 17, 2013               [Page 11]