Internet Engineering Task Force                                   AVT WG
INTERNET-DRAFT                                              Ladan Gharai
Intended status: Standards Track                            22 July 2007
draft-ietf-avt-tfrc-profile-10.txt
Expires: January 2008


                   RTP with TCP Friendly Rate Control


Status of this Memo

   By submitting this Internet-Draft, each author represents that any
   applicable patent or other IPR claims of which he or she is aware
   have been or will be disclosed, and any of which he or she becomes
   aware will be disclosed, in accordance with Section 6 of BCP 79.

   Internet-Drafts are working documents of the Internet Engineering
   Task Force (IETF), its areas, and its working groups. Note that other
   groups may also distribute working documents as Internet-Drafts.

   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."

   The list of current Internet-Drafts can be accessed at
   http://www.ietf.org/ietf/1id-abstracts.txt

   The list of Internet-Draft Shadow Directories can be accessed at
   http://www.ietf.org/shadow.html.

Copyright Notice

   Copyright (C) The IETF Trust  (2007).

Abstract

   This memo specifies how the TCP Friendly Rate Control (TFRC) of RTP
   flows can be supported using the RTP/AVPF profile and the general RTP
   header extension mechanism.  AVPF feedback packets and RTP header
   extensions are defined to support the exchange of control information
   between RTP TFRC senders and receivers. TFRC is an equation-based
   congestion control scheme for unicast flows operating in a best
   effort Internet environment.



Gharai                                                          [Page 1]


INTERNET-DRAFT            Expires: January 2008                July 2007


1.  Introduction

   [Note to RFC Editor: All references to RFC XXXX are to be replaced
   with the RFC number of this memo, when published]

   This memo specifies how the TCP Friendly Rate Control (TFRC) of RTP
   flows can be supported using the RTP/AVPF [RFC3550][RFC4585] profile
   and RTP header extensions, by defining a new header extension and
   AVPF feedback packet, and related parameters. Any of the AVPF based
   RTP profiles, such as SAVPF, can be used to support TFRC RTP flows.

   TFRC is an equation-based congestion control scheme for unicast flows
   operating in a best effort Internet environment and competing with
   TCP traffic. TFRC computes a TCP-friendly data rate based on current
   network conditions, as represented by the latest round trip time and
   packet loss calculations. The complete TFRC mechanism is described in
   detail in [RFC3448bis].

   To calculate a TCP-friendly data rate and keep track of round trip
   times and packet losses, TFRC senders and receivers rely on
   exchanging specific information between each other, i.e: the sender
   provides the receiver with the latest updates to round trip time
   calculations, while the receiver provides feedback needed to compute
   round trip times and on packet losses. This memo defines how this
   information can be exchanged between TFRC senders and receiver with
   RTP header extensions and an AVPF feedback packet.

2.  Terminology

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

3.  Relation to the Datagram Congestion Control Protocol

   The TFRC congestion control mechanism is one of a set of congestion
   control methods provided by the Datagram Congestion Control Protocol
   (DCCP) [RFC4340] In this section we detail the pros and cons of using
   TFRC with RTP versus DCCP.

   DCCP is a minimal general purpose transport-layer protocol with
   unreliable yet congestion controlled packet delivery semantics and
   reliable connection setup and teardown [RFC4336]. DCCP currently
   supports both TFRC [RFC4342] and TCP-like [RFC4341] congestion
   control, and the protocol is structured to support new congestion
   control mechanisms defined in the future.  A DCCP mapping for RTP has
   been standardized for media applications [RFCxRTP]. In addition DCCP
   supports a host of other features, such as: use of Explicit



Gharai                                                          [Page 2]


INTERNET-DRAFT            Expires: January 2008                July 2007


   Congestion Notification (ECN) and the ECN Nonce, flexible options
   processing, reliable option negotiation, DTLS [ID.DTLS], Path Maximum
   Transfer Unit (PMTU) and Service Codes.  Naturally an application
   using RTP/DCCP as its transport protocol will benefit from the
   protocol features supported by DCCP.

   However there are a number of benefits to be gained by the
   development and standardization of the use RTP with TFRC:

     o Media applications lacking congestion control can incorporate
       congestion controlled transport without delay by using RTP with
       TFRC. Widespread deployment of the DCCP protocol is not currently
       in place.

     o Use of RTP with TFRC is not contingent on any OS level changes
       and can be quickly deployed, because RTP is implemented at the
       application layer.

     o RTP/UDP flows face the same restrictions in firewall traversal
       as do UDP flows and do not require NATs and firewall
       modifications.   DCCP flows, on the other hand, do require NAT
       and firewall modifications, however once these modifications are
       in place, they can result in easier NAT and firewall traversal
       for RTP/DCCP flows in the future.

     o Use of RTP with TFRC with various media applications will give
       researchers, implementors and developers a better understanding
       of the intricate relationship between media quality and
       equation-based congestion control.  Hopefully this experience
       with congestion control and TFRC will ease the migration of media
       applications to DCCP once DCCP is deployed.

   Using the AVPF/RTP profile and header extension to support TFRC
   provides an immediate means for congestion control in media streams,
   in the time until DCCP is deployed.

   Additionally, there are also a number of technical differences as to
   how (and which) congestion control information is exchanged between
   DCCP with CCID3 and RTP:

     o Using header extensions the RTP TFRC sender transmits a
       send timestamp to the RTP TFRC receiver with every data packet.
       In addition to congestion control the send timestamp can be
       used by the receiver for jitter calculations.

       In contrast DCCP with CCID3 transmits a quad round trip
       counter to the receiver.




Gharai                                                          [Page 3]


INTERNET-DRAFT            Expires: January 2008                July 2007


     o The RTP TFRC receiver only provides the RTP TFRC sender
       with the loss event rate as computed by the receiver.

       In contrast DCCP with CCID3, provides 2 other options for the
       transport of loss event rate. A sender may choose to receive
       loss intervals or an Ack Vector. These two options provide the
       sender with the necessary information to compute the loss event
       rate.

     o Sequence number: DCCP supports a 48 bit and a 24 bit sequence
       number, whereas RTP only supports a 16 bit sequence number. While
       this makes RTP susceptible to data injection attacks, it can be
       avoided by using the SRTP [RFC3711] profile.


4.  The TFRC Information Exchange Loop

   TFRC depends on the exchange of congestion control information
   between a sender and receiver.  In this section we reiterate which
   items are exchanged between a TFRC sender and receiver as discussed
   in [RFC3448bis]. We note how RTP can accommodates these exchanges.

4.1.  Data Packets

   As stated in [RFC3448bis] a TFRC sender transmits the following
   information in each data packet to the receiver:

    o A sequence number, incremented by one for each data packet
      transmitted.

    o A timestamp indicating the packet send time and the sender's
      current estimate of the round-trip time, RTT. This information
      is then used by the receiver to compute the TFRC loss intervals.
      - or -
      A course-grained timestamp incrementing every quarter of a
      round trip time, which is then used to determine the TFRC loss
      intervals.

   The standard RTP sequence number suffices for the functionality
   provided by TFRC.  A RTP header extension [hdrtxt] is used to
   transmit the send timestamp and RTT.  This extension is defined in
   Section 5.


4.2.  Feedback Packets

   As stated in [RFC3448bis] a TFRC receiver provides the following
   feedback to the sender at least once per RTT or per data packet



Gharai                                                          [Page 4]


INTERNET-DRAFT            Expires: January 2008                July 2007


   received (which ever time interval is larger):

    o The send timestamp of the last data packet received, t_i.

    o The amount of time elapsed between the receipt of the last
      data packet at the receiver, and the generation of this feedback
      report, t_delay. This is used by the sender for RTT computations.

    o The rate at which the receiver estimates that data was received
      since the last feedback report was sent, x_recv.

    o The receiver's current estimate of the loss event rate, p, a real
      value between 0 and 1.0.

   To accommodate the feedback of these values a new AVPF transport
   layer feedback message is defined, as detailed in Section 6. The
   timing interval between the feedback packets is discussed in Section
   7.




5.  The Header Extension

   The form of the extension block is depicted in Figure 1. The length
   field for the extension takes the value 6 to indicate that the
   payload is 7 bytes. Two header extension fields are defined and used
   as follows:


   Send timestamp (t_i): 32 bits
     The timestamp indicating when the packet is sent. This timestamp
     is measured in microseconds and is used for round trip time
     calculations.

   Round trip time (RTT): 24 bits
     The round trip time as measured by the RTP TFRC sender in
     microseconds.













Gharai                                                          [Page 5]


INTERNET-DRAFT            Expires: January 2008                July 2007


       0                   1                   2                   3
       0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |      0xBE     |      0xDE     |            length=2           |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |  ID   | len=6 |                RTT                            |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                     send timestamp  (t_i)                     |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
     Figure 1: RTP Sender Header Extension Block



6.  TFRC-FB: A New AVPF Transport Layer Feedback Message

   A new transport layer AVPF feedback message is defined to support
   feedback from the receivers: TFRC-FB.  Figure 2 depicts the both the
   common packet format (the first 12 octets)  and the feedback control
   information (FCI) for the TFRC-FB packet.

   We note that the TFRC related feedback, is specific to one media
   stream sender, therefore all messages in the compound RTCP packet
   MUST share the same media source SSRC. In the case where a sender is
   sending multiple media streams to a receiver, each media flow will be
   allocated its own AVPF feedback flow.

   We define four FCI fields for the TFRC-FB message as follows:

   Send timestamp (t_i): 32 bits
     The send timestamp of the last data packet received by the
     RTP TFRC receiver, t_i, in microseconds.

   Delay (t_delay): 32 bits
     The amount of time elapsed between the receipt of the last data
     packet at the RTP TFRC receiver, and the generation of this
     feedback report in microseconds. This is used by the RTP TFRC
     sender for RTT computations.

   Data rate (X_recv): 32 bits
     The rate at which the receiver estimates that data was received
     since the last feedback report was sent in bytes per second. X_recv
     is computed per [RFC3448bis].

   Loss event rate (p): 32 bits
     The receiver's current estimate of the loss event rate, p,
     expressed as a fixed point number with the binary point at the
     left edge of the field. (That is equivalent to taking the integer
     part after multiplying the loss event rate by 2^32.) The value of



Gharai                                                          [Page 6]


INTERNET-DRAFT            Expires: January 2008                July 2007


     the loss event rate is computed per [RFC3448bis] Section 5.

      0                   1                   2                   3
       0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |V=2|P|  FMT=2  |  PT=RTPFB     |             length            |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                     SSRC of packet sender                     |
      +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+
      |                     SSRC of media source                      |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                      send timestamp (t_i)                     |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                      delay  (t_delay)                         |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                  data rate at the receiver (x_recv)           |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                    loss event rate (p)                        |
      +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+
     Figure 2: The AVPF TFRC-FB RTCP Transport Layer Feedback Message



7.  RTCP Transmission Intervals and Bandwidth Requirements

   When using TFRC rate controlled RTP, the RTCP transmission intervals
   must be set according to the requirements of the TFRC algorithm. TFRC
   requires a receiver to generate a feedback ack packet at least once
   per RTT or per packet received (based on the larger time interval).
   These requirements are to ensure timely reaction to congestion.

   The TFRC requirements of receiving feedback once per RTT can at times
   conflict with the AVP RTCP bandwidth constraints, particularly at
   small RTTs of 20 ms or less.  Assuming only one TFRC-FB report per
   RTCP compound packet, Table 1 lists the RTCP bandwidths at RTTs of 2,
   5, 10 and 20 ms and the minimum corresponding RTP data rates, where
   RTCP(X) <= (0.05)*RTP(X) is true.   For example, according to Table
   1, a TFRC RTP flow of less than 3.2 Mbps and a RTT of 5 ms, can not
   comply with the 5% RTCP bandwidth constraints (Table 1 assumes each
   RTCP packet is 100 bytes).   RTP flows facing such circumstances
   should take into account the additional RTCP bandwidth needed when
   signaling their bandwidth information in SDP [RFC4566].

   Based on initial assumptions on round trip time if more than the
   recommended 5% is needed for RTCP bandwidth, the applications SHOULD
   use the SDP bandwidth modifiers RS and RR [RFC3556] to signal the
   amount of RTCP bandwidth needed. If the round trip time assumptions
   change after the RTP flows start running, the application MAY



Gharai                                                          [Page 7]


INTERNET-DRAFT            Expires: January 2008                July 2007


   recalculate the amount of RTCP bandwidth needed and re-signal this
   new value using its signaling protocol of choice.


                        RTT      RTCP(X)   RTP(X)
                    +------------------------------+
                    |  20 ms |  40 kbps | 0.8 Mbps |
                    |  10 ms |  80 kbps | 1.6 Mbps |
                    |   5 ms | 160 kbps | 3.2 Mbps |
                    |   2 ms | 400 kbps | 8.0 Mbps |
                    +------------------------------+
     Table 1: RTCP bandwidth for TFRC flows with corresponding
     RTTs of 20, 10, 5 and 2 ms. Assuming, 100 byte RTCP packets
     and one RTCP packet per RTT.


Additionally, to support the transmission of a feedback packet once per
RTT,  the AVPF T_rr_interval variable MUST NOT be set to a value larger
than the current round trip time, RTT, as this would prevent generating
feedback packets at least once per RTT (see RFC 4585, Section 3.4,m).


8.  SDP Usage

   RTP flows using TFRC congestion control MUST signal their use of the
   AVPF profile and RTCP feedback packets, the round trip time (RTT) and
   send timestamp extension, and MAY also signal an initial RTCP
   bandwidth usage:

           v=0
           o=alice 2890844526 2890844526 IN IP4 host.example.com
           s=congestion control with TFRC
           c=IN IP4 host.example.com
           m=video 5400 RTP/AVPF 112
           a=rtpmap:112 H261/90000
           a=extmap:4 urn:ietf:params:rtp-hdtext:rtt-sendts
           a=rtcp-fb * tfrc
           b=AS:400
           b=RS:800
           b=RR:4000



8.1.  Usage with the SDP Offer/Answer Model

   TBC





Gharai                                                          [Page 8]


INTERNET-DRAFT            Expires: January 2008                July 2007


9.  IANA Considerations

   In this section we detail IANA registry values that need to be
   registered. Two new values must be reistered for the AVPF profile:

   The new RTP/AVPF feedback packet, TFRC-FB.  The following format
   (FMT) values must be registerd in the FMT sub-registry of the RTPFB
   payload type:

     Value name:  TFRC-FB
     Long name:   TFRC feedback
     Value:  5
     Reference:   RFC XXXX

   The new rtcp-fd-id "tfrc" must be registered with the "rtcp-fb"
   attribute registry:

     Value name:     tfrc
     Long name:      TFRC Feedback
     Reference:      RFC XXXX

   For the new header extension, the name rtt-sendts must be registered
   into the rtp-hdrext section of the urn:ietf: namespace, referring to
   RFC XXXX.


10.  Security Considerations

   This memo defines how to use the RTP AVPF profile and the general RTP
   header extensions to support TFRC congestion control. Therefore RTP
   packets using these mechanisms are subject to the security
   considerations discussed in the RTP specification [RFC3550], the
   RTP/AVPF profile specification [RFC4585] and the general header
   extensions mechanism [hdrtxt]. Combining these mechanisms does not
   pose any additional security implications.  Applications requiring
   authentication and integrity protection, or applictions operating in
   environments that must strictly adhere to the TFRC send rate (and
   fear manipulation of the feedback messsages) can use the SAVPF
   [SAVPF] profile.


11.  Acknowledgments

   This memo is based upon work supported by the U.S. National Science
   Foundation (NSF) under Grant No. 0334182. Any opinions, findings and
   conclusions or recommendations expressed in this material are those
   of the authors and do not necessarily reflect the views of NSF.




Gharai                                                          [Page 9]


INTERNET-DRAFT            Expires: January 2008                July 2007


12.  Author's Address

     Ladan Gharai <ladan@gharai.org>


Normative References

   [RFC2119] S. Bradner, "Key words for use in RFCs to Indicate
             Requirement Levels", Internet Engineering Task Force,
             RFC 2119, March 1997.

   [RFC3550] H. Schulzrinne, S. Casner, R. Frederick and V. Jacobson,
             "RTP: A Transport Protocol for Real-Time Applications",
             Internet Engineering Task Force, RFC 3550 (STD0064), July
             2003.

   [RFC3556] S. Casner, "Session Description Protocol (SDP) Bandwidth
             Modifiers for RTP Control Protocol (RTCP) Bandwidth", RFC
             3556, July 2003.

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

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

   [RFC4585] J. Ott, S. Wenger, A. Sato, C. Burmeister and J. Ray,
             "Extended RTP Profile for RTCP-based Feedback (RTP/AVPF)",
               RFC 4585, July 2006.

   [RFC3448bis]  M. Handley, S. Floyed, J. Padhye and J. widmer,
             "TCP Friendly Rate Control (TRFC): Protocol Specification",
             draft-ietf-dccp-rfc3448bis-01.txt, March 2007.

   [hdrext]  D. Singer, "A general mechanism for RTP Header Extensions",
          IETF Work in Progress, draft-ietf-avt-rtp-hdrext-12.txt.

   [SAVPF]   J. Ott, E. Carrara, "Extended Secure RTP Profile for
             RTCP-based Feedback (RTP/SAVPF)," IETF Work in Progress,
             draft-ietf-avt-profile-savpf-10.txt.


Informative References

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



Gharai                                                         [Page 10]


INTERNET-DRAFT            Expires: January 2008                July 2007


   [RFC4336] Floyd, S., Handley, M., and E. Kohler, "Problem Statement
             for the Datagram Congestion Control Protocol (DCCP)", RFC
             4336, March 2006.

   [RFC4340] Kohler, E., Handley, M., and S. Floyd, "Datagram Congestion
             Control Protocol (DCCP)", RFC 4340, March 2006.

   [RFC4341] Floyd, S. and E. Kohler, "Profile for Datagram Congestion
             Control Protocol (DCCP) Congestion Control ID 2: TCP-like
             Congestion Control", RFC 4341, March 2006.

   [RFC4342] Floyd, S., Kohler, E., and J. Padhye, "Profile for Datagram
             Congestion Control Protocol (DCCP) Congestion Control ID 3:
             TCP-Friendly Rate Control (TFRC)", RFC 4342, March 2006.

   [RFCxRTP] C. Perkins, "RTP and the Datagram Congestion Control
             Protocol (DCCP)", IETF Work in Progress, draft-ietf-dccp-
             rtp-07.txt.

   [ID.DTLS] T.Phelan, "Datagram Transport Layer Security (DTLS)
             over the Datagram Congestion Control Protocol (DCCP)", IETF
             Work in Progress, draft-ietf-dccp-dtls-00.txt.

13.  IPR Notice

   The IETF takes no position regarding the validity or scope of any
   Intellectual Property Rights or other rights that might be claimed to
   pertain to the implementation or use of the technology described in
   this document or the extent to which any license under such rights
   might or might not be available; nor does it represent that it has
   made any independent effort to identify any such rights.  Information
   on the procedures with respect to rights in RFC documents can be
   found in BCP 78 and BCP 79.

   Copies of IPR disclosures made to the IETF Secretariat and any
   assurances of licenses to be made available, or the result of an
   attempt made to obtain a general license or permission for the use of
   such proprietary rights by implementers or users of this
   specification can be obtained from the IETF on-line IPR repository at
   http://www.ietf.org/ipr.

   The IETF invites any interested party to bring to its attention any
   copyrights, patents or patent applications, or other proprietary
   rights that may cover technology that may be required to implement
   this standard.  Please address the information to the IETF at ietf-
   ipr@ietf.org.





Gharai                                                         [Page 11]


INTERNET-DRAFT            Expires: January 2008                July 2007


14.  Full Copyright Statement

   Copyright (C) The IETF Trust (2007).

   This document is subject to the rights, licenses and restrictions
   contained in BCP 78, and except as set forth therein, the authors
   retain all their rights.

   This document and the information contained herein are provided on an
   "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS
   OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND
   THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS
   OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF
   THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED
   WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.




































Gharai                                                         [Page 12]