Internet Engineering Task Force (IETF)                         R. Gilman
Request for Comments: 6871                                   Independent
Updates: 5939                                                    R. Even
Category: Standards Track                            Huawei Technologies
ISSN: 2070-1721                                             F. Andreasen
                                                          Cisco Systems
                                                          February 2013


  Session Description Protocol (SDP) Media Capabilities Negotiation

Abstract

  Session Description Protocol (SDP) capability negotiation provides a
  general framework for indicating and negotiating capabilities in SDP.
  The base framework defines only capabilities for negotiating
  transport protocols and attributes.  This documents extends the
  framework by defining media capabilities that can be used to
  negotiate media types and their associated parameters.

  This document updates the IANA Considerations of RFC 5939.

Status of This Memo

  This is an Internet Standards Track document.

  This document is a product of the Internet Engineering Task Force
  (IETF).  It represents the consensus of the IETF community.  It has
  received public review and has been approved for publication by the
  Internet Engineering Steering Group (IESG).  Further information on
  Internet Standards is available in Section 2 of RFC 5741.

  Information about the current status of this document, any errata,
  and how to provide feedback on it may be obtained at
  http://www.rfc-editor.org/info/rfc6871.
















Gilman, et al.               Standards Track                    [Page 1]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


Copyright Notice

  Copyright (c) 2013 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.

  This document may contain material from IETF Documents or IETF
  Contributions published or made publicly available before November
  10, 2008.  The person(s) controlling the copyright in some of this
  material may not have granted the IETF Trust the right to allow
  modifications of such material outside the IETF Standards Process.
  Without obtaining an adequate license from the person(s) controlling
  the copyright in such materials, this document may not be modified
  outside the IETF Standards Process, and derivative works of it may
  not be created outside the IETF Standards Process, except to format
  it for publication as an RFC or to translate it into languages other
  than English.

























Gilman, et al.               Standards Track                    [Page 2]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


Table of Contents

  1. Introduction ....................................................4
  2. Terminology .....................................................4
  3. SDP Media Capabilities ..........................................6
     3.1. Requirements ...............................................6
     3.2. Solution Overview ..........................................7
     3.3. New Capability Attributes .................................13
          3.3.1. The Media Format Capability Attributes .............13
          3.3.2. The Media Format Parameter Capability Attribute ....16
          3.3.3. The Media-Specific Capability Attribute ............19
          3.3.4. New Configuration Parameters .......................21
          3.3.5. The Latent Configuration Attribute .................23
          3.3.6. Enhanced Potential Configuration Attribute .........25
          3.3.7. Substitution of Media Payload Type Numbers
                 in Capability ......................................29
          3.3.8. The Session Capability Attribute ...................30
     3.4. Offer/Answer Model Extensions .............................35
          3.4.1. Generating the Initial Offer .......................35
          3.4.2. Generating the Answer ..............................39
          3.4.3. Offerer Processing of the Answer ...................43
          3.4.4. Modifying the Session ..............................43
  4. Examples .......................................................44
     4.1. Alternative Codecs ........................................44
     4.2. Alternative Combinations of Codecs (Session
          Configurations) ...........................................47
     4.3. Latent Media Streams ......................................47
  5. IANA Considerations ............................................49
     5.1. New SDP Attributes ........................................49
     5.2. New SDP Capability Negotiation Option Tag .................50
     5.3. SDP Capability Negotiation Configuration
          Parameters Registry .......................................50
     5.4. SDP Capability Negotiation Configuration Parameter
          Registrations .............................................52
  6. Security Considerations ........................................52
  7. Acknowledgements ...............................................53
  8. References .....................................................54
     8.1. Normative References ......................................54
     8.2. Informative References ....................................54












Gilman, et al.               Standards Track                    [Page 3]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


1.  Introduction

  "Session Description Protocol (SDP) Capability Negotiation" [RFC5939]
  provides a general framework for indicating and negotiating
  capabilities in SDP [RFC4566].  The base framework defines only
  capabilities for negotiating transport protocols and attributes.

  RFC 5939 [RFC5939] lists some of the issues with the current SDP
  capability negotiation process.  An additional real-life problem is
  to be able to offer one media stream (e.g., audio) but list the
  capability to support another media stream (e.g., video) without
  actually offering it concurrently.

  In this document, we extend the framework by defining media
  capabilities that can be used to indicate and negotiate media types
  and their associated format parameters.  This document also adds the
  ability to declare support for media streams, the use of which can be
  offered and negotiated later, and the ability to specify session
  configurations as combinations of media stream configurations.  The
  definitions of new attributes for media capability negotiation are
  chosen to make the translation from these attributes to
  "conventional" SDP [RFC4566] media attributes as straightforward as
  possible in order to simplify implementation.  This goal is intended
  to reduce processing in two ways: each proposed configuration in an
  offer may be easily translated into a conventional SDP media stream
  record for processing by the receiver and the construction of an
  answer based on a selected proposed configuration would be
  straightforward.

  This document updates RFC 5939 [RFC5939] by updating the IANA
  considerations.  All other extensions defined in this document are
  considered extensions above and beyond RFC 5939 [RFC5939].

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] and
  indicate requirement levels for compliant implementations.

  Actual Configuration: An actual configuration specifies which
  combinations of SDP session parameters and media stream components
  can be used in the current offer/answer exchange and with what
  parameters.  Use of an actual configuration does not require any
  further negotiation in the offer/answer exchange.  See RFC 5939
  [RFC5939] for further details.





Gilman, et al.               Standards Track                    [Page 4]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  Base Attributes: Conventional SDP attributes appearing in the base
  configuration of a media block.

  Base Configuration: The media configuration represented by a media
  block exclusive of all the capability negotiation attributes defined
  in this document, the base capability negotiation document [RFC5939],
  or any other capability negotiation document.  In an offer SDP, the
  base configuration corresponds to the actual configuration as defined
  in RFC 5939 [RFC5939].

  Conventional Attribute: Any SDP attribute other than those defined by
  the series of capability negotiation specifications.

  Conventional SDP: An SDP record devoid of capability negotiation
  attributes.

  Media Format Capability: A media format, typically a media subtype
  such as PCMU, H263-1998, or T38, expressed in the form of a
  capability.

  Media Format Parameter Capability: A media format parameter ("a=fmtp"
  in conventional SDP) expressed in the form of a capability.  The
  media format parameter capability is associated with a media format
  capability.

  Media Capability: The combined set of capabilities associated with
  expressing a media format and its relevant parameters (e.g., media
  format parameters and media specific parameters).

  Potential Configuration: A potential configuration indicates which
  combinations of capabilities can be used for the session and its
  associated media stream components.  Potential configurations are not
  ready for use; however, they are offered for potential use in the
  current offer/answer exchange.  They provide an alternative that may
  be used instead of the actual configuration, subject to negotiation
  in the current offer/answer exchange.  See RFC 5939 [RFC5939] for
  further details.

  Latent Configuration: A latent configuration indicates which
  combinations of capabilities could be used in a future negotiation
  for the session and its associated media stream components.  Latent
  configurations are neither ready for use nor offered for actual or
  potential use in the current offer/answer exchange.  Latent
  configurations merely inform the other side of possible
  configurations supported by the entity.  Those latent configurations
  may be used to guide subsequent offer/answer exchanges, but they are
  not offered for use as part of the current offer/answer exchange.




Gilman, et al.               Standards Track                    [Page 5]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


3.  SDP Media Capabilities

  The SDP capability negotiation [RFC5939] discusses the use of any SDP
  [RFC4566] attribute (a=) under the attribute capability "acap".  The
  limitations of using "acap" for "fmtp" and "rtpmap" in a potential
  configuration are described in RFC 5939 [RFC5939]; for example, they
  can be used only at the media level since they are media-level
  attributes.  RFC 5939 [RFC5939] does not provide a way to exchange
  media-level capabilities prior to the actual offer of the associated
  media stream.  This section provides an overview of extensions
  providing an SDP media capability negotiation solution offering more
  robust capabilities negotiation.  This is followed by definitions of
  new SDP attributes for the solution and its associated updated
  offer/answer procedures [RFC3264].

3.1.  Requirements

  The capability negotiation extensions requirements considered herein
  are as follows.

  REQ-01:  Support the specification of alternative (combinations of)
           media formats (codecs) in a single media block.

  REQ-02:  Support the specification of alternative media format
           parameters for each media format.

  REQ-03:  Retain backward compatibility with conventional SDP.  Ensure
           that each and every offered configuration can be easily
           translated into a corresponding SDP media block expressed
           with conventional SDP lines.

  REQ-04:  Ensure that the scheme operates within the offer/answer
           model in such a way that media formats and parameters can be
           agreed upon with a single exchange.

  REQ-05:  Provide the ability to express offers in such a way that the
           offerer can receive media as soon as the offer is sent.
           (Note that the offerer may not be able to render received
           media prior to exchange of keying material.)

  REQ-06:  Provide the ability to offer latent media configurations for
           future negotiation.

  REQ-07:  Provide reasonable efficiency in the expression of
           alternative media formats and/or format parameters,
           especially in those cases in which many combinations of
           options are offered.




Gilman, et al.               Standards Track                    [Page 6]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  REQ-08:  Retain the extensibility of the base capability negotiation
           mechanism.

  REQ-09:  Provide the ability to specify acceptable combinations of
           media streams and media formats.  For example, offer a PCMU
           audio stream with an H264 video stream or a G729 audio
           stream with an H263 video stream.  This ability would give
           the offerer a means to limit processing requirements for
           simultaneous streams.  This would also permit an offer to
           include the choice of an audio/T38 stream or an image/T38
           stream, but not both.

  Other possible extensions have been discussed, but have not been
  treated in this document.  They may be considered in the future.
  Three such extensions are:

  FUT-01:  Provide the ability to mix, or change, media types within a
           single media block.  Conventional SDP does not support this
           capability explicitly; the usual technique is to define a
           media subtype that represents the actual format within the
           nominal media type.  For example, T.38 FAX as an alternative
           to audio/PCMU within an audio stream is identified as
           audio/T38; a separate FAX stream would use image/T38.

  FUT-02:  Provide the ability to support multiple transport protocols
           within an active media stream without reconfiguration.  This
           is not explicitly supported by conventional SDP.

  FUT-03:  Provide capability negotiation attributes for all media-
           level SDP line types in the same manner as already done for
           the attribute type, with the exception of the media line
           type itself.  The media line type is handled in a special
           way to permit compact expression of media coding/format
           options.  The line types are bandwidth ("b="), information
           ("i="), connection data ("c="), and, possibly, the
           deprecated encryption key ("k=").

3.2.  Solution Overview

  The solution consists of new capability attributes corresponding to
  conventional SDP line types, new parameters for the "pcfg", "acfg",
  and the new "lcfg" attributes extending the base attributes from RFC
  5939 [RFC5939], and a use of the "pcfg" attribute to return
  capability information in the SDP answer.

  Several new attributes are defined in a manner that can be related to
  the capabilities specified in a media line, and its corresponding
  "rtpmap" and "fmtp" attributes.



Gilman, et al.               Standards Track                    [Page 7]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  o  A new attribute ("a=rmcap") defines RTP-based media format
     capabilities in the form of a media subtype (e.g., "PCMU"), and
     its encoding parameters (e.g., "/8000/2").  Each resulting media
     format type/subtype capability has an associated handle called a
     media capability number.  The encoding parameters are as specified
     for the "rtpmap" attribute defined in SDP [RFC4566], without the
     payload type number part.

  o  A new attribute ("a=omcap") defines other (non-RTP-based) media
     format capabilities in the form of a media subtype only (e.g.,
     "T38").  Each resulting media format type/subtype capability has
     an associated handle called a media capability number.

  o  A new attribute ("a=mfcap") specifies media format parameters
     associated with one or more media format capabilities.  The
     "mfcap" attribute is used primarily to associate the media format
     parameters normally carried in the "fmtp" attribute.  Note that
     media format parameters can be used with RTP and non-RTP-based
     media formats.

  o  A new attribute ("a=mscap") specifies media parameters associated
     with one or more media format capabilities.  The "mscap" attribute
     is used to associate capabilities with attributes other than
     "fmtp" or "rtpmap", for example, the "rtcp-fb" attribute defined
     in RFC 4585 [RFC4585].

  o  A new attribute ("a=lcfg") specifies latent media stream
     configurations when no corresponding media line ("m=") is offered.
     An example is the offer of latent configurations for video even
     though no video is currently offered.  If the peer indicates
     support for one or more offered latent configurations, the
     corresponding media stream(s) may be added via a new offer/answer
     exchange.

  o  A new attribute ("a=sescap") is used to specify an acceptable
     combination of simultaneous media streams and their configurations
     as a list of potential and/or latent configurations.

  New parameters are defined for the potential configuration ("pcfg"),
  latent configuration ("lcfg"), and accepted configuration ("acfg")
  attributes to associate the new attributes with particular
  configurations.

  o  A new parameter type ("m=") is added to the potential
     configuration ("a=pcfg:") attribute and the actual configuration
     ("a=acfg:") attribute defined in RFC 5939 [RFC5939] and to the new
     latent configuration ("a=lcfg:") attribute.  This permits
     specification of media capabilities (including their associated



Gilman, et al.               Standards Track                    [Page 8]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


     parameters) and combinations thereof for the configuration.  For
     example, the "a=pcfg:" line might specify PCMU and telephone
     events [RFC4733] or G.729B and telephone events as acceptable
     configurations.  The "a=acfg:" line in the answer would specify
     the configuration chosen.

  o  A new parameter type ("pt=") is added to the potential
     configuration, actual configuration, and latent configuration
     attributes.  This parameter associates RTP payload type numbers
     with the referenced RTP-based media format capabilities and is
     appropriate only when the transport protocol uses RTP.

  o  A new parameter type ("mt=") is used to specify the media type for
     latent configurations.

  Special processing rules are defined for capability attribute
  arguments in order to reduce the need to replicate essentially
  identical attribute lines for the base configuration and potential
  configurations.

  o  A substitution rule is defined for any capability attribute to
     permit the replacement of the (escaped) media capability number
     with the media format identifier (e.g., the payload type number in
     audio/video profiles).

  o  Replacement rules are defined for the conventional SDP equivalents
     of the "mfcap" and "mscap" capability attributes.  This reduces
     the necessity to use the deletion qualifier in the "a=pcfg"
     parameter in order to ignore "rtpmap", "fmtp", and certain other
     attributes in the base configuration.

  o  An argument concatenation rule is defined for "mfcap" attributes
     that refer to the same media capability number.  This makes it
     convenient to combine format options concisely by associating
     multiple mfcap lines with multiple media format capabilities.

  This document extends the base protocol extensions to the
  offer/answer model that allow for capabilities and potential
  configurations to be included in an offer.  Media capabilities
  constitute capabilities that can be used in potential and latent
  configurations.  Whereas potential configurations constitute
  alternative offers that may be accepted by the answerer instead of
  the actual configuration(s) included in the "m=" line(s) and
  associated parameters, latent configurations merely inform the other
  side of possible configurations supported by the entity.  Those
  latent configurations may be used to guide subsequent offer/answer
  exchanges, but they are not part of the current offer/answer
  exchange.



Gilman, et al.               Standards Track                    [Page 9]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  The mechanism is illustrated by the offer/answer exchange below,
  where Alice sends an offer to Bob:

                  Alice                            Bob
                 | (1) Offer (SRTP and RTP)         |
                 |--------------------------------->|
                 |                                  |
                 | (2) Answer (RTP)                 |
                 |<---------------------------------|
                 |                                  |

  Alice's offer includes RTP and Secure Real-time Transport Protocol
  (SRTP) as alternatives.  RTP is the default, but SRTP is the
  preferred one (long lines are folded to fit the margins):

     v=0
     o=- 25678 753849 IN IP4 192.0.2.1
     s=
     c=IN IP4 192.0.2.1
     t=0 0
     a=creq:med-v0
     m=audio 3456 RTP/AVP 0 18
     a=tcap:1 RTP/SAVP RTP/AVP
     a=rtpmap:0 PCMU/8000/1
     a=rtpmap:18 G729/8000/1
     a=fmtp:18 annexb=yes
     a=rmcap:1,4 G729/8000/1
     a=rmcap:2 PCMU/8000/1
     a=rmcap:5 telephone-event/8000
     a=mfcap:1 annexb=no
     a=mfcap:4 annexb=yes
     a=mfcap:5 0-11
     a=acap:1 crypto:1 AES_CM_128_HMAC_SHA1_32 \
     inline:NzB4d1BINUAvLEw6UzF3WSJ+PSdFcGdUJShpX1Zj|2^20|1:32
     a=pcfg:1 m=4,5|1,5 t=1 a=1 pt=1:100,4:101,5:102
     a=pcfg:2 m=2 t=1 a=1 pt=2:103
     a=pcfg:3 m=4 t=2 pt=4:18

  The required base and extensions are provided by the "a=creq"
  attribute defined in RFC 5939 [RFC5939], with the option tag
  "med-v0", which indicates that the extension framework defined here
  must be supported.  The base-level capability negotiation support
  ("cap-v0" [RFC5939]) is implied since it is required for the
  extensions.

  The "m=" line indicates that Alice is offering to use plain RTP with
  PCMU or G.729B.  The media line implicitly defines the default




Gilman, et al.               Standards Track                   [Page 10]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  transport protocol (RTP/AVP in this case) and the default actual
  configuration.

  The "a=tcap:1" line, specified in the SDP capability negotiation base
  protocol [RFC5939], defines transport protocol capabilities, in this
  case Secure RTP (SAVP profile) as the first option and RTP (AVP
  profile) as the second option.

  The "a=rmcap:1,4" line defines two G.729 RTP-based media format
  capabilities, numbered 1 and 4, and their encoding rate.  The
  capabilities are of media type "audio" and subtype G729.  Note that
  the media subtype is explicitly specified here, rather than RTP
  payload type numbers.  This permits the assignment of payload type
  numbers in the media stream configuration specification.  In this
  example, two G.729 subtype capabilities are defined.  This permits
  the declaration of two sets of formatting parameters for G.729.

  The "a=rmcap:2" line defines a G.711 mu-law capability, numbered 2.

  The "a=rmcap:5" line defines an audio telephone-event capability,
  numbered 5.

  The "a=mfcap:1" line specifies the "fmtp" formatting parameters for
  capability 1 (offerer will not accept G.729 Annex B packets).

  The "a=mfcap:4" line specifies the "fmtp" formatting parameters for
  capability 4 (offerer will accept G.729 Annex B packets).

  The "a=mfcap:5" line specifies the "fmtp" formatting parameters for
  capability 5 (the dual-tone multi-frequency (DTMF) touchtones
  0-9,*,#).

  The "a=acap:1" line specified in the base protocol provides the
  "crypto" attribute that provides the keying material for SRTP using
  SDP security descriptions.

  The "a=pcfg:" attributes provide the potential configurations
  included in the offer by reference to the media capabilities,
  transport capabilities, attribute capabilities, and specified payload
  type number mappings.  Three explicit alternatives are provided; the
  lowest-numbered one is the preferred one.  The "a=pcfg:1 ..." line
  specifies media capabilities 4 and 5, i.e., G.729B and DTMF
  (including their associated media format parameters), or media
  capability 1 and 5, i.e., G.729 and DTMF (including their associated
  media format parameters).  Furthermore, it specifies transport
  protocol capability 1 (i.e., the RTP/SAVP profile - secure RTP), and
  the attribute capability 1, i.e., the "crypto" attribute provided.
  Last, it specifies a payload type number mapping for (RTP-based)



Gilman, et al.               Standards Track                   [Page 11]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  media capabilities 1, 4, and 5, thereby permitting the offerer to
  distinguish between encrypted media and unencrypted media received
  prior to receipt of the answer.

  Use of unique payload type numbers in alternative configurations is
  not required; codecs such as Adaptive Multi-Rate Wideband (AMR-WB)
  [RFC4867] have the potential for so many combinations of options that
  it may be impractical to define unique payload type numbers for all
  supported combinations.  If unique payload type numbers cannot be
  specified, then the offerer will be obliged to wait for the SDP
  answer before rendering received media.  For SRTP using Security
  Descriptions (SDES) inline keying [RFC4568], the offerer will still
  need to receive the answer before being able to decrypt the stream.

  The second alternative ("a=pcfg:2 ...") specifies media capability 2,
  i.e., PCMU, under the RTP/SAVP profile, with the same SRTP key
  material.

  The third alternative ("a=pcfg:3 ...") offers G.729B unsecured; its
  only purpose in this example is to show a preference for G.729B over
  PCMU.

  Per RFC 5939 [RFC5939], the media line, with any qualifying
  attributes such as "fmtp" or "rtpmap", is itself considered a valid
  configuration (the current actual configuration); it has the lowest
  preference (per RFC 5939 [RFC5939]).

  Bob receives the SDP offer from Alice.  Bob supports G.729B, PCMU,
  and telephone events over RTP, but not SRTP, hence he accepts the
  potential configuration 3 for RTP provided by Alice.  Bob generates
  the following answer:

     v=0
     o=- 24351 621814 IN IP4 192.0.2.2
     s=
     c=IN IP4 192.0.2.2
     t=0 0
     a=csup:med-v0
     m=audio 4567 RTP/AVP 18
     a=rtpmap:18 G729/8000
     a=fmtp:18 annexb=yes
     a=acfg:3 m=4 t=2 pt=4:18

  Bob includes the "a=csup" and "a=acfg" attributes in the answer to
  inform Alice that he can support the med-v0 level of capability
  negotiations.  Note that in this particular example, the answerer
  supported the capability extensions defined here; however, had he
  not, he would simply have processed the offer based on the offered



Gilman, et al.               Standards Track                   [Page 12]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  PCMU and G.729 codecs under the RTP/AVP profile only.  Consequently,
  the answer would have omitted the "a=csup" attribute line and chosen
  one or both of the PCMU and G.729 codecs instead.  The answer carries
  the accepted configuration in the "m=" line along with corresponding
  "rtpmap" and/or "fmtp" parameters, as appropriate.

  Note that per the base protocol, after the above, Alice MAY generate
  a new offer with an actual configuration ("m=" line, etc.)
  corresponding to the actual configuration referenced in Bob's answer
  (not shown here).

3.3.  New Capability Attributes

  In this section, we present the new attributes associated with
  indicating the media capabilities for use by the SDP capability
  negotiation.  The approach taken is to keep things similar to the
  existing media capabilities defined by the existing media
  descriptions ("m=" lines) and the associated "rtpmap" and "fmtp"
  attributes.  We use media subtypes and "media capability numbers" to
  link the relevant media capability parameters.  This permits the
  capabilities to be defined at the session level and be used for
  multiple streams, if desired.  For RTP-based media formats, payload
  types are then specified at the media level (see Section 3.3.4.2).

  A media capability merely indicates possible support for the media
  type and media format(s) and parameters in question.  In order to
  actually use a media capability in an offer/answer exchange, it MUST
  be referenced in a potential configuration.

  Media capabilities, i.e., the attributes associated with expressing
  media capability formats, parameters, etc., can be provided at the
  session level and/or the media level.  Media capabilities provided at
  the session level may be referenced in any "pcfg" or "lcfg" attribute
  at the media level (consistent with the media type), whereas media
  capabilities provided at the media level may be referenced only by
  the "pcfg" or "lcfg" attribute within that media stream.  In either
  case, the scope of the <med-cap-num> is the entire session
  description.  This enables each media capability to be uniquely
  referenced across the entire session description (e.g., in a
  potential configuration).

3.3.1.  The Media Format Capability Attributes

  Media subtypes can be expressed as media format capabilities by use
  of the "a=rmcap" and "a=omcap" attributes.  The "a=rmcap" attribute
  MUST be used for RTP-based media, whereas the "a=omcap" attribute
  MUST be used for non-RTP-based (other) media formats.  The two
  attributes are defined as follows:



Gilman, et al.               Standards Track                   [Page 13]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  a=rmcap:<media-cap-num-list> <encoding-name>/<clock-rate>
                               [/<encoding-parms>]

  a=omcap:<media-cap-num-list> <format-name>

  where <media-cap-num-list> is a (list of) media capability number(s)
  used to number a media format capability, the <encoding name> or
  <format-name> is the media subtype, e.g., H263-1998, PCMU, or T38,
  <clock rate> is the encoding rate, and <encoding parms> are the media
  encoding parameters for the media subtype.  All media format
  capabilities in the list are assigned to the same media type/subtype.
  Each occurrence of the "rmcap" and "omcap" attribute MUST use unique
  values in their <media-cap-num-list>; the media capability numbers
  are shared between the two attributes and the numbers MUST be unique
  across the entire SDP session.  In short, the "rmcap" and "omcap"
  attributes define media format capabilities and associate them with a
  media capability number in the same manner as the "rtpmap" attribute
  defines them and associates them with a payload type number.
  Additionally, the attributes allow multiple capability numbers to be
  defined for the media format in question by specifying a range of
  media capability numbers.  This permits the media format to be
  associated with different media parameters in different
  configurations.  When a range of capability numbers is specified, the
  first (leftmost) capability number MUST be strictly smaller than the
  second (rightmost), i.e., the range increases and covers at least two
  numbers.

  In ABNF [RFC5234], we have:

  media-capability-line = rtp-mcap / non-rtp-mcap

  rtp-mcap           = "a=rmcap:" media-cap-num-list
                          1*WSP encoding-name "/" clock-rate
                          ["/" encoding-parms]
  non-rtp-mcap       = "a=omcap:" media-cap-num-list 1*WSP format-name
  media-cap-num-list = media-cap-num-element
                       *("," media-cap-num-element)
  media-cap-num-element = media-cap-num
                               / media-cap-num-range
  media-cap-num-range = media-cap-num "-" media-cap-num
  media-cap-num      = NonZeroDigit *9(DIGIT)
  encoding-name      = token ;defined in RFC 4566
  clock-rate         = NonZeroDigit *9(DIGIT)
  encoding-parms     = token
  format-name        = token ;defined in RFC 4566
  NonZeroDigit       = %x31-39    ; 1-9





Gilman, et al.               Standards Track                   [Page 14]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  The encoding-name, clock-rate, and encoding-params are as defined to
  appear in an "rtpmap" attribute for each media type/subtype.  Thus,
  it is easy to convert an "rmcap" attribute line into one or more
  "rtpmap" attribute lines, once a payload type number is assigned to a
  media-cap-num (see Section 3.3.5).

  The format-name is a media format description for non-RTP-based media
  as defined for the <fmt> part of the media description ("m=" line) in
  SDP [RFC4566].  In simple terms, it is the name of the media format,
  e.g., "t38".  This form can also be used in cases such as Binary
  Floor Control Protocol (BFCP) [RFC4585] where the fmt list in the
  "m=" line is effectively ignored (BFCP uses "*").

  The "rmcap" and "omcap" attributes can be provided at the session
  level and/or the media level.  There can be more than one "rmcap" and
  more than one "omcap" attribute at both the session and media levels
  (i.e., more than one of each at the session level and more than one
  of each in each media description).  Media capability numbers cannot
  include leading zeroes, and each media-cap-num MUST be unique within
  the entire SDP record; it is used to identify that media capability
  in potential, latent, and actual configurations, and in other
  attribute lines as explained below.  Note that the media-cap-num
  values are shared between the "rmcap" and "omcap" attributes; hence,
  the uniqueness requirement applies to the union of them.  When the
  media capabilities are used in a potential, latent, or actual
  configuration, the media formats referred by those configurations
  apply at the media level, irrespective of whether the media
  capabilities themselves were specified at the session or media level.
  In other words, the media capability applies to the specific media
  description associated with the configuration that invokes it.





















Gilman, et al.               Standards Track                   [Page 15]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  For example:

     v=0
     o=- 24351 621814 IN IP4 192.0.2.2
     s=
     c=IN IP4 192.0.2.2
     t=0 0
     a=rmcap:1 L16/8000/1
     a=rmcap:2 L16/16000/2
     a=rmcap:3 H263-1998/90000
     a=omcap:4 example
     m=audio 54320 RTP/AVP 0
     a=pcfg:1 m=1|2, pt=1:99,2:98
     m=video 66544 RTP/AVP 100
     a=rtpmap:100 H264/90000
     a=pcfg:10 m=3 pt=3:101
     a=tcap:1 TCP
     a=pcfg:11 m=4 t=1

3.3.2.  The Media Format Parameter Capability Attribute

  This attribute is used to associate media format specific parameters
  with one or more media format capabilities.  The form of the
  attribute is

     a=mfcap:<media-caps> <list of parameters>

  where <media-caps> permits the list of parameters to be associated
  with one or more media format capabilities and the format parameters
  are specific to the type of media format.  The mfcap lines map to a
  single traditional SDP "fmtp" attribute line (one for each entry in
  <media-caps>) of the form

     a=fmtp:<fmt> <list of parameters>

  where <fmt> is the media format parameter defined in RFC 4566
  [RFC4566], as appropriate for the particular media stream.  The
  "mfcap" attribute MUST be used to encode attributes for media
  capabilities, which would conventionally appear in an "fmtp"
  attribute.  The existing "acap" attribute MUST NOT be used to encode
  "fmtp" attributes.

  The "mfcap" attribute adheres to SDP [RFC4566] attribute production
  rules with

     media-format-parameter-capability =
            "a=mfcap:" media-cap-num-list 1*WSP fmt-specific-param-list
     fmt-specific-param-list = text ; defined in RFC 4566



Gilman, et al.               Standards Track                   [Page 16]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  Note that media format parameters can be used with RTP-based and non-
  RTP-based media formats.

3.3.2.1.  Media Format Parameter Concatenation Rule

  The appearance of media subtypes with a large number of formatting
  options (e.g., AMR-WB [RFC4867]), coupled with the restriction that
  only a single "fmtp" attribute can appear per media format, suggests
  that it is useful to create a combining rule for "mfcap" parameters
  that are associated with the same media capability number.
  Therefore, different mfcap lines MAY include the same media-cap-num
  in their media-cap-num-list.  When a particular media capability is
  selected for processing, the parameters from each mfcap line that
  references the particular capability number in its media-cap-num-list
  are concatenated together via ";", in the order the "mfcap"
  attributes appear in the SDP record, to form the equivalent of a
  single "fmtp" attribute line.  This permits one to define a separate
  mfcap line for a single parameter and value that is to be applied to
  each media capability designated in the media-cap-num-list.  This
  provides a compact method to specify multiple combinations of format
  parameters when using codecs with multiple format options.  Note that
  order-dependent parameters SHOULD be placed in a single mfcap line to
  avoid possible problems with line rearrangement by a middlebox.

  Format parameters are not parsed by SDP; their content is specific to
  the media type/subtype.  When format parameters for a specific media
  capability are combined from multiple "a=mfcap" lines that reference
  that media capability, the format-specific parameters are
  concatenated together and separated by ";" for construction of the
  corresponding format attribute ("a=fmtp").  The resulting format
  attribute will look something like the following (without line
  breaks):

       a=fmtp:<fmt> <fmt-specific-param-list1>;
                    <fmt-specific-param-list2>;
                    ...

  where <fmt> depends on the transport protocol in the manner defined
  in RFC 4566 [RFC4566].  SDP cannot assess the legality of the
  resulting parameter list in the "a=fmtp" line; the user must take
  care to ensure that legal parameter lists are generated.

  The "mfcap" attribute can be provided at the session level and the
  media level.  There can be more than one "mfcap" attribute at the
  session or media level.  The unique media-cap-num is used to
  associate the parameters with a media capability.





Gilman, et al.               Standards Track                   [Page 17]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  As a simple example, a G.729 capability is, by default, considered to
  support comfort noise as defined by Annex B.  Capabilities for G.729
  with and without comfort noise support may thus be defined by:

     a=rmcap:1,2 G729/8000
     a=mfcap:2 annexb:no

  Media capability 1 supports G.729 with Annex B, whereas media
  capability 2 supports G.729 without Annex B.

  Example for H.263 video:

     a=rmcap:1 H263-1998/90000
     a=rmcap:2 H263-2000/90000
     a=mfcap:1 CIF=4;QCIF=2;F=1;K=1
     a=mfcap:2 profile=2;level=2.2

  Finally, for six format combinations of the Adaptive Multi-Rate
  codec:

     a=rmcap:1-3 AMR/8000/1
     a=rmcap:4-6 AMR-WB/16000/1
     a=mfcap:1,2,3,4 mode-change-capability=1
     a=mfcap:5,6 mode-change-capability=2
     a=mfcap:1,2,3,5 max-red=220
     a=mfcap:3,4,5,6 octet-align=1
     a=mfcap:1,3,5 mode-set=0,2,4,7
     a=mfcap:2,4,6 mode-set=0,3,5,6

  So that AMR codec #1, when specified in a "pcfg" attribute within an
  audio stream block (and assigned payload type number 98) as in:

     a=pcfg:1 m=1 pt=1:98

  is essentially equivalent to the following:

     m=audio 49170 RTP/AVP 98
     a=rtpmap:98 AMR/8000/1
     a=fmtp:98 mode-change-capability=1; \
     max-red=220; mode-set=0,2,4,7

  and AMR codec #4 with payload type number 99, depicted by the
  potential configuration:

     a=pcfg:4 m=4, pt=4:99






Gilman, et al.               Standards Track                   [Page 18]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  is equivalent to the following:

     m=audio 49170 RTP/AVP 99
     a=rtpmap:99 AMR-WB/16000/1
     a=fmtp:99 mode-change-capability=1; octet-align=1; \
     mode-set=0,3,5,6

  and so on for the other four combinations.  SDP could thus convert
  the media capabilities specifications into one or more alternative
  media stream specifications, one of which can be chosen for the
  answer.

3.3.3.  The Media-Specific Capability Attribute

  Attributes and parameters associated with a media format are
  typically specified using the "rtpmap" and "fmtp" attributes in SDP,
  and the similar "rmcap" and "mfcap" attributes in SDP media
  capabilities.  Some SDP extensions define other attributes that need
  to be associated with media formats, for example, the "rtcp-fb"
  attribute defined in RFC 4585 [RFC4585].  Such media-specific
  attributes, beyond the "rtpmap" and "fmtp" attributes, may be
  associated with media capability numbers via a new media-specific
  attribute, "mscap", of the following form:

        a=mscap:<media caps star> <att field> <att value>


  where <media caps star> is a (list of) media capability number(s),
  <att field> is the attribute name, and <att value> is the value field
  for the named attribute.  Note that the media capability numbers
  refer to media format capabilities specified elsewhere in the SDP
  ("rmcap" and/or "omcap").  If a range of capability numbers is
  specified, the first (leftmost) capability number MUST be strictly
  smaller than the second (rightmost).  The media capability numbers
  may include a wildcard ("*"), which will be used instead of any
  payload type mappings in the resulting SDP (see, e.g., RFC 4585
  [RFC4585] and the example below).  In ABNF, we have:

         media-specific-capability = "a=mscap:"
                                      media-caps-star
                                      1*WSP att-field ; from RFC 4566
                                      1*WSP att-value ; from RFC 4566
         media-caps-star           =  media-cap-star-element
                                        *("," media-cap-star-element)
         media-cap-star-element    = (media-cap-num [wildcard])
                                   / (media-cap-num-range [wildcard])
         wildcard                  = "*"




Gilman, et al.               Standards Track                   [Page 19]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  Given an association between a media capability and a payload type
  number as specified by the "pt=" parameters in a "pcfg" attribute
  line, a mscap line may be translated easily into a conventional SDP
  attribute line of the form:

     a=<att field>":"<fmt> <att value> ; <fmt> defined in SDP [RFC4566]

  A resulting attribute that is not a legal SDP attribute, as specified
  by RFC 4566, MUST be ignored by the receiver.

  If a media capability number (or range) contains a wildcard character
  at the end, any payload type mapping specified for that media-
  specific capability (or range of capabilities) will use the wildcard
  character in the resulting SDP instead of the payload type specified
  in the payload type mapping ("pt" parameter) in the configuration
  attribute.

  A single mscap line may refer to multiple media capabilities by use
  of a capability number range; this is equivalent to multiple mscap
  lines, each with the same attribute values (but different media
  capability numbers), one line per media capability.

  Multiple mscap lines may refer to the same media capability, but,
  unlike the "mfcap" attribute, no concatenation operation is defined.
  Hence, multiple mscap lines applied to the same media capability are
  equivalent to multiple lines of the specified attribute in a
  conventional media record.

  Here is an example with the "rtcp-fb" attribute, modified from an
  example in RFC 5104 [RFC5104] (with the session level and audio media
  omitted).  If the offer contains a media block like the following
  (note the wildcard character),

     m=video 51372 RTP/AVP 98
     a=rtpmap:98 H263-1998/90000
     a=tcap:1 RTP/AVPF
     a=rmcap:1 H263-1998/90000
     a=mscap:1 rtcp-fb ccm tstr
     a=mscap:1 rtcp-fb ccm fir
     a=mscap:1* rtcp-fb ccm tmmbr smaxpr=120
     a=pcfg:1 t=1 m=1 pt=1:98

  and if the proposed configuration is chosen, then the equivalent
  media block would look like the following







Gilman, et al.               Standards Track                   [Page 20]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


     m=video 51372 RTP/AVPF 98
     a=rtpmap:98 H263-1998/90000
     a=rtcp-fb:98 ccm tstr
     a=rtcp-fb:98 ccm fir
     a=rtcp-fb:* ccm tmmbr smaxpr=120

3.3.4.  New Configuration Parameters

  Along with the new attributes for media capabilities, new extension
  parameters are defined for use in the potential configuration, the
  actual configuration, and/or the new latent configuration defined in
  Section 3.3.5.

3.3.4.1.  The Media Configuration Parameter (m=)

  The media configuration parameter is used to specify the media
  format(s) and related parameters for a potential, actual, or latent
  configuration.  Adhering to the ABNF for extension-config-list in RFC
  5939 [RFC5939] with

            ext-cap-name = "m"
            ext-cap-list = media-cap-num-list
                           [*(BAR media-cap-num-list)]

  we have

             media-config-list = ["+"] "m=" media-cap-num-list
                                            *(BAR media-cap-num-list)
                                  ;BAR is defined in RFC 5939
                                  ;media-cap-num-list is defined above

  Alternative media configurations are separated by a vertical bar
  ("|").  The alternatives are ordered by preference, most-preferred
  first.  When media capabilities are not included in a potential
  configuration at the media level, the media type and media format
  from the associated "m=" line will be used.  The use of the plus sign
  ("+") is described in RFC 5939.

3.3.4.2.  The Payload Type Number Mapping Parameter (pt=)

  The payload type number mapping parameter is used to specify the
  payload type number to be associated with each RTP-based media format
  in a potential, actual, or latent configuration.  We define the
  payload type number mapping parameter, payload-number-config-list, in
  accordance with the extension-config-list format defined in RFC 5939
  [RFC5939].  In ABNF:





Gilman, et al.               Standards Track                   [Page 21]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  payload-number-config-list = ["+"] "pt=" media-map-list
  media-map-list      = media-map *("," media-map)
  media-map           = media-cap-num ":" payload-type-number
                           ; media-cap-num is defined in Section 3.3.1
  payload-type-number = NonZeroDigit *2(DIGIT) ; RTP payload
                                               ; type number

  The example in Section 3.3.7 shows how the parameters from the rmcap
  line are mapped to payload type numbers from the "pcfg" "pt"
  parameter.  The use of the plus sign ("+") is described in RFC 5939
  [RFC5939].

  A latent configuration represents a future capability; hence, the
  "pt=" parameter is not directly meaningful in the "lcfg" attribute
  because no actual media session is being offered or accepted.  It is
  permitted in order to tie any payload type number parameters within
  attributes to the proper media format.  A primary example is the case
  of format parameters for the Redundant Audio Data (RED) [RFC2198]
  payload, which are payload type numbers.  Specific payload type
  numbers used in a latent configuration MAY be interpreted as
  suggestions to be used in any future offer based on the latent
  configuration, but they are not binding; the offerer and/or answerer
  may use any payload type numbers each deems appropriate.  The use of
  explicit payload type numbers for latent configurations can be
  avoided by use of the parameter substitution rule of Section 3.3.7.
  Future extensions are also permitted.  Note that leading zeroes are
  not permitted.

3.3.4.3.  The Media Type Parameter

  When a latent configuration is specified (always at the media level),
  indicating the ability to support an additional media stream, it is
  necessary to specify the media type (audio, video, etc.)  as well as
  the format and transport type.  The media type parameter is defined
  in ABNF as

           media-type = ["+"] "mt=" media; media defined in RFC 4566

  At present, the media-type parameter is used only in the latent
  configuration attribute, and the use of the "+" prefix to specify
  that the entire attribute line is to be ignored if the mt= parameter
  is not understood is unnecessary.  However, if the media-type
  parameter is later added to an existing capability attribute such as
  "pcfg", then the "+" would be useful.  The media format(s) and
  transport type(s) are specified using the media configuration
  parameter ("+m=") defined above, and the transport parameter ("t=")
  defined in RFC 5939 [RFC5939], respectively.




Gilman, et al.               Standards Track                   [Page 22]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


3.3.5.  The Latent Configuration Attribute

  One of the goals of this work is to permit the exchange of
  supportable media configurations in addition to those offered or
  accepted for immediate use.  Such configurations are referred to as
  "latent configurations".  For example, a party may offer to establish
  a session with an audio stream, and, at the same time, announce its
  ability to support a video stream as part of the same session.  The
  offerer can supply its video capabilities by offering one or more
  latent video configurations along with the media stream for audio;
  the responding party may indicate its ability and willingness to
  support such a video session by returning a corresponding latent
  configuration.

  Latent configurations returned in SDP answers MUST match offered
  latent configurations (or parameter subsets thereof).  Therefore, it
  is appropriate for the offering party to announce most, if not all,
  of its capabilities in the initial offer.  This choice has been made
  in order to keep the size of the answer more compact by not requiring
  acap, rmcap, tcap, etc.  lines in the answer.

  Latent configurations may be announced by use of the latent
  configuration attribute, which is defined in a manner very similar to
  the potential configuration attribute.  The latent configuration
  attribute combines the properties of a media line and a potential
  configuration.  A latent configuration MUST include a media type
  (mt=) and a transport protocol configuration parameter since the
  latent configuration is independent of any media line present.  In
  most cases, the media configuration (m=) parameter needs to be
  present as well (see Section 4 for examples).  The "lcfg" attribute
  is a media-level attribute.

     The "lcfg" attribute is defined as a media-level attribute since
     it specifies a possible future media stream.  However, the "lcfg"
     attribute is not necessarily related to the media description
     within which it is provided.  Session capability attributes
     ("a=sescap") may be used to indicate supported media stream
     configurations.

  Each media line in an SDP description represents an offered
  simultaneous media stream, whereas each latent configuration
  represents an additional stream that may be negotiated in a future
  offer/answer exchange.  Session capability attributes may be used to
  determine whether a latent configuration may be used to form an offer
  for an additional simultaneous stream or to reconfigure an existing
  stream in a subsequent offer/answer exchange.





Gilman, et al.               Standards Track                   [Page 23]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  The latent configuration attribute is of the form:

       a=lcfg:<config-number> <latent-cfg-list>


  which adheres to the SDP [RFC4566] "attribute" production with
  att-field and att-value defined as:

     att-field  = "lcfg"
     att-value  = config-number 1*WSP lcfg-cfg-list
     config-number = NonZeroDigit *9(DIGIT)  ;DIGIT defined in RFC 5234
     lcfg-cfg-list = media-type 1*WSP pot-cfg-list
                                 ; as defined in RFC 5939
                                 ; and extended herein

  The media-type (mt=) parameter identifies the media type (audio,
  video, etc.)  to be associated with the latent media stream, and it
  MUST be present.  The pot-cfg-list MUST contain a transport-protocol-
  config-list (t=) parameter and a media-config-list (m=) parameter.
  The pot-cfg-list MUST NOT contain more than one instance of each type
  of parameter list.  As specified in RFC 5939 [RFC5939], the use of
  the "+" prefix with a parameter indicates that the entire
  configuration MUST be ignored if the parameter is not understood;
  otherwise, the parameter itself may be ignored.

  Media stream payload numbers are not assigned by a latent
  configuration.  Assignment will take place if and when the
  corresponding stream is actually offered via an "m=" line in a later
  exchange.  The payload-number-config-list is included as a parameter
  to the "lcfg" attribute in case it is necessary to tie payload
  numbers in attribute capabilities to specific media capabilities.

  If an "lcfg" attribute invokes an "acap" attribute that appears at
  the session level, then that attribute will be expected to appear at
  the session level of a subsequent offer when and if a corresponding
  media stream is offered.  Otherwise, "acap" attributes that appear at
  the media level represent media-level attributes.  Note, however,
  that "rmcap", omcap, "mfcap", "mscap", and "tcap" attributes may
  appear at the session level because they always result in media-level
  attributes or "m=" line parameters.

  The configuration numbers for latent configurations do not imply a
  preference; the offerer will imply a preference when actually
  offering potential configurations derived from latent configurations
  negotiated earlier.  Note, however, that the offerer of latent
  configurations MAY specify preferences for combinations of potential
  and latent configurations by use of the "sescap" attribute defined in
  Section 3.3.8.  For example, if an SDP offer contains, say, an audio



Gilman, et al.               Standards Track                   [Page 24]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  stream with "pcfg:1", and two latent video configurations, "lcfg:2"
  and "lcfg:3", then a session with one audio stream and one video
  stream could be specified by including "a=sescap:1 1,2|3".  One audio
  stream and two video streams could be specified by including
  "a=sescap:2 1,2,3" in the offer.  In order to permit combinations of
  latent and potential configurations in session capabilities, latent
  configuration numbers MUST be different from those used for potential
  configurations.  This restriction is especially important if the
  offerer does not require cmed-v0 capability and the recipient of the
  offer doesn't support it.  If the "lcfg" attribute is not recognized,
  the capability attributes intended to be associated with it may be
  confused with those associated with a potential configuration of some
  other media stream.  Note also that leading zeroes are not permitted
  in configuration numbers.

  If a cryptographic attribute, such as the SDES "a=crypto:" attribute
  [RFC4568], is referenced by a latent configuration through an "acap"
  attribute, any keying material required in the conventional
  attribute, such as the SDES key/salt string, MUST be included in
  order to satisfy formatting rules for the attribute.  Since the
  keying material will be visible but not actually used at this stage
  (since it's a latent configuration), the value(s) of the keying
  material MUST NOT be a real value used for real exchange of media,
  and the receiver of the "lcfg" attribute MUST ignore the value(s).

3.3.6.  Enhanced Potential Configuration Attribute

  The present work requires new extensions (parameters) for the "pcfg"
  attribute defined in the SDP capability negotiation base protocol
  [RFC5939].  The parameters and their definitions are "borrowed" from
  the definitions provided for the latent configuration attribute in
  Section 3.3.5.  The expanded ABNF definition of the "pcfg" attribute
  is

       a=pcfg: <config-number> [<pot-cfg-list>]

  where

       config-number = 1*DIGIT ;defined in [RFC5234]
       pot-cfg-list  = pot-config *(1*WSP pot-config)
       pot-config    =  attribute-config-list / ;def in [RFC5939]
            transport-protocol-config-list / ;defined in [RFC5939]
            extension-config-list / ;[RFC5939]
            media-config-list / ; Section 3.3.4.1
            payload-number-config-list ; Section 3.3.4.2

  Except for the extension-config-list, the pot-cfg-list MUST NOT
  contain more than one instance of each parameter list.



Gilman, et al.               Standards Track                   [Page 25]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


3.3.6.1.  Returning Capabilities in the Answer

  Potential and/or latent configuration attributes may be returned
  within an answer SDP to indicate the ability of the answerer to
  support alternative configurations of the corresponding stream(s).
  For example, an offer may include multiple potential configurations
  for a media stream and/or latent configurations for additional
  streams.  The corresponding answer will indicate (via an "acfg"
  attribute) the configuration accepted and used to construct the base
  configuration for each active media stream in the reply, but the
  reply MAY also contain potential and/or latent configuration
  attributes, with parameters, to indicate which other offered
  configurations would be acceptable.  This information is useful if it
  becomes desirable to reconfigure a media stream, e.g., to reduce
  resource consumption.

  When potential and/or latent configurations are returned in an
  answer, all numbering MUST refer to the configuration and capability
  attribute numbering of the offer.  The offered capability attributes
  need not be returned in the answer.  The answer MAY include
  additional capability attributes and/or configurations (with distinct
  numbering).  The parameter values of any returned "pcfg" or "lcfg"
  attributes MUST be a subset of those included in the offered
  configurations and/or those added by the answerer; values MAY be
  omitted only if they were indicated as alternative sets, or optional,
  in the original offer.  The parameter set indicated in the returned
  "acfg" attribute need not be repeated in a returned "pcfg" attribute.
  The answerer MAY return more than one "pcfg" attribute with the same
  configuration number if it is necessary to describe selected
  combinations of optional or alternative parameters.

  Similarly, one or more session capability attributes ("a=sescap") MAY
  be returned to indicate which of the offered session capabilities
  is/are supportable by the answerer (see Section 3.3.8).

  Note that, although the answerer MAY return capabilities beyond those
  included by the offerer, these capabilities MUST NOT be used to form
  any base level media description in the answer.  For this reason, it
  is advisable for the offerer to include most, if not all, potential
  and latent configurations it can support in the initial offer, unless
  the size of the resulting SDP is a concern.  Either party MAY later
  announce additional capabilities by renegotiating the session in a
  second offer/answer exchange.








Gilman, et al.               Standards Track                   [Page 26]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


3.3.6.2.  Payload Type Number Mapping

  When media format capabilities defined in "rmcap" attributes are used
  in potential configuration lines, the transport protocol uses RTP and
  it is necessary to assign payload type numbers.  In some cases, it is
  desirable to assign different payload type numbers to the same media
  format capability when used in different potential configurations.
  One example is when configurations for AVP and SAVP are offered: the
  offerer would like the answerer to use different payload type numbers
  for encrypted and unencrypted media, so the offerer can decide
  whether or not to render early media that arrives before the answer
  is received.

     For example, if use of AVP was selected by the answerer, then
     media received by the offerer is not encrypted; hence, it can be
     played out prior to receiving the answer.  Conversely, if SAVP was
     selected, cryptographic parameters and keying material present in
     the answer may be needed to decrypt received media.  If the offer
     configuration indicated that AVP media uses one set of payload
     types and SAVP a different set, then the offerer will know whether
     media received prior to the answer is encrypted or not by simply
     looking at the RTP payload type number in the received packet.

  This association of distinct payload type number(s) with different
  transport protocols requires a separate pcfg line for each protocol.
  Clearly, this technique cannot be used if the number of potential
  configurations exceeds the number of possible payload type numbers.

3.3.6.3.  Processing of Media-Format-Related Conventional Attributes for
         Potential Configurations

  When media capabilities negotiation is employed, SDP records are
  likely to contain conventional attributes such as "rtpmap", "fmtp",
  and other media-format-related lines, as well as capability
  attributes such as "rmcap", omcap, "mfcap", and "mscap" that map into
  those conventional attributes when invoked by a potential
  configuration.  In such cases, it MAY be appropriate to employ the
  delete-attributes option [RFC5939] in the attribute configuration
  list parameter in order to avoid the generation of conflicting "fmtp"
  attributes for a particular configuration.  Any media-specific
  attributes in the media block that refer to media formats not used by
  the potential configuration MUST be ignored.









Gilman, et al.               Standards Track                   [Page 27]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  For example:

     v=0
     o=- 25678 753849 IN IP4 192.0.2.1
     s=
     c=IN IP4 192.0.2.1
     t=0 0
     a=creq:med-v0
     m=audio 3456 RTP/AVP 0 18 100
     a=rtpmap:100 telephone-event
     a=fmtp:100 0-11
     a=rmcap:1 PCMU/8000
     a=rmcap:2 G729/8000
     a=rmcap:3 telephone-event/8000
     a=mfcap:3 0-15
     a=pcfg:1 m=2,3|1,3 a=-m pt=1:0,2:18,3:100
     a=pcfg:2

  In this example, PCMU is media capability 1, G729 is media capability
  2, and telephone-event is media capability 3.  The a=pcfg:1 line
  specifies that the preferred configuration is G.729 with extended
  DTMF events, second is G.711 mu-law with extended DTMF events, and
  the base media-level attributes are to be deleted.  Intermixing of
  G.729, G.711, and "commercial" DTMF events is least preferred (the
  base configuration provided by the "m=" line, which is, by default,
  the least preferred configuration).  The "rtpmap" and "fmtp"
  attributes of the base configuration are replaced by the "rmcap" and
  "mfcap" attributes when invoked by the proposed configuration.

  If the preferred configuration is selected, the SDP answer will look
  like the following

     v=0
     o=- 25678 753849 IN IP4 192.0.2.1
     s=
     c=IN IP4 192.0.2.1
     t=0 0
     a=csup:med-v0
     m=audio 3456 RTP/AVP 18 100
     a=rtpmap:100 telephone-event/8000
     a=fmtp:100 0-15
     a=acfg:1 m=2,3 pt=1:0,2:18,3:100









Gilman, et al.               Standards Track                   [Page 28]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


3.3.7.  Substitution of Media Payload Type Numbers in Capability
       Attribute Parameters

  In some cases, for example, when an RFC 2198 [RFC2198] redundancy
  audio subtype (RED) capability is defined in an "mfcap" attribute,
  the parameters to an attribute may contain payload type numbers.  Two
  options are available for specifying such payload type numbers.  They
  may be expressed explicitly, in which case they are bound to actual
  payload types by means of the payload type number parameter (pt=) in
  the appropriate potential or latent configuration.  For example, the
  following SDP fragment defines a potential configuration with
  redundant G.711 mu-law

     m=audio 45678 RTP/AVP 0
     a=rtpmap:0 PCMU/8000
     a=rmcap:1 PCMU/8000
     a=rmcap:2 RED/8000
     a=mfcap:2 0/0
     a=pcfg:1 m=2,1 pt=2:98,1:0

  The potential configuration is then equivalent to

     m=audio 45678 RTP/AVP 98 0
     a=rtpmap:0 PCMU/8000
     a=rtpmap:98 RED/8000
     a=fmtp:98 0/0

  A more general mechanism is provided via the parameter substitution
  rule.  When an "mfcap", "mscap", or "acap" attribute is processed,
  its arguments will be scanned for a payload type number escape
  sequence of the following form (in ABNF):

     ptn-esc = "%m=" media-cap-num "%" ; defined in Section 3.3.1

  If the sequence is found, the sequence is replaced by the payload
  type number assigned to the media capability number, as specified by
  the "pt=" parameter in the selected potential configuration; only
  actual payload type numbers are supported -- wildcards are excluded.
  The sequence "%%" (null digit string) is replaced by a single percent
  sign and processing continues with the next character, if any.











Gilman, et al.               Standards Track                   [Page 29]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  For example, the above offer sequence could have been written as

     m=audio 45678 RTP/AVP 0
     a=rtpmap:0 PCMU/8000
     a=rmcap:1 PCMU/8000
     a=rmcap:2 RED/8000
     a=mfcap:2 %m=1%/%m=1%
     a=pcfg:1 m=2,1 pt=2:98,1:0

  and the equivalent SDP is the same as above.

3.3.8.  The Session Capability Attribute

  Potential and latent configurations enable offerers and answerers to
  express a wide range of alternative configurations for current and
  future negotiation.  However, in practice, it may not be possible to
  support all combinations of these configurations.

  The session capability attribute provides a means for the offerer
  and/or the answerer to specify combinations of specific media stream
  configurations that it is willing and able to support.  Each session
  capability in an offer or answer MAY be expressed as a list of
  required potential configurations, and MAY include a list of optional
  potential and/or latent configurations.

  The choices of session capabilities may be based on processing load,
  total bandwidth, or any other criteria of importance to the
  communicating parties.  If the answerer supports media capabilities
  negotiation, and session configurations are offered, it MUST accept
  one of the offered configurations, or it MUST refuse the session.
  Therefore, if the offer includes any session capabilities, it SHOULD
  include all the session capabilities the offerer is willing to
  support.

  The session capability attribute is a session-level attribute
  described by

      "a=sescap:" <session num> <list of configs>













Gilman, et al.               Standards Track                   [Page 30]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  which corresponds to the standard value attribute definition with

          att-field        = "sescap"
          att-value        = session-num 1*WSP list-of-configs
                             [1*WSP optional-configs]
          session-num      = NonZeroDigit *9(DIGIT)  ; DIGIT defined
                                                     ; in RFC 5234
          list-of-configs  = alt-config *("," alt-config)
          optional-configs = "[" list-of-configs "]"
          alt-config       = config-number *("|" config-number)

  The session-num identifies the session: a lower-number session is
  preferred over a higher-number session, and leading zeroes are not
  permitted.  Each alt-config list specifies alternative media
  configurations within the session; preference is based on config-num
  as specified in RFC 5939 [RFC5939].  Note that the session preference
  order, when present, takes precedence over the individual media
  stream configuration preference order.

  Use of session capability attributes requires that configuration
  numbers assigned to potential and latent configurations MUST be
  unique across the entire session; RFC 5939 [RFC5939] requires only
  that "pcfg" configuration numbers be unique within a media
  description.  Also, leading zeroes are not permitted.

  As an example, consider an endpoint that is capable of supporting an
  audio stream with either one H.264 video stream or two H.263 video
  streams with a floor control stream.  In the latter case, the second
  video stream is optional.  The SDP offer might look like the
  following (offering audio, an H.263 video streams, BFCP and another
  optional H.263 video stream) -- the empty lines are added for
  readability only (not part of valid SDP):

     v=0
     o=- 25678 753849 IN IP4 192.0.2.1
     s=
     c=IN IP4 192.0.2.1
     t=0 0
     a=creq:med-v0
     a=sescap:2 1,2,5,[3]
     a=sescap:1 1,4

     m=audio 54322 RTP/AVP 0
     a=rtpmap:0 PCMU/8000
     a=pcfg:1






Gilman, et al.               Standards Track                   [Page 31]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


     m=video 22344 RTP/AVP 102
     a=rtpmap:102 H263-1998/90000
     a=fmtp:102 CIF=4;QCIF=2;F=1;K=1
     i=main video stream
     a=label:11
     a=pcfg:2
     a=rmcap:1 H264/90000
     a=mfcap:1 profile-level-id=42A01E; packetization-mode=2
     a=acap:1 label:13
     a=pcfg:4 m=1 a=1 pt=1:104

     m=video 33444 RTP/AVP 103
     a=rtpmap:103 H263-1998/90000
     a=fmtp:103 CIF=4;QCIF=2;F=1;K=1
     i=secondary video (slides)
     a=label:12
     a=pcfg:3

     m=application 33002 TCP/BFCP *
     a=setup:passive
     a=connection:new
     a=floorid:1 m-stream:11 12
     a=floor-control:s-only
     a=confid:4321
     a=userid:1234
     a=pcfg:5

  If the answerer understands MediaCapNeg, but cannot support the
  Binary Floor Control Protocol, then it would respond with (invalid
  empty lines in SDP included again for readability):

     v=0
     o=- 25678 753849 IN IP4 192.0.2.1
     s=
     c=IN IP4 192.0.2.22
     t=0 0
     a=csup:med-v0
     a=sescap:1 1,4

     m=audio 23456 RTP/AVP 0
     a=rtpmap:0 PCMU/8000
     a=acfg:1

     m=video 41234 RTP/AVP 104
     a=rtpmap:104 H264/90000
     a=fmtp:104 profile-level-id=42A01E; packetization-mode=2
     a=acfg:4 m=1 a=1 pt=1:104




Gilman, et al.               Standards Track                   [Page 32]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


     m=video 0 RTP/AVP 103
     a=acfg:3

     m=application 0 TCP/BFCP *
     a=acfg:5

  An endpoint that doesn't support media capabilities negotiation, but
  does support H.263 video, would respond with one or two H.263 video
  streams.  In the latter case, the answerer may issue a second offer
  to reconfigure the session to one audio and one video channel using
  H.264 or H.263.

  Session capabilities can include latent capabilities as well.  Here's
  a similar example in which the offerer wishes to initially establish
  an audio stream, and prefers to later establish two video streams
  with chair control.  If the answerer doesn't understand Media CapNeg,
  or cannot support the dual video streams or flow control, then it may
  support a single H.264 video stream.  Note that establishment of the
  most favored configuration will require two offer/answer exchanges.

     v=0
     o=- 25678 753849 IN IP4 192.0.2.1
     s=
     c=IN IP4 192.0.2.1
     t=0 0
     a=creq:med-v0
     a=sescap:1 1,3,4,5
     a=sescap:2 1,2
     a=sescap:3 1

     a=rmcap:1 H263-1998/90000
     a=mfcap:1 CIF=4;QCIF=2;F=1;K=1
     a=tcap:1 RTP/AVP TCP/BFCP
     m=audio 54322 RTP/AVP 0
     a=rtpmap:0 PCMU/8000
     a=pcfg:1
     m=video 22344 RTP/AVP 102
     a=rtpmap:102 H264/90000
     a=fmtp:102 profile-level-id=42A01E; packetization-mode=2
     a=label:11
     a=content:main
     a=pcfg:2
     a=lcfg:3 mt=video t=1 m=1 a=31,32
     a=acap:31 label:12
     a=acap:32 content:main
     a=lcfg:4 mt=video t=1 m=1 a=41,42
     a=acap:41 label:13
     a=acap:42 content:slides



Gilman, et al.               Standards Track                   [Page 33]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


     a=lcfg:5 mt=application m=51 t=51
     a=tcap:51 TCP/BFCP
     a=omcap:51 *
     a=acap:51 setup:passive
     a=acap:52 connection:new
     a=acap:53 floorid:1 m-stream:12 13
     a=acap:54 floor-control:s-only
     a=acap:55 confid:4321
     a=acap:56 userid:1234

  In this example, the default offer, as seen by endpoints that do not
  understand capabilities negotiation, proposes a PCMU audio stream and
  an H.264 video stream.  Note that the offered lcfg lines for the
  video streams don't carry "pt=" parameters because they're not needed
  (payload type numbers will be assigned in the offer/answer exchange
  that establishes the streams).  Note also that the three "rmcap",
  "mfcap", and "tcap" attributes used by "lcfg:3" and "lcfg:4" are
  included at the session level so they may be referenced by both
  latent configurations.  As per Section 3.3, the media attributes
  generated from the "rmcap", "mfcap", and "tcap" attributes are always
  media-level attributes.  If the answerer supports Media CapNeg, and
  supports the most desired configuration, it would return the
  following SDP:

     v=0
     o=- 25678 753849 IN IP4 192.0.2.1
     s=
     c=IN IP4 192.0.2.22
     t=0 0
     a=csup:med-v0
     a=sescap:1 1,3,4,5
     a=sescap:2 1,2
     a=sescap:3 1
     m=audio 23456 RTP/AVP 0
     a=rtpmap:0 PCMU/8000
     a=acfg:1
     m=video 0 RTP/AVP 102
     a=pcfg:2
     a=lcfg:3 mt=video t=1 m=1 a=31,32
     a=lcfg:4 mt=video t=1 m=1 a=41,42
     a=lcfg:5 mt=application t=2

  This exchange supports immediate establishment of an audio stream for
  preliminary conversation.  This exchange would presumably be followed
  at the appropriate time with a "reconfiguration" offer/answer
  exchange to add the video and chair control streams.





Gilman, et al.               Standards Track                   [Page 34]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


3.4.  Offer/Answer Model Extensions

  In this section, we define extensions to the offer/answer model
  defined in RFC 3264 [RFC3264] and RFC 5939 [RFC5939] to allow for
  media format and associated parameter capabilities, latent
  configurations, and acceptable combinations of media stream
  configurations to be used with the SDP capability negotiation
  framework.  Note that the procedures defined in this section extend
  the offer/answer procedures defined in RFC 5939 [RFC5939] Section 6;
  those procedures form a baseline set of capability negotiation
  offer/answer procedures that MUST be followed, subject to the
  extensions defined here.

  SDP capability negotiation [RFC5939] provides a relatively compact
  means to offer the equivalent of an ordered list of alternative
  configurations for offered media streams (as would be described by
  separate "m=" lines and associated attributes).  The attributes
  "acap", "mscap", "mfcap", "omcap", and "rmcap" are designed to map
  somewhat straightforwardly into equivalent "m=" lines and
  conventional attributes when invoked by a "pcfg", "lcfg", or "acfg"
  attribute with appropriate parameters.  The "a=pcfg:" lines, along
  with the "m=" line itself, represent offered media configurations.
  The "a=lcfg:" lines represent alternative capabilities for future
  use.

3.4.1.  Generating the Initial Offer

  The media capabilities negotiation extensions defined in this
  document cover the following categories of features:

  o  Media format capabilities and associated parameters ("rmcap",
     "omcap", "mfcap", and "mscap" attributes)

  o  Potential configurations using those media format capabilities and
     associated parameters

  o  Latent media streams ("lcfg" attribute)

  o  Acceptable combinations of media stream configurations ("sescap"
     attribute).

  The high-level description of the operation is as follows:

  When an endpoint generates an initial offer and wants to use the
  functionality described in the current document, it SHOULD identify
  and define the media formats and associated parameters it can support
  via the "rmcap", "omcap", "mfcap", and "mscap" attributes.  The SDP
  media line(s) ("m=") should be made up with the actual configuration



Gilman, et al.               Standards Track                   [Page 35]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  to be used if the other party does not understand capability
  negotiations (by default, this is the least preferred configuration).
  Typically, the media line configuration will contain the minimum
  acceptable configuration from the offerer's point of view.

  Preferred configurations for each media stream are identified
  following the media line.  The present offer may also include latent
  configuration ("lcfg") attributes, at the media level, describing
  media streams and/or configurations the offerer is not now offering
  but that it is willing to support in a future offer/answer exchange.
  A simple example might be the inclusion of a latent video
  configuration in an offer for an audio stream.

  Lastly, if the offerer wishes to impose restrictions on the
  combinations of potential configurations to be used, it will include
  session capability ("sescap") attributes indicating those.

  If the offerer requires the answerer to understand the media
  capability extensions, the offerer MUST include a "creq" attribute
  containing the value "med-v0".  If media capability negotiation is
  required only for specific media descriptions, the "med-v0" value
  MUST be provided only in "creq" attributes within those media
  descriptions, as described in RFC 5939 [RFC5939].

  Below, we provide a more detailed description of how to construct the
  offer SDP.

3.4.1.1.  Offer with Media Capabilities

  For each RTP-based media format the offerer wants to include as a
  media format capability, the offer MUST include an "rmcap" attribute
  for the media format as defined in Section 3.3.1.

  For each non-RTP-based media format the offer wants to include as a
  media format capability, the offer MUST include an "omcap" attribute
  for the media format as defined in Section 3.3.1.

  Since the media capability number space is shared between the "rmcap"
  and "omcap" attributes, each media capability number provided
  (including ranges) MUST be unique in the entire SDP.

  If an "fmtp" parameter value is needed for a media format (whether or
  not it is RTP based) in a media capability, then the offer MUST
  include one or more "mfcap" parameters with the relevant "fmtp"
  parameter values for that media format as defined in Section 3.3.2.
  When multiple "mfcap" parameters are provided for a given media
  capability, they MUST be provided in accordance with the
  concatenation rules in Section 3.3.2.1.



Gilman, et al.               Standards Track                   [Page 36]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  For each of the media format capabilities above, the offer MAY
  include one or more "mscap" parameters with attributes needed for
  those specific media formats as defined in Section 3.3.3.  Such
  attributes will be instantiated at the media level; hence, session-
  level-only attributes MUST NOT be used in the "mscap" parameter.  The
  "mscap" parameter MUST NOT include an "rtpmap" or "fmtp" attribute
  ("rmcap" and "mfcap" are used instead).

  If the offerer wants to limit the relevance (and use) of a media
  format capability or parameter to a particular media stream, the
  media format capability or parameter MUST be provided within the
  corresponding media description.  Otherwise, the media format
  capabilities and parameters MUST be provided at the session level.
  Note, however, that the attribute or parameter embedded in these will
  always be instantiated at the media level.

     This is due to those parameters being effectively media-level
     parameters.  If session-level attributes are needed, the "acap"
     attribute defined in RFC 5939 [RFC5939] can be used; however, it
     does not provide for media-format-specific instantiation.

  Inclusion of the above does not constitute an offer to use the
  capabilities; a potential configuration is needed for that.  If the
  offerer wants to offer one or more of the media capabilities above,
  they MUST be included as part of a potential configuration ("pcfg")
  attribute as defined in Section 3.3.4.  Each potential configuration
  MUST include a config-number, and each config-number MUST be unique
  in the entire SDP (note that this differs from RFC 5939 [RFC5939],
  which only requires uniqueness within a media description).  Also,
  the config-number MUST NOT overlap with any config-number used by a
  latent configuration in the SDP.  As described in RFC 5939 [RFC5939],
  lower config-numbers indicate a higher preference; the ordering still
  applies within a given media description only though.

  For a media capability to be included in a potential configuration,
  there MUST be an "m=" parameter in the "pcfg" attribute referencing
  the media capability number in question.  When one or more media
  capabilities are included in an offered potential configuration
  ("pcfg"), they completely replace the list of media formats offered
  in the actual configuration ("m=" line).  Any attributes included for
  those formats remain in the SDP though (e.g., "rtpmap", "fmtp",
  etc.).  For non-RTP-based media formats, the format-name (from the
  "omcap" media capability) is simply added to the "m=" line as a media
  format (e.g., t38).  For RTP-based media, payload type mappings MUST
  be provided by use of the "pt" parameter in the potential
  configuration (see Section 3.3.4.2); payload type escaping may be
  used in "mfcap", "mscap", and "acap" attributes as defined in Section
  3.3.7.



Gilman, et al.               Standards Track                   [Page 37]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  Note that the "mt" parameter MUST NOT be used with the "pcfg"
  attribute (since it is defined for the "lcfg" attribute only); the
  media type in a potential configuration cannot be changed from that
  of the encompassing media description.

3.4.1.2.  Offer with Latent Configuration

  If the offerer wishes to offer one or more latent configurations for
  future use, the offer MUST include a latent configuration attribute
  ("lcfg") for each as defined in Section 3.3.6.

  Each "lcfg" attribute

  o  MUST be specified at the media level

  o  MUST include a config-number that is unique in the entire SDP
     (including for any potential configuration attributes).  Note that
     config-numbers in latent configurations do not indicate any
     preference order

  o  MUST include a media type ("mt")

  o  MUST reference a valid transport capability ("t")

  Each "lcfg" attribute MAY include additional capability references,
  which may refer to capabilities anywhere in the session description,
  subject to any restrictions normally associated with such
  capabilities.  For example, a media-level attribute capability must
  be present at the media level in some media description in the SDP.
  Note that this differs from the potential configuration attribute,
  which cannot validly refer to media-level capabilities in another
  media description (per RFC 5939 [RFC5939], Section 3.5.1).

     Potential configurations constitute an actual offer and may
     instantiate a referenced capability.  Latent configurations are
     not actual offers; hence, they cannot instantiate a referenced
     capability.  Therefore, it is safe for those to refer to
     capabilities in another media description.

3.4.1.3.  Offer with Configuration Combination Restrictions

  If the offerer wants to indicate restrictions or preferences among
  combinations of potential and/or latent configurations, a session
  capability ("sescap") attribute MUST be provided at the session level
  for each such combination as described in Section 3.3.8.  Each
  "sescap" attribute MUST include a session-num that is unique in the





Gilman, et al.               Standards Track                   [Page 38]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  entire SDP; the lower the session-num the more preferred that
  combination is.  Furthermore, "sescap" preference order takes
  precedence over any order specified in individual "pcfg" attributes.

     For example, if we have pcfg-1 and pcfg-2, and sescap-1 references
     pcfg-2, whereas sescap-2 references pcfg-1, then pcfg-2 will be
     the most preferred potential configuration.  Without the sescap,
     pcfg-1 would be the most preferred.

3.4.2.  Generating the Answer

  When receiving an offer, the answerer MUST check the offer for "creq"
  attributes containing the value "med-v0"; answerers compliant with
  this specification will support this value in accordance with the
  procedures specified in RFC 5939 [RFC5939].

  The SDP MAY contain

  o  Media format capabilities and associated parameters ("rmcap",
     "omcap", "mfcap", and "mscap" attributes)

  o  Potential configurations using those media format capabilities and
     associated parameters

  o  Latent media streams ("lcfg" attribute)

  o  Acceptable combinations of media stream configurations ("sescap"
     attribute)

  The high-level informative description of the operation is as
  follows:

  When the answering party receives the offer, if it supports the
  required capability negotiation extensions, it should select the
  most-preferred configuration it can support for each media stream,
  and build its answer accordingly.  The configuration selected for
  each accepted media stream is placed into the answer as a media line
  with associated parameters and attributes.  If a proposed
  configuration is chosen for a given media stream, the answer must
  contain an actual configuration ("acfg") attribute for that media
  stream to indicate which offered "pcfg" attribute was used to build
  the answer.  The answer should also include any potential or latent
  configurations the answerer can support, especially any
  configurations compatible with other potential or latent
  configurations received in the offer.  The answerer should make note
  of those configurations it might wish to offer in the future.





Gilman, et al.               Standards Track                   [Page 39]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  Below we provide a more detailed normative description of how the
  answerer processes the offer SDP and generates an answer SDP.

3.4.2.1.  Processing Media Capabilities and Potential Configurations

  The answerer MUST first determine if it needs to perform media
  capability negotiation by examining the SDP for valid and preferred
  potential configuration attributes that include media configuration
  parameters (i.e., an "m" parameter in the "pcfg" attribute).

  Such a potential configuration is valid if

  1.  It is valid according to the rules defined in RFC 5939 [RFC5939].

  2.  It contains a config-number that is unique in the entire SDP and
      does not overlap with any latent configuration config-numbers.

  3.  All media format capabilities ("rmcap" or "omcap"), media format
      parameter capabilities ("mfcap"), and media-specific capabilities
      ("mscap") referenced by the potential configuration ("m"
      parameter) are valid themselves (as defined in Sections 3.3.1,
      3.3.2, and 3.3.3) and each of them is provided either at the
      session level or within this particular media description.

  4.  All RTP-based media format capabilities ("rmcap") have a
      corresponding payload type ("pt") parameter in the potential
      configuration that results in mapping to a valid payload type
      that is unique within the resulting SDP.

  5.  Any concatenation (see Section 3.3.2.1) and substitution (see
      Section 3.3.7) applied to any capability ("mfcap", "mscap", or
      "acap") referenced by this potential configuration results in a
      valid SDP.

  Note that since SDP does not interpret the value of "fmtp"
  parameters, any resulting "fmtp" parameter value will be considered
  valid.

  Secondly, the answerer MUST determine the order in which potential
  configurations are to be negotiated.  In the absence of any session
  capability ("sescap") attributes, this simply follows the rules of
  RFC 5939 [RFC5939], with a lower config-number within a media
  description being preferred over a higher one.  If a valid "sescap"
  attribute is present, the preference order provided in the "sescap"
  attribute MUST take precedence.  A "sescap" attribute is considered
  valid if





Gilman, et al.               Standards Track                   [Page 40]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  1.  It adheres to the rules provided in Section 3.3.8.

  2.  All the configurations referenced by the "sescap" attribute are
      valid themselves (note that this can include the actual,
      potential, and latent configurations).

  The answerer MUST now process the offer for each media stream based
  on the most preferred valid potential configuration in accordance
  with the procedures specified in RFC 5939 [RFC5939], Section 3.6.2,
  and further extended below:

  o  If one or more media format capabilities are included in the
     potential configuration, then they replace all media formats
     provided in the "m=" line for that media description.  For non-
     RTP-based media formats ("omcap"), the format-name is added.  For
     RTP-based media formats ("rmcap"), the payload-type specified in
     the payload-type mapping ("pt") is added and a corresponding
     "rtpmap" attribute is added to the media description.

  o  If one or more media format parameter capabilities are included in
     the potential configuration, then the corresponding "fmtp"
     attributes are added to the media description.  Note that this
     inclusion is done indirectly via the media format capability.

  o  If one or more media-specific capabilities are included in the
     potential configuration, then the corresponding attributes are
     added to the media description.  Note that this inclusion is done
     indirectly via the media format capability.

  o  When checking to see if the answerer supports a given potential
     configuration that includes one or more media format capabilities,
     the answerer MUST support at least one of the media formats
     offered.  If he does not, the answerer MUST proceed to the next
     potential configuration based on the preference order that
     applies.

  o  If session capability ("sescap") preference ordering is included,
     then the potential configuration selection process MUST adhere to
     the ordering provided.  Note that this may involve coordinated
     selection of potential configurations between media descriptions.
     The answerer MUST accept one of the offered sescap combinations
     (i.e., all the required potential configurations specified) or it
     MUST reject the entire session.








Gilman, et al.               Standards Track                   [Page 41]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  Once the answerer has selected a valid and supported offered
  potential configuration for all of the media streams (or has fallen
  back to the actual configuration plus any added session attributes),
  the answerer MUST generate a valid answer SDP as described in RFC
  5939 [RFC5939], Section 3.6.2, and further extended below:

  o  Additional answer capabilities and potential configurations MAY be
     returned in accordance with Section 3.3.6.1.  Capability numbers
     and configuration numbers for those MUST be distinct from the ones
     used in the offer SDP.

  o  Latent configuration processing and answer generation MUST be
     performed, as specified below.

  o  Session capability specification for the potential and latent
     configurations in the answer MAY be included (see Section 3.3.8).

3.4.2.2.  Latent Configuration Processing

  The answerer MUST determine if it needs to perform any latent
  configuration processing by examining the SDP for valid latent
  configuration attributes ("lcfg").  An "lcfg" attribute is considered
  valid if:

  o  It adheres to the description in Section 3.3.5.

  o  It includes a config-number that is unique in the entire SDP and
     does not overlap with any potential configuration config-number.

  o  It includes a valid media type ("mt=").

  o  It references a valid transport capability ("t=").

  o  All other capabilities referenced by it are valid.

  For each such valid latent configuration in the offer, the answerer
  checks to see if it could support the latent configuration in a
  subsequent offer/answer exchange.  If so, it includes the latent
  configuration with the same configuration number in the answer,
  similar to the way potential configurations are processed and the
  selected one returned in an actual configuration attribute (see RFC
  5939 [RFC5939]).  If the answerer supports only a (non-mandatory)
  subset of the parameters offered in a latent configuration, the
  answer latent configuration will include only those parameters
  supported (similar to "acfg" processing).  Note that latent
  configurations do not constitute an actual offer at this point in
  time; they merely indicate additional configurations that could be
  supported.



Gilman, et al.               Standards Track                   [Page 42]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  If a session capability ("sescap") attribute is included and it
  references a latent configuration, then the answerer processing of
  that latent configuration must be done within the constraints
  specified by that session capability.  That is, it must be possible
  to support it at the same time as any required (i.e., non-optional)
  potential configurations in the session capability.  The answerer may
  in turn add his own sescap indications in the answer as well.

3.4.3.  Offerer Processing of the Answer

  The offerer MUST process the answer in accordance with Section 3.6.3
  of RFC 5939 [RFC5939] and the further explanation below.

  When the offerer processes the answer SDP based on a valid actual
  configuration attribute in the answer, and that valid configuration
  includes one or more media capabilities, the processing MUST
  furthermore be done as if the offer was sent using those media
  capabilities instead of the actual configuration.  In particular, the
  media formats in the "m=" line, and any associated payload type
  mappings ("rtpmap"), "fmtp" parameters ("mfcap"), and media-specific
  attributes ("mscap") MUST be used.  Note that this may involve use of
  concatenation and substitution rules (see Sections 3.3.2.1 and
  3.3.7).  The actual configuration attribute may also be used to infer
  the lack of acceptability of higher-preference configurations that
  were not chosen, subject to any constraints provided by a session
  capability ("sescap") attribute in the offer.  Note that the SDP
  capability negotiation base specification [RFC5939] requires the
  answerer to choose the highest-preference configuration it can
  support, subject to local policies.

  When the offerer receives the answer, it SHOULD furthermore make note
  of any capabilities and/or latent configurations included for future
  use, and any constraints on how those may be combined.

3.4.4.  Modifying the Session

  If, at a later time, one of the parties wishes to modify the
  operating parameters of a session, e.g., by adding a new media
  stream, or by changing the properties used on an existing stream, it
  can do so via the mechanisms defined for offer/answer [RFC3264].  If
  the initiating party has remembered the codecs, potential
  configurations, latent configurations, and session capabilities
  provided by the other party in the earlier negotiation, it MAY use
  this knowledge to maximize the likelihood of a successful
  modification of the session.  Alternatively, the initiator MAY
  perform a new capabilities exchange as part of the reconfiguration.





Gilman, et al.               Standards Track                   [Page 43]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  In such a case, the new capabilities will replace the previously
  negotiated capabilities.  This may be useful if conditions change on
  the endpoint.

4.  Examples

  In this section, we provide examples showing how to use the media
  capabilities with the SDP capability negotiation.

4.1.  Alternative Codecs

  This example provides a choice of one of six variations of the
  Adaptive Multi-Rate codec.  In this example, the default
  configuration as specified by the media line is the same as the most
  preferred configuration.  Each configuration uses a different payload
  type number so the offerer can interpret early media.

     v=0
     o=- 25678 753849 IN IP4 192.0.2.1
     s=
     c=IN IP4 192.0.2.1
     t=0 0
     a=creq:med-v0
     m=audio 54322 RTP/AVP 96
     a=rtpmap:96 AMR-WB/16000/1
     a=fmtp:96 mode-change-capability=1; max-red=220; \
     mode-set=0,2,4,7
     a=rmcap:1,3,5 audio AMR-WB/16000/1
     a=rmcap:2,4,6 audio AMR/8000/1
     a=mfcap:1,2,3,4 mode-change-capability=1
     a=mfcap:5,6 mode-change-capability=2
     a=mfcap:1,2,3,5 max-red=220
     a=mfcap:3,4,5,6 octet-align=1
     a=mfcap:1,3,5 mode-set=0,2,4,7
     a=mfcap:2,4,6 mode-set=0,3,5,6
     a=pcfg:1 m=1 pt=1:96
     a=pcfg:2 m=2 pt=2:97
     a=pcfg:3 m=3 pt=3:98
     a=pcfg:4 m=4 pt=4:99
     a=pcfg:5 m=5 pt=5:100
     a=pcfg:6 m=6 pt=6:101

  In the above example, media capability 1 could have been excluded
  from the first "rmcap" declaration and from the corresponding "mfcap"
  attributes, and the "pcfg:1" attribute line could have been simply
  "pcfg:1".





Gilman, et al.               Standards Track                   [Page 44]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  The next example offers a video stream with three options of H.264
  and four transports.  It also includes an audio stream with different
  audio qualities: four variations of AMR, or AC3.  The offer looks
  something like the following:

     v=0
     o=- 25678 753849 IN IP4 192.0.2.1
     s=An SDP Media NEG example
     c=IN IP4 192.0.2.1
     t=0 0
     a=creq:med-v0
     a=ice-pwd:speEc3QGZiNWpVLFJhQX
     m=video 49170 RTP/AVP 100
     c=IN IP4 192.0.2.56
     a=maxprate:1000
     a=rtcp:51540
     a=sendonly
     a=candidate 12345 1 UDP 9 192.0.2.56 49170 host
     a=candidate 23456 2 UDP 9 192.0.2.56 51540 host
     a=candidate 34567 1 UDP 7 198.51.100.1 41345 srflx raddr \
     192.0.2.56 rport 49170
     a=candidate 45678 2 UDP 7 198.51.100.1 52567 srflx raddr \
     192.0.2.56 rport 51540
     a=candidate 56789 1 UDP 3 192.0.2.100 49000 relay raddr \
     192.0.2.56 rport 49170
     a=candidate 67890 2 UDP 3 192.0.2.100 49001 relay raddr \
     192.0.2.56 rport 51540
     b=AS:10000
     b=TIAS:10000000
     b=RR:4000
     b=RS:3000
     a=rtpmap:100 H264/90000
     a=fmtp:100 profile-level-id=42A01E; packetization-mode=2; \
     sprop-parameter-sets=Z0IACpZTBYmI,aMljiA==; \
     sprop-interleaving-depth=45; sprop-deint-buf-req=64000; \
     sprop-init-buf-time=102478; deint-buf-cap=128000
     a=tcap:1 RTP/SAVPF RTP/SAVP RTP/AVPF
     a=rmcap:1-3,7-9 H264/90000
     a=rmcap:4-6 rtx/90000
     a=mfcap:1-9 profile-level-id=42A01E
     a=mfcap:1-9 aMljiA==
     a=mfcap:1,4,7 packetization-mode=0
     a=mfcap:2,5,8 packetization-mode=1
     a=mfcap:3,6,9 packetization-mode=2
     a=mfcap:1-9 sprop-parameter-sets=Z0IACpZTBYmI
     a=mfcap:1,7 sprop-interleaving-depth=45; \
     sprop-deint-buf-req=64000; sprop-init-buf-time=102478; \
     deint-buf-cap=128000



Gilman, et al.               Standards Track                   [Page 45]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


     a=mfcap:4 apt=100
     a=mfcap:5 apt=99
     a=mfcap:6 apt=98
     a=mfcap:4-6 rtx-time=3000
     a=mscap:1-6 rtcp-fb nack
     a=acap:1 crypto:1 AES_CM_128_HMAC_SHA1_80 \
     inline:d0RmdmcmVCspeEc3QGZiNWpVLFJhQX1cfHAwJSoj|220|1:32
     a=pcfg:1 t=1 m=1,4 a=1 pt=1:100,4:97
     a=pcfg:2 t=1 m=2,5 a=1 pt=2:99,4:96
     a=pcfg:3 t=1 m=3,6 a=1 pt=3:98,6:95
     a=pcfg:4 t=2 m=7 a=1 pt=7:100
     a=pcfg:5 t=2 m=8 a=1 pt=8:99
     a=pcfg:6 t=2 m=9 a=1 pt=9:98
     a=pcfg:7 t=3 m=1,3 pt=1:100,4:97
     a=pcfg:8 t=3 m=2,4 pt=2:99,4:96
     a=pcfg:9 t=3 m=3,6 pt=3:98,6:95
     m=audio 49176 RTP/AVP 101 100 99 98
     c=IN IP4 192.0.2.56
     a=ptime:60
     a=maxptime:200
     a=rtcp:51534
     a=sendonly
     a=candidate 12345 1 UDP 9 192.0.2.56 49176 host
     a=candidate 23456 2 UDP 9 192.0.2.56 51534 host
     a=candidate 34567 1 UDP 7 198.51.100.1 41348 srflx \
     raddr 192.0.2.56 rport 49176
     a=candidate 45678 2 UDP 7 198.51.100.1 52569 srflx \
     raddr 192.0.2.56 rport 51534
     a=candidate 56789 1 UDP 3 192.0.2.100 49002 relay \
     raddr 192.0.2.56 rport 49176
     a=candidate 67890 2 UDP 3 192.0.2.100 49003 relay \
     raddr 192.0.2.56 rport 51534
     b=AS:512
     b=TIAS:512000
     b=RR:4000
     b=RS:3000
     a=maxprate:120
     a=rtpmap:98 AMR-WB/16000
     a=fmtp:98 octet-align=1; mode-change-capability=2
     a=rtpmap:99 AMR-WB/16000
     a=fmtp:99 octet-align=1; crc=1; mode-change-capability=2
     a=rtpmap:100 AMR-WB/16000/2
     a=fmtp:100 octet-align=1; interleaving=30
     a=rtpmap:101 AMR-WB+/72000/2
     a=fmtp:101 interleaving=50; int-delay=160000;
     a=rmcap:14 ac3/48000/6
     a=acap:23 crypto:1 AES_CM_128_HMAC_SHA1_80 \
     inline:d0RmdmcmVCspeEc3QGZiNWpVLFJhQX1cfHAwJSoj|220|1:32



Gilman, et al.               Standards Track                   [Page 46]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


     a=tcap:4 RTP/SAVP
     a=pcfg:10 t=4 a=23
     a=pcfg:11 t=4 m=14 a=23 pt=14:102

  This offer illustrates the advantage in compactness that arises if
  one can avoid deleting the base configuration attributes and
  recreating them in "acap" attributes for the potential
  configurations.

4.2.  Alternative Combinations of Codecs (Session Configurations)

  If an endpoint has limited signal processing capacity, it might be
  capable of supporting, say, a G.711 mu-law audio stream in
  combination with an H.264 video stream, or a G.729B audio stream in
  combination with an H.263-1998 video stream.  It might then issue an
  offer like the following:

     v=0
     o=- 25678 753849 IN IP4 192.0.2.1
     s=
     c=IN IP4 192.0.2.1
     t=0 0
     a=creq:med-v0
     a=sescap:1 2,4
     a=sescap:2 1,3
     m=audio 54322 RTP/AVP 18
     a=rtpmap:18 G729/8000
     a=fmtp:18 annexb=yes
     a=rmcap:1 PCMU/8000
     a=pcfg:1 m=1 pt=1:0
     a=pcfg:2
     m=video 54344 RTP/AVP 100
     a=rtpmap:100 H263-1998/90000
     a=rmcap:2 H264/90000
     a=mfcap:2 profile-level-id=42A01E; packetization-mode=2
     a=pcfg:3 m=2 pt=2:101
     a=pcfg:4

  Note that the preferred session configuration (and the default as
  well) is G.729B with H.263.  This overrides the individual media
  stream preferences that are PCMU and H.264 by the potential
  configuration numbering rule.

4.3.  Latent Media Streams

  Consider a case in which the offerer can support either G.711 mu-law
  or G.729B, along with DTMF telephony events for the 12 common
  touchtone signals, but is willing to support simple G.711 mu-law



Gilman, et al.               Standards Track                   [Page 47]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  audio as a last resort.  In addition, the offerer wishes to announce
  its ability to support video and Message Session Relay Protocol
  (MSRP) in the future, but does not wish to offer a video stream or an
  MSRP stream at present.  The offer might look like the following:

     v=0
     o=- 25678 753849 IN IP4 192.0.2.1
     s=
     c=IN IP4 192.0.2.1
     t=0 0
     a=creq:med-v0
     m=audio 23456 RTP/AVP 0
     a=rtpmap:0 PCMU/8000
     a=rmcap:1 PCMU/8000
     a=rmcap:2 G729/8000
     a=rmcap:3 telephone-event/8000
     a=mfcap:3 0-11
     a=pcfg:1 m=1,3|2,3 pt=1:0,2:18,3:100
     a=lcfg:2 mt=video t=1 m=10|11
     a=rmcap:10 H263-1998/90000
     a=rmcap:11 H264/90000
     a=tcap:1 RTP/AVP
     a=lcfg:3 mt=message t=2 m=20
     a=tcap:2 TCP/MSRP
     a=omcap:20 *

  The first "lcfg" attribute line ("lcfg:2") announces support for
  H.263 and H.264 video (H.263 preferred) for future negotiation.  The
  second "lcfg" attribute line ("lcfg:3") announces support for MSRP
  for future negotiation.  The "m=" line and the "rtpmap" attribute
  offer an audio stream and provide the lowest precedence configuration
  (PCMU without any DTMF encoding).  The rmcap lines define the RTP-
  based media format capabilities (PCMU, G729, telephone-event,
  H263-1998, and H264) and the omcap line defines the non-RTP-based
  media format capability (wildcard).  The "mfcap" attribute provides
  the format parameters for telephone-event, specifying the 12
  commercial DTMF 'digits'.  The "pcfg" attribute line defines the
  most-preferred media configuration as PCMU plus DTMF events and the
  next-most-preferred configuration as G.729B plus DTMF events.

  If the answerer is able to support all the potential configurations,
  and also support H.263 video (but not H.264), it would reply with an
  answer like the following:








Gilman, et al.               Standards Track                   [Page 48]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


     v=0
     o=- 24351 621814 IN IP4 192.0.2.2
     s=
     c=IN IP4 192.0.2.2
     t=0 0
     a=csup:med-v0
     m=audio 54322 RTP/AVP 0 100
     a=rtpmap:0 PCMU/8000
     a=rtpmap:100 telephone-event/8000
     a=fmtp:100 0-11
     a=acfg:1 m=1,3 pt=1:0,3:100
     a=pcfg:1 m=2,3 pt=2:18,3:100
     a=lcfg:2 mt=video t=1 m=10

  The "lcfg" attribute line announces the capability to support H.263
  video at a later time.  The media line and subsequent "rtpmap" and
  "fmtp" attribute lines present the selected configuration for the
  media stream.  The "acfg" attribute line identifies the potential
  configuration from which it was taken, and the "pcfg" attribute line
  announces the potential capability to support G.729 with DTMF events
  as well.  If, at some later time, congestion becomes a problem in the
  network, either party may, with expectation of success, offer a
  reconfiguration of the media stream to use G.729 in order to reduce
  packet sizes.

5.  IANA Considerations

5.1.  New SDP Attributes

  IANA has registered the following new SDP attributes:

     Attribute name: rmcap
     Long form name: RTP-based media format capability
     Type of attribute: session-level and media-level
     Subject to charset: no
     Purpose: associate RTP-based media capability number(s) with
     media subtype and encoding parameters
     Appropriate Values: see Section 3.3.1
     Contact name: Flemming Andreasen, [email protected]

     Attribute name: omcap
     Long form name: non-RTP-based media format capability
     Type of attribute: session-level and media-level
     Subject to charset: no
     Purpose: associate non-RTP-based media capability number(s) with
     media subtype and encoding parameters
     Appropriate Values: see Section 3.3.1
     Contact name: Flemming Andreasen, [email protected]



Gilman, et al.               Standards Track                   [Page 49]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


     Attribute name: mfcap
     Long form name: media format parameter capability
     Type of attribute: session-level and media-level
     Subject to charset: no
     Purpose: associate media format attributes and
     parameters with media format capabilities
     Appropriate Values: see Section 3.3.2
     Contact name: Flemming Andreasen, [email protected]

     Attribute name: mscap
     Long form name: media-specific capability
     Type of attribute: session-level and media-level
     Subject to charset: no
     Purpose: associate media-specific attributes and
     parameters with media capabilities
     Appropriate Values: see Section 3.3.3
     Contact name: Flemming Andreasen, [email protected]

     Attribute name: lcfg
     Long form name: latent configuration
     Type of attribute: media-level
     Subject to charset: no
     Purpose: to announce supportable media streams
     without offering them for immediate use.
     Appropriate Values: see Section 3.3.5
     Contact name: Flemming Andreasen, [email protected]

     Attribute name: sescap
     Long form name: session capability
     Type of attribute: session-level
     Subject to charset: no
     Purpose: to specify and prioritize acceptable
     combinations of media stream configurations.
     Appropriate Values: see Section 3.3.8
     Contact name: Flemming Andreasen, [email protected]

5.2.  New SDP Capability Negotiation Option Tag

  IANA has added the new option tag "med-v0", defined in this document,
  to the "SDP Capability Negotiation Option Capability Tags" registry
  created for RFC 5939 [RFC5939].

5.3.  SDP Capability Negotiation Configuration Parameters Registry

  IANA has changed the "SDP Capability Negotiation Potential
  Configuration Parameters" registry, currently registered and defined
  by RFC 5939 [RFC5939], as follows:




Gilman, et al.               Standards Track                   [Page 50]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  The name of the registry should be "SDP Capability Negotiation
  Configuration Parameters Registry" and it should contain a table with
  the following column headings:

  o  Encoding Name: The syntactical value used for the capability
     negotiation configuration parameter, as defined in RFC 5939
     [RFC5939], Section 3.5.

  o  Descriptive Name: The name commonly used to refer to the
     capability negotiation configuration parameter.

  o  Potential Configuration Definition: A reference to the RFC that
     defines the configuration parameter in the context of a potential
     configuration attribute.  If the configuration parameter is not
     defined for potential configurations, the string "N/A" (Not
     Applicable) MUST be present instead.

  o  Actual Configuration Definition: A reference to the RFC that
     defines the configuration parameter in the context of an actual
     configuration attribute.  If the configuration parameter is not
     defined for actual configurations, the string "N/A" (Not
     Applicable) MUST be present instead.

  o  Latent Configuration Definition: A reference to the RFC that
     defines the configuration parameter in the context of a latent
     configuration attribute.  If the configuration parameter is not
     defined for latent configurations, the string "N/A" (Not
     Applicable) MUST be present instead.

  An IANA SDP Capability Negotiation Configuration registration MUST be
  documented in an RFC in accordance with the IETF Review policy
  [RFC5226].  Furthermore:

  o  The RFC MUST define the syntax and semantics of each new potential
     configuration parameter.

  o  The syntax MUST adhere to the syntax provided for extension
     configuration lists in RFC 5939 [RFC5939], Section 3.5.1, and the
     semantics MUST adhere to the semantics provided for extension
     configuration lists in RFC 5939 [RFC5939], Sections 3.5.1 and
     3.5.2.

  o  Configuration parameters that apply to latent configurations MUST
     furthermore adhere to the syntax provided in Section 3.3.5 and the
     semantics defined overall in this document.

  o  Associated with each registration MUST be the encoding name for
     the parameter as well as a short descriptive name for it.



Gilman, et al.               Standards Track                   [Page 51]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  o  Each registration MUST specify if it applies to

     *  Potential configurations

     *  Actual configurations

     *  Latent configurations

5.4.  SDP Capability Negotiation Configuration Parameter Registrations

  IANA has registered the following capability negotiation
  configuration parameters:

     Encoding Name: a
     Descriptive Name: Attribute Configuration
     Potential Configuration Definition: [RFC5939]
     Actual Configuration Definition: [RFC5939]
     Latent Configuration Definition: [RFC6871]

     Encoding Name: t
     Descriptive Name: Transport Protocol Configuration
     Potential Configuration Definition: [RFC5939]
     Actual Configuration Definition: [RFC5939]
     Latent Configuration Definition: [RFC6871]

     Encoding Name: m
     Descriptive Name: Media Configuration
     Potential Configuration Definition: [RFC6871]
     Actual Configuration Definition: [RFC6871]
     Latent Configuration Definition: [RFC6871]

     Encoding Name: pt
     Descriptive Name: Payload Type Number Mapping
     Potential Configuration Definition: [RFC6871]
     Actual Configuration Definition: [RFC6871]
     Latent Configuration Definition: [RFC6871]

     Encoding Name: mt
     Descriptive Name: Media Type
     Potential Configuration Definition: N/A
     Actual Configuration Definition: N/A
     Latent Configuration Definition: [RFC6871]

6.  Security Considerations

  The security considerations of RFC 5939 [RFC5939] apply for this
  document.




Gilman, et al.               Standards Track                   [Page 52]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  In RFC 5939 [RFC5939], it was noted that negotiation of transport
  protocols (e.g., secure and non-secure) and negotiation of keying
  methods and material are potential security issues that warrant
  integrity protection to remedy.  Latent configuration support
  provides hints to the other side about capabilities supported for
  further offer/answer exchanges, including transport protocols and
  attribute capabilities, e.g., for keying methods.  If an attacker can
  remove or alter latent configuration information to suggest that only
  non-secure or less-secure alternatives are supported, then he may be
  able to force negotiation of a less secure session than would
  otherwise have occurred.  While the specific attack, as described
  here, differs from those described in RFC 5939 [RFC5939], the
  considerations and mitigation strategies are similar to those
  described in RFC 5939 [RFC5939].

  Another variation on the above attack involves the session capability
  ("sescap") attribute defined in this document.  The "sescap" enables
  a preference order to be specified for all the potential
  configurations, and that preference will take precedence over any
  preference indication provided in individual potential configuration
  attributes.  Consequently, an attacker that can insert or modify a
  "sescap" attribute may be able to force negotiation of an insecure or
  less secure alternative than would otherwise have occurred.  Again,
  the considerations and mitigation strategies are similar to those
  described in RFC 5939 [RFC5939].

  The addition of negotiable media formats and their associated
  parameters, defined in this specification can cause problems for
  middleboxes that attempt to control bandwidth utilization, media
  flows, and/or processing resource consumption as part of network
  policy, but that do not understand the media capability negotiation
  feature.  As for the initial SDP capability negotiation work
  [RFC5939], the SDP answer is formulated in such a way that it always
  carries the selected media encoding for every media stream selected.
  Pending an understanding of capabilities negotiation, the middlebox
  should examine the answer SDP to obtain the best picture of the media
  streams being established.  As always, middleboxes can best do their
  job if they fully understand media capabilities negotiation.

7.  Acknowledgements

  This document is heavily influenced by the discussions and work done
  by the SDP Capability Negotiation design team.  The following people
  in particular provided useful comments and suggestions to either the
  document itself or the overall direction of the solution defined
  herein: Cullen Jennings, Matt Lepinski, Joerg Ott, Colin Perkins, and
  Thomas Stach.




Gilman, et al.               Standards Track                   [Page 53]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  We thank Ingemar Johansson and Magnus Westerlund for examples that
  stimulated this work, and for critical reading of the document.  We
  also thank Cullen Jennings, Christer Holmberg, and Miguel Garcia for
  their review of the document.

8.  References

8.1.  Normative References

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

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

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

  [RFC5226]  Narten, T. and H. Alvestrand, "Guidelines for Writing an
             IANA Considerations Section in RFCs", BCP 26, RFC 5226,
             May 2008.

  [RFC5234]  Crocker, D. and P. Overell, "Augmented BNF for Syntax
             Specifications: ABNF", STD 68, RFC 5234, January 2008.

  [RFC5939]  Andreasen, F., "Session Description Protocol (SDP)
             Capability Negotiation", RFC 5939, September 2010.

8.2.  Informative References

  [RFC2198]  Perkins, C., Kouvelas, I., Hodson, O., Hardman, V.,
             Handley, M., Bolot, J., Vega-Garcia, A., and S. Fosse-
             Parisis, "RTP Payload for Redundant Audio Data", RFC 2198,
             September 1997.

  [RFC4568]  Andreasen, F., Baugher, M., and D. Wing, "Session
             Description Protocol (SDP) Security Descriptions for Media
             Streams", RFC 4568, July 2006.

  [RFC4585]  Ott, J., Wenger, S., Sato, N., Burmeister, C., and J. Rey,
             "Extended RTP Profile for Real-time Transport Control
             Protocol (RTCP)-Based Feedback (RTP/AVPF)", RFC 4585, July
             2006.

  [RFC4733]  Schulzrinne, H. and T. Taylor, "RTP Payload for DTMF
             Digits, Telephony Tones, and Telephony Signals", RFC 4733,
             December 2006.



Gilman, et al.               Standards Track                   [Page 54]

RFC 6871           SDP Media Capabilities Negotiation      February 2013


  [RFC4867]  Sjoberg, J., Westerlund, M., Lakaniemi, A., and Q. Xie,
             "RTP Payload Format and File Storage Format for the
             Adaptive Multi-Rate (AMR) and Adaptive Multi-Rate Wideband
             (AMR-WB) Audio Codecs", RFC 4867, April 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.

Authors' Addresses

  Robert R Gilman
  Independent
  3243 W. 11th Ave. Dr.
  Broomfield, CO 80020
  USA

  EMail: [email protected]


  Roni Even
  Huawei Technologies
  14 David Hamelech
  Tel Aviv  64953
  Israel

  EMail: [email protected]


  Flemming Andreasen
  Cisco Systems
  Iselin, NJ
  USA

  EMail: [email protected]
















Gilman, et al.               Standards Track                   [Page 55]