RFC 9071




Internet Engineering Task Force (IETF)                      G. Hellström
Request for Comments: 9071                                      GHAccess
Updates: 4103                                                  July 2021
Category: Standards Track
ISSN: 2070-1721


           RTP-Mixer Formatting of Multiparty Real-Time Text

Abstract



   This document provides enhancements of real-time text (as specified
   in RFC 4103) suitable for mixing in a centralized conference model,
   enabling source identification and rapidly interleaved transmission
   of text from different sources.  The intended use is for real-time
   text mixers and participant endpoints capable of providing an
   efficient presentation or other treatment of a multiparty real-time
   text session.  The specified mechanism builds on the standard use of
   the Contributing Source (CSRC) list in the Real-time Transport
   Protocol (RTP) packet for source identification.  The method makes
   use of the same "text/t140" and "text/red" formats as for two-party
   sessions.

   Solutions using multiple RTP streams in the same RTP session are
   briefly mentioned, as they could have some benefits over the RTP-
   mixer model.  The RTP-mixer model was selected to be used for the
   fully specified solution in this document because it can be applied
   to a wide range of existing RTP implementations.

   A capability exchange is specified so that it can be verified that a
   mixer and a participant can handle the multiparty-coded real-time
   text stream using the RTP-mixer method.  The capability is indicated
   by the use of a Session Description Protocol (SDP) (RFC 8866) media
   attribute, "rtt-mixer".

   This document updates RFC 4103 ("RTP Payload for Text Conversation").

   A specification for how a mixer can format text for the case when the
   endpoint is not multiparty aware is also provided.

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 7841.

   Information about the current status of this document, any errata,
   and how to provide feedback on it may be obtained at
   https://www.rfc-editor.org/info/rfc9071.

Copyright Notice



   Copyright (c) 2021 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
   (https://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.

Table of Contents



   1.  Introduction
     1.1.  Terminology
     1.2.  Main Method, Fallback Method, and Considered Alternatives
     1.3.  Intended Application
   2.  Overview of the Two Specified Solutions and Selection of Method
     2.1.  The RTP-Mixer-Based Solution for Multiparty-Aware Endpoints
     2.2.  Mixing for Multiparty-Unaware Endpoints
     2.3.  Offer/Answer Considerations
     2.4.  Actions Depending on Capability Negotiation Result
   3.  Details for the RTP-Mixer-Based Mixing Method for
           Multiparty-Aware Endpoints
     3.1.  Use of Fields in the RTP Packets
     3.2.  Initial Transmission of a BOM Character
     3.3.  Keep-Alive
     3.4.  Transmission Interval
     3.5.  Only One Source per Packet
     3.6.  Do Not Send Received Text to the Originating Source
     3.7.  Clean Incoming Text
     3.8.  Principles of Redundant Transmission
     3.9.  Text Placement in Packets
     3.10. Empty T140blocks
     3.11. Creation of the Redundancy
     3.12. Timer Offset Fields
     3.13. Other RTP Header Fields
     3.14. Pause in Transmission
     3.15. RTCP Considerations
     3.16. Reception of Multiparty Contents
     3.17. Performance Considerations
     3.18. Security for Session Control and Media
     3.19. SDP Offer/Answer Examples
     3.20. Packet Sequence Example from Interleaved Transmission
     3.21. Maximum Character Rate "cps" Setting
   4.  Presentation-Level Considerations
     4.1.  Presentation by Multiparty-Aware Endpoints
     4.2.  Multiparty Mixing for Multiparty-Unaware Endpoints
   5.  Relationship to Conference Control
     5.1.  Use with SIP Centralized Conferencing Framework
     5.2.  Conference Control
   6.  Gateway Considerations
     6.1.  Gateway Considerations with Textphones
     6.2.  Gateway Considerations with WebRTC
   7.  Updates to RFC 4103
   8.  Congestion Considerations
   9.  IANA Considerations
     9.1.  Registration of the "rtt-mixer" SDP Media Attribute
   10. Security Considerations
   11. References
     11.1.  Normative References
     11.2.  Informative References
   Acknowledgements

   Author's Address



1.  Introduction



   "RTP Payload for Text Conversation" [RFC4103] specifies the use of
   the Real-time Transport Protocol (RTP) [RFC3550] for transmission of
   real-time text (often called RTT) and the "text/t140" format.  It
   also specifies a redundancy format, "text/red", for increased
   robustness.  The "text/red" format is registered in [RFC4102].

   Real-time text is usually provided together with audio and sometimes
   with video in conversational sessions.

   A requirement related to multiparty sessions from the presentation-
   level standard T.140 [T140] for real-time text is as follows:

   |  The display of text from the members of the conversation should be
   |  arranged so that the text from each participant is clearly
   |  readable, and its source and the relative timing of entered text
   |  is visualized in the display.

   Another requirement is that the mixing procedure must not introduce
   delays in the text streams that could be perceived as disruptive to
   the real-time experience of the receiving users.

   The use of real-time text is increasing, and specifically, use in
   emergency calls is increasing.  Emergency call use requires
   multiparty mixing, because it is common that one agent needs to
   transfer the call to another specialized agent but is obliged to stay
   on the call to at least verify that the transfer was successful.
   Mixer implementations for RFC 4103 ("RTP Payload for Text
   Conversation") can use traditional RTP functions (RFC 3550) for
   mixing and source identification, but the performance of the mixer
   when giving turns for the different sources to transmit is limited
   when using the default transmission characteristics with redundancy.

   The redundancy scheme described in [RFC4103] enables efficient
   transmission of earlier transmitted redundant text in packets
   together with new text.  However, the redundancy header format has no
   source indicators for the redundant transmissions.  The redundant
   parts in a packet must therefore be from the same source as the new
   text.  The recommended transmission is one new and two redundant
   generations of text (T140blocks) in each packet, and the recommended
   transmission interval for two-party use is 300 ms.

   Real-time text mixers for multiparty sessions need to include the
   source with each transmitted group of text from a conference
   participant so that the text can be transmitted interleaved with text
   groups from different sources at the rate at which they are created.
   This enables the text groups to be presented by endpoints in a
   suitable grouping with other text from the same source.

   The presentation can then be arranged so that text from different
   sources can be presented in real time and easily read.  At the same
   time, it is possible for a reading user to perceive approximately
   when the text was created in real time by the different parties.  The
   transmission and mixing are intended to be done in a general way, so
   that presentation can be arranged in a layout decided upon by the
   receiving endpoint.

   Existing implementations of RFC 4103 in endpoints that do not
   implement the updates specified in this document cannot be expected
   to properly present real-time text mixed for multiparty-aware
   endpoints.

   A negotiation mechanism is therefore needed to verify if the parties
   (1) are able to handle a common method for multiparty transmissions
   and (2) can agree on using that method.

   A fallback mixing procedure is also needed for cases when the
   negotiation result indicates that a receiving endpoint is not capable
   of handling the mixed format.  Multiparty-unaware endpoints would
   possibly otherwise present all received multiparty mixed text as if
   it came from the same source regardless of any accompanying source
   indication coded in fields in the packet.  Or, they may have other
   undesirable ways of acting on the multiparty content.  The fallback
   method is called the mixing procedure for multiparty-unaware
   endpoints.  The fallback method is naturally not expected to meet all
   performance requirements placed on the mixing procedure for
   multiparty-aware endpoints.

   This document updates [RFC4103] by introducing an attribute for
   declaring support of the RTP-mixer-based multiparty-mixing case and
   rules for source indications and interleaving of text from different
   sources.

1.1.  Terminology



   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and
   "OPTIONAL" in this document are to be interpreted as described in
   BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all
   capitals, as shown here.

   The terms "Source Description" (SDES), "Canonical Name" (CNAME),
   "Name" (NAME), "Synchronization Source" (SSRC), "Contributing Source"
   (CSRC), "CSRC list", "CSRC count" (CC), "RTP Control Protocol"
   (RTCP), and "RTP mixer" are defined in [RFC3550].

   "real-time text" (RTT) is text transmitted instantly as it is typed
   or created.  Recipients can immediately read the message while it is
   being written, without waiting.

   The term "T140block" is defined in [RFC4103] to contain one or more
   T.140 code elements.

   "TTY" stands for a textphone type used in North America.

   Web Real-Time Communication (WebRTC) is specified by the World Wide
   Web Consortium (W3C) and the IETF.  See [RFC8825].

   "DTLS-SRTP" is a Datagram Transport Layer Security (DTLS) extension
   for use with the Secure Real-time Transport Protocol / Secure Real-
   time Transport Control Protocol (SRTP/SRTCP) as specified in
   [RFC5764].

   The term "multiparty aware" describes an endpoint that (1) receives
   real-time text from multiple sources through a common conference
   mixer, (2) is able to present the text in real time, separated by
   source, and (3) presents the text so that a user can get an
   impression of the approximate relative timing of text from different
   parties.

   The term "multiparty unaware" describes an endpoint that cannot
   itself separate text from different sources when the text is received
   through a common conference mixer.

1.2.  Main Method, Fallback Method, and Considered Alternatives



   A number of alternatives were considered when searching for an
   efficient and easily implemented multiparty method for real-time
   text.  This section briefly explains a few of them.

   Multiple RTP streams, one per participant:
      One RTP stream per source would be sent in the same RTP session
      with the "text/red" format.  From some points of view, the use of
      multiple RTP streams, one for each source, sent in the same RTP
      session would be efficient and would use exactly the same packet
      format as [RFC4103] and the same payload type.  A couple of
      relevant scenarios using multiple RTP streams are specified in
      "RTP Topologies" [RFC7667].  One possibility of special interest
      is the Selective Forwarding Middlebox (SFM) topology specified in
      Section 3.7 of [RFC7667], which could enable end-to-end
      encryption.  In contrast to audio and video, real-time text is
      only transmitted when the users actually transmit information.
      Thus, an SFM solution would not need to exclude any party from
      transmission under normal conditions.  In order to allow the mixer
      to convey the packets with the payload preserved and encrypted, an
      SFM solution would need to act on some specific characteristics of
      the "text/red" format.  The redundancy headers are part of the
      payload, so the receiver would need to just assume that the
      payload type number in the redundancy header is for "text/t140".
      The characters per second ("cps") parameter would need to act per
      stream.  The relationship between the SSRC and the source would
      need to be conveyed in some specified way, e.g., in the CSRC.
      Recovery and loss detection would preferably be based on RTP
      sequence number gap detection.  Thus, sequence number gaps in the
      incoming stream to the mixer would need to be reflected in the
      stream to the participant, with no new gaps created by the mixer.
      However, the RTP implementation in both mixers and endpoints needs
      to support multiple streams in the same RTP session in order to
      use this mechanism.  To provide the best opportunities for
      deployment, it should be possible to upgrade existing endpoint
      solutions to be multiparty aware with a reasonable amount of
      effort.  There is currently a lack of support for multi-stream RTP
      in certain implementations.  This fact led to only brief mention
      of this solution in this document as an option for further study.

   RTP-mixer-based method for multiparty-aware endpoints:
      The "text/red" format as defined in RFC 4102 and applied in RFC
      4103 is sent with the RTP-mixer method indicating the source in
      the CSRC field.  The "text/red" format with a "text/t140" payload
      in a single RTP stream can be sent when text is available from the
      call participants instead of at the regular 300 ms intervals.
      Transmission of packets with text from different sources can then
      be done smoothly while simultaneous transmission occurs as long as
      it is not limited by the maximum character rate "cps" value.  With
      ten participants sending text simultaneously, the switching and
      transmission performance is good.  With more simultaneously
      sending participants and with receivers at default capacity, there
      will be a noticeable jerkiness and delay in text presentation.
      The more participants who send text simultaneously, the more
      jerkiness will occur.  Two seconds of jerkiness will be noticeable
      and slightly unpleasant, but it corresponds in time to what typing
      humans often cause by hesitating or changing position while
      typing.  A benefit of this method is that no new packet format
      needs to be introduced and implemented.  Since simultaneous typing
      by more than two parties is expected to be very rare -- as
      described in Section 1.3 -- this method can be used successfully
      with good performance.  Recovery of text in the case of packet
      loss is based on analysis of timestamps of received redundancy
      versus earlier received text.  Negotiation is based on a new SDP
      media attribute, "rtt-mixer".  This method was selected to be the
      main method specified in this document.

   Multiple sources per packet:
      A new "text" media subtype would be specified with up to 15
      sources in each packet.  The mechanism would make use of the RTP-
      mixer model specified in RTP [RFC3550].  The sources would be
      indicated in strict order in the CSRC list of the RTP packets.
      The CSRC list can have up to 15 members.  Therefore, text from up
      to 15 sources can be included in each packet.  Packets are
      normally sent at 300 ms intervals.  The mean delay would be 150
      ms.  A new redundancy packet format would be specified.  This
      method would result in good performance but would require
      standardization and implementation of new releases in the target
      technologies; these would take more time than desirable to
      complete.  It was therefore not selected to be included in this
      document.

   Mixing for multiparty-unaware endpoints:
      The presentation of text from multiple parties is prepared by the
      mixer in one single stream.  It is desirable to have a method that
      does not require any modifications in existing user devices
      implementing RFC 4103 for real-time text without explicit support
      of multiparty sessions.  This is made possible by having the mixer
      insert a new line and a text-formatted source label before each
      switch of text source in the stream.  Switching the source can
      only be done in places in the text where it does not disturb the
      perception of the contents.  Text from only one source at a time
      can be presented in real time.  The delay will therefore vary.  In
      calls where parties take turns properly by ending their entries
      with a new line, the limitations will have limited influence on
      the user experience.  When only two parties send text, these two
      will see the text in real time with no delay.  Although this
      method also has other limitations, it is included in this document
      as a fallback method.

   Real-time text transport in WebRTC:
      [RFC8865] specifies how the WebRTC data channel can be used to
      transport real-time text.  That specification contains a section
      briefly describing its use in multiparty sessions.  The focus of
      this document is RTP transport.  Therefore, even if the WebRTC
      transport provides good multiparty performance, it is only
      mentioned in this document in relation to providing gateways with
      multiparty capabilities between RTP and WebRTC technologies.

1.3.  Intended Application



   The method for multiparty real-time text specified in this document
   is primarily intended for use in transmissions between mixers and
   endpoints in centralized mixing configurations.  It is also
   applicable between mixers.  An often-mentioned application is for
   emergency service calls with real-time text and voice, where a call
   taker wants to make an attended handover of a call to another agent
   and stay on the call to observe the session.  Multimedia conference
   sessions with support for participants to contribute with text is
   another example.  Conferences with central support for speech-to-text
   conversion represent yet another example.

   In all these applications, normally only one participant at a time
   will send long text comments.  In some cases, one other participant
   will occasionally contribute with a longer comment simultaneously.
   That may also happen in some rare cases when text is translated to
   text in another language in a conference.  Apart from these cases,
   other participants are only expected to contribute with very brief
   comments while others are sending text.

   Users expect the text they send to be presented in real time in a
   readable way to the other participants even if they send
   simultaneously with other users and even when they make brief edit
   operations of their text by backspacing and correcting their text.

   Text is supposed to be human generated, by some means of text input,
   such as typing on a keyboard or using speech-to-text technology.
   Occasional small cut-and-paste operations may appear even if that is
   not the initial purpose of real-time text.

   The real-time characteristics of real-time text are essential for the
   participants to be able to contribute to a conversation.  If the text
   is delayed too much between the typing of a character and its
   presentation, then, in some conference situations, the opportunity to
   comment will be gone and someone else will grab the turn.  A delay of
   more than one second in such situations is an obstacle to good
   conversation.

2.  Overview of the Two Specified Solutions and Selection of Method



   This section contains a brief introduction of the two methods
   specified in this document.

2.1.  The RTP-Mixer-Based Solution for Multiparty-Aware Endpoints



   This method specifies the negotiated use of the formats described in
   RFC 4103, for multiparty transmissions in a single RTP stream.  The
   main purpose of this document is to specify a method for true
   multiparty real-time text mixing for multiparty-aware endpoints that
   can be widely deployed.  The RTP-mixer-based method makes use of the
   current format for real-time text as provided in [RFC4103].  This
   method updates RFC 4103 by clarifying one way to use it in the
   multiparty situation.  That is done by completing a negotiation for
   this kind of multiparty capability and by interleaving packets from
   different sources.  The source is indicated in the CSRC element in
   the RTP packets.  Specific considerations are made regarding the
   ability to recover text after packet loss.

   The detailed procedures for the RTP-mixer-based multiparty-aware case
   are specified in Section 3.

   Please refer to [RFC4103] when reading this document.

2.2.  Mixing for Multiparty-Unaware Endpoints



   This document also specifies a method to be used in cases when the
   endpoint participating in a multiparty call does not itself implement
   any solution or does not implement the same solution as the mixer.
   This method requires the mixer to insert text dividers and readable
   labels and only send text from one source at a time until a suitable
   point appears for changing the source.  This solution is a fallback
   method with functional limitations.  It operates at the presentation
   level.

   A mixer SHOULD by default format and transmit text to a call
   participant so that the text is suitable for presentation on a
   multiparty-unaware endpoint that has not negotiated any method for
   true multiparty real-time text handling but has negotiated a "text/
   red" or "text/t140" format in a session.  This SHOULD be done if
   nothing else is specified for the application, in order to maintain
   interoperability.  Section 4.2 specifies how this mixing is done.

2.3.  Offer/Answer Considerations



   "RTP Payload for Text Conversation" [RFC4103] specifies the use of
   RTP [RFC3550] and a redundancy format ("text/red", as defined in
   [RFC4102]) for increased robustness of real-time text transmission.
   This document updates [RFC4103] by introducing a capability
   negotiation for handling multiparty real-time text, a way to indicate
   the source of transmitted text, and rules for efficient timing of the
   transmissions interleaved from different sources.

   The capability negotiation for the RTP-mixer-based multiparty method
   is based on the use of the SDP media attribute "rtt-mixer".

   The syntax is as follows:

      a=rtt-mixer

   If in the future any other method for RTP-based multiparty real-time
   text is specified by additional work, it is assumed that it will be
   recognized by some specific SDP feature exchange.

2.3.1.  Initial Offer



   A party that intends to set up a session and is willing to use the
   RTP-mixer-based method provided in this specification for sending,
   receiving, or both sending and receiving real-time text SHALL include
   the "rtt-mixer" SDP attribute in the corresponding "text" media
   section in the initial offer.

   The party MAY indicate its capability regarding both the RTP-mixer-
   based method provided in this specification and other methods.

   When the offerer has sent the offer, which includes the "rtt-mixer"
   attribute, it MUST be prepared to receive and handle real-time text
   formatted according to both the method for multiparty-aware parties
   specified in Section 3 and two-party formatted real-time text.

2.3.2.  Answering the Offer



   A party that receives an offer containing the "rtt-mixer" SDP
   attribute and is willing to use the RTP-mixer-based method provided
   in this specification for sending, receiving, or both sending and
   receiving real-time text SHALL include the "rtt-mixer" SDP attribute
   in the corresponding "text" media section in the answer.

   If the offer did not contain the "rtt-mixer" attribute, the answer
   MUST NOT contain the "rtt-mixer" attribute.

   Even when the "rtt-mixer" attribute is successfully negotiated, the
   parties MAY send and receive two-party coded real-time text.

   An answer MUST NOT include acceptance of more than one method for
   multiparty real-time text in the same RTP session.

   When the answer, which includes acceptance, is transmitted, the
   answerer MUST be prepared to act on received text in the negotiated
   session according to the method for multiparty-aware parties
   specified in Section 3.  Reception of text for a two-party session
   SHALL also be supported.

2.3.3.  Offerer Processing the Answer



   When the answer is processed by the offerer, the offerer MUST follow
   the requirements listed in Section 2.4.

2.3.4.  Modifying a Session



   A session MAY be modified at any time by any party offering a
   modified SDP with or without the "rtt-mixer" SDP attribute expressing
   a desired change in the support of multiparty real-time text.

   If the modified offer adds the indication of support for multiparty
   real-time text by including the "rtt-mixer" SDP attribute, the
   procedures specified in the previous subsections SHALL be applied.

   If the modified offer deletes the indication of support for
   multiparty real-time text by excluding the "rtt-mixer" SDP attribute,
   the answer MUST NOT contain the "rtt-mixer" attribute.  After
   processing this SDP exchange, the parties MUST NOT send real-time
   text formatted for multiparty-aware parties according to this
   specification.

2.4.  Actions Depending on Capability Negotiation Result



   A transmitting party SHALL send text according to the RTP-mixer-based
   multiparty method only when the negotiation for that method was
   successful and when it conveys text for another source.  In all other
   cases, the packets SHALL be populated and interpreted as for a two-
   party session.

   A party that has negotiated the "rtt-mixer" SDP media attribute and
   acts as an RTP mixer sending multiparty text MUST (1) populate the
   CSRC list and (2) format the packets according to Section 3.

   A party that has negotiated the "rtt-mixer" SDP media attribute MUST
   interpret the contents of the CC field, the CSRC list, and the
   packets according to Section 3 in received RTP packets in the
   corresponding RTP stream.

   A party that has not successfully completed the negotiation of the
   "rtt-mixer" SDP media attribute MUST NOT transmit packets interleaved
   from different sources in the same RTP stream, as specified in
   Section 3.  If the party is a mixer and did declare the "rtt-mixer"
   SDP media attribute, it SHOULD perform the procedure for multiparty-
   unaware endpoints.  If the party is not a mixer, it SHOULD transmit
   as in a two-party session according to [RFC4103].

3.  Details for the RTP-Mixer-Based Mixing Method for Multiparty-Aware
    Endpoints



3.1.  Use of Fields in the RTP Packets



   The CC field SHALL show the number of members in the CSRC list, which
   SHALL be one (1) in transmissions from a mixer when conveying text
   from other sources in a multiparty session, and otherwise 0.

   When text is conveyed by a mixer during a multiparty session, a CSRC
   list SHALL be included in the packet.  The single member in the CSRC
   list SHALL contain the SSRC of the source of the T140blocks in the
   packet.

   When redundancy is used, the RECOMMENDED level of redundancy is to
   use one primary and two redundant generations of T140blocks.  In some
   cases, a primary or redundant T140block is empty but is still
   represented by a member in the redundancy header.

   In other respects, the contents of the RTP packets will be as
   specified in [RFC4103].

3.2.  Initial Transmission of a BOM Character



   As soon as a participant is known to participate in a session with
   another entity and is available for text reception, a Unicode byte
   order mark (BOM) character SHALL be sent to it by the other entity
   according to the procedures in this section.  This is useful in many
   configurations for opening ports and firewalls and for setting up the
   connection between the application and the network.  If the
   transmitter is a mixer, then the source of this character SHALL be
   indicated to be the mixer itself.

   Note that the BOM character SHALL be transmitted with the same
   redundancy procedures as any other text.

3.3.  Keep-Alive



   After that, the transmitter SHALL send keep-alive traffic to the
   receiver(s) at regular intervals when no other traffic has occurred
   during that interval, if that is decided upon for the actual
   connection.  It is RECOMMENDED to use the keep-alive solution
   provided in [RFC6263].  The consent check [RFC7675] is a possible
   alternative if it is used anyway for other reasons.

3.4.  Transmission Interval



   A "text/red" or "text/t140" transmitter in a mixer SHALL send packets
   distributed over time as long as there is something (new or redundant
   T140blocks) to transmit.  The maximum transmission interval between
   text transmissions from the same source SHALL then be 330 ms, when no
   other limitations cause a longer interval to be temporarily used.  It
   is RECOMMENDED to send the next packet to a receiver as soon as new
   text to that receiver is available, as long as the mean character
   rate of new text to the receiver calculated over the last 10 one-
   second intervals does not exceed the "cps" value of the receiver.
   The intention is to keep the latency low and network load limited
   while keeping good protection against text loss in bursty packet loss
   conditions.  The main purpose of the 330 ms interval is for the
   timing of redundant transmissions, when no new text from the same
   source is available.

   The value of 330 ms is used, because many sources of text will
   transmit new text at 300 ms intervals during periods of continuous
   user typing, and then reception in the mixer of such new text will
   cause a combined transmission of the new text and the unsent
   redundancy from the previous transmission.  Only when the user stops
   typing will the 330 ms interval be applied to send the redundancy.

   If the characters per second ("cps") value is reached, a longer
   transmission interval SHALL be applied for text from all sources as
   specified in [RFC4103] and only as much of the text queued for
   transmission SHALL be sent at the end of each transmission interval
   as can be allowed without exceeding the "cps" value.  Division of
   text for partial transmission MUST then be made at T140block borders.
   When the transmission rate falls below the "cps" value again, the
   transmission intervals SHALL be reset to 330 ms and transmission of
   new text SHALL again be made as soon as new text is available.

      |  NOTE: Extending the transmission intervals during periods of
      |  high load does not change the number of characters to be
      |  conveyed.  It just evens out the load over time and reduces the
      |  number of packets per second.  With human-created
      |  conversational text, the sending user will eventually take a
      |  pause, letting transmission catch up.

   See also Section 8.

   For a transmitter not acting as a mixer, the transmission interval
   principles provided in [RFC4103] apply, and the normal transmission
   interval SHALL be 300 ms.

3.5.  Only One Source per Packet



   New text and redundant copies of earlier text from one source SHALL
   be transmitted in the same packet if available for transmission at
   the same time.  Text from different sources MUST NOT be transmitted
   in the same packet.

3.6.  Do Not Send Received Text to the Originating Source



   Text received by a mixer from a participant SHOULD NOT be included in
   transmissions from the mixer to that participant, because for text
   that is produced locally, the normal behavior of the endpoint is to
   present such text directly when it is produced.

3.7.  Clean Incoming Text



   A mixer SHALL handle reception, recovery from packet loss, deletion
   of superfluous redundancy, marking of possible text loss, and
   deletion of BOM characters from each participant before queueing
   received text for transmission to receiving participants as specified
   in [RFC4103] for single-party sources and Section 3.16 for multiparty
   sources (chained mixers).

3.8.  Principles of Redundant Transmission



   A transmitting party using redundancy SHALL send redundant
   repetitions of T140blocks already transmitted in earlier packets.

   The number of redundant generations of T140blocks to include in
   transmitted packets SHALL be deduced from the SDP negotiation.  It
   SHALL be set to the minimum of the number declared by the two parties
   negotiating a connection.  It is RECOMMENDED to declare and transmit
   one original and two redundant generations of the T140blocks, because
   this provides good protection against text loss in the case of packet
   loss and also provides low overhead.

3.9.  Text Placement in Packets



   The mixer SHALL compose and transmit an RTP packet to a receiver when
   one or more of the following conditions have occurred:

   *  The transmission interval is the normal 330 ms (no matter whether
      the transmission interval has passed or not), and there is newly
      received unsent text available for transmission to that receiver.

   *  The current transmission interval has passed and is longer than
      the normal 330 ms, and there is newly received unsent text
      available for transmission to that receiver.

   *  The current transmission interval (normally 330 ms) has passed
      since already-transmitted text was queued for transmission as
      redundant text.

   The principles provided in [RFC4103] apply for populating the header,
   the redundancy header, and the data in the packet with specific
   information, as detailed here and in the following sections.

   At the time of transmission, the mixer SHALL populate the RTP packet
   with all T140blocks queued for transmission originating from the
   source selected for transmission as long as this is not in conflict
   with the allowed number of characters per second ("cps") or the
   maximum packet size.  In this way, the latency of the latest received
   text is kept low even in moments of simultaneous transmission from
   many sources.

   Redundant text SHALL also be included, and the assessment of how much
   new text can be included within the maximum packet size MUST take
   into account that the redundancy has priority to be transmitted in
   its entirety.  See Section 3.4.

   The SSRC of the source SHALL be placed as the only member in the CSRC
   list.

      |  Note: The CSRC list in an RTP packet only includes the
      |  participant whose text is included in text blocks.  It is not
      |  the same as the total list of participants in a conference.
      |  With audio and video media, the CSRC list would often contain
      |  all participants who are not muted, whereas text participants
      |  that don't type are completely silent and thus are not
      |  represented in RTP packet CSRC lists.

3.10.  Empty T140blocks



   If no unsent T140blocks were available for a source at the time of
   populating a packet but already-transmitted T140blocks are available
   that have not yet been sent the full intended number of redundant
   transmissions, then the primary area in the packet is composed of an
   empty T140block and included (without taking up any length) in the
   packet for transmission.  The corresponding SSRC SHALL be placed as
   usual in its place in the CSRC list.

   The first packet in the session, the first after a source switch, and
   the first after a pause SHALL be populated with the available
   T140blocks for the source selected to be sent as the primary, and
   empty T140blocks for the agreed-upon number of redundancy
   generations.

3.11.  Creation of the Redundancy



   The primary T140block from a source in the latest transmitted packet
   is saved for populating the first redundant T140block for that source
   in the next transmission of text from that source.  The first
   redundant T140block for that source from the latest transmission is
   saved for populating the second redundant T140block in the next
   transmission of text from that source.

   Usually, this is the level of redundancy used.  If a higher level of
   redundancy is negotiated, then the procedure SHALL be continued until
   all available redundant levels of T140blocks are placed in the
   packet.  If a receiver has negotiated a lower number of "text/red"
   generations, then that level SHALL be the maximum used by the
   transmitter.

   The T140blocks saved for transmission as redundant data are assigned
   a planned transmission time of 330 ms after the current time but
   SHOULD be transmitted earlier if new text for the same source gets
   selected for transmission before that time.

3.12.  Timer Offset Fields



   The timestamp offset values SHALL be inserted in the redundancy
   header, with the time offset from the RTP timestamp in the packet
   when the corresponding T140block was sent as the primary.

   The timestamp offsets are expressed in the same clock tick units as
   the RTP timestamp.

   The timestamp offset values for empty T140blocks have no relevance
   but SHOULD be assigned realistic values.

3.13.  Other RTP Header Fields



   The number of members in the CSRC list (0 or 1) SHALL be placed in
   the CC header field.  Only mixers place value 1 in the CC field.  A
   value of 0 indicates that the source is the transmitting device
   itself and that the source is indicated by the SSRC field.  This
   value is used by endpoints and also by mixers sending self-sourced
   data.

   The current time SHALL be inserted in the timestamp.

   The SSRC header field SHALL contain the SSRC of the RTP session where
   the packet will be transmitted.

   The M-bit SHALL be handled as specified in [RFC4103].

3.14.  Pause in Transmission



   When there is no new T140block to transmit and no redundant T140block
   that has not been retransmitted the intended number of times from any
   source, the transmission process SHALL be stopped until either new
   T140blocks arrive or a keep-alive method calls for transmission of
   keep-alive packets.

3.15.  RTCP Considerations



   A mixer SHALL send RTCP reports with SDES, CNAME, and NAME
   information about the sources in the multiparty call.  This makes it
   possible for participants to compose a suitable label for text from
   each source.

   Privacy considerations SHALL be taken when composing these fields.
   They contain name and address information that may be considered
   sensitive if the information is transmitted in its entirety, e.g., to
   unauthenticated participants.

3.16.  Reception of Multiparty Contents



   The "text/red" receiver included in an endpoint with presentation
   functions will receive RTP packets in the single stream from the
   mixer and SHALL distribute the T140blocks for presentation in
   presentation areas for each source.  Other receiver roles, such as
   gateways or chained mixers, are also feasible.  Whether the stream
   will only be forwarded or will be distributed based on the different
   sources must be taken into consideration.

3.16.1.  Acting on the Source of the Packet Contents



   If the CC field value of a received packet is 1, it indicates that
   the text is conveyed from a source indicated in the single member in
   the CSRC list, and the receiver MUST act on the source according to
   its role.  If the CC value is 0, the source is indicated in the SSRC
   field.

3.16.2.  Detection and Indication of Possible Text Loss



   The receiver SHALL monitor the RTP sequence numbers of the received
   packets for gaps and for packets received out of order.  If a
   sequence number gap appears and still exists after some defined short
   time for jitter and reordering resolution, the packets in the gap
   SHALL be regarded as lost.

   If it is known that only one source is active in the RTP session,
   then it is likely that a gap equal to or larger than the agreed-upon
   number of redundancy generations (including the primary) causes text
   loss.  In that case, the receiver SHALL create a T140block with a
   marker for possible text loss [T140ad1], associate it with the
   source, and insert it in the reception buffer for that source.

   If it is known that more than one source is active in the RTP
   session, then it is not possible in general to evaluate if text was
   lost when packets were lost.  With two active sources and the
   recommended number of redundancy generations (one original and two
   redundant), it can take a gap of five consecutive lost packets before
   any text may be lost, but text loss can also appear if three non-
   consecutive packets are lost when they contained consecutive data
   from the same source.  A simple method for deciding when there is a
   risk of resulting text loss is to evaluate if three or more packets
   were lost within one second.  If this simple method is used, then a
   T140block SHOULD be created with a marker for possible text loss
   [T140ad1] and associated with the SSRC of the RTP session as a
   general input from the mixer.

   Implementations MAY apply more refined methods for more reliable
   detection of whether text was lost or not.  Any refined method SHOULD
   prefer marking possible loss rather than not marking when it is
   uncertain if there was loss.

3.16.3.  Extracting Text and Handling Recovery



   When applying the following procedures, the effects of possible
   timestamp wraparound and the RTP session possibly changing the SSRC
   MUST be considered.

   When a packet is received in an RTP session using the packetization
   for multiparty-aware endpoints, its T140blocks SHALL be extracted as
   described below.

   The source SHALL be extracted from the CSRC list if available, and
   otherwise from the SSRC.

   If the received packet is the first packet received from the source,
   then all T140blocks in the packet SHALL be retrieved and assigned to
   a receive buffer for that source, beginning with the oldest available
   redundant generation, continuing with the younger redundant
   generations in age order, and finally ending with the primary.

      |  Note: The normal case is that in the first packet, only the
      |  primary data has contents.  The redundant data has contents in
      |  the first received packet from a source only after initial
      |  packet loss.

   If the packet is not the first packet from a source, then if
   redundant data is available, the process SHALL start with the oldest
   generation.  The timestamp of that redundant data SHALL be created by
   subtracting its timestamp offset from the RTP timestamp.  If the
   resulting timestamp is later than the latest retrieved data from the
   same source, then the redundant data SHALL be retrieved and appended
   to the receive buffer.  The process SHALL be continued in the same
   way for all younger generations of redundant data.  After that, the
   timestamp of the packet SHALL be compared with the timestamp of the
   latest retrieved data from the same source and if it is later, then
   the primary data SHALL be retrieved from the packet and appended to
   the receive buffer for the source.

3.16.4.  Delete BOM



   The Unicode BOM character is used as a start indication and is
   sometimes used as a filler or keep-alive by transmission
   implementations.  Any BOM characters SHALL be deleted after
   extraction from received packets.

3.17.  Performance Considerations



   This solution has good performance with low text delays, as long as
   the mean number of characters per second sent during any 10-second
   interval from a number of simultaneously sending participants to a
   receiving participant does not reach the "cps" value.  At higher
   numbers of sent characters per second, a jerkiness is visible in the
   presentation of text.  The solution is therefore suitable for
   emergency service use, relay service use, and small or well-managed
   larger multimedia conferences.  In large unmanaged conferences with a
   high number of participants only, on very rare occasions, situations
   might arise where many participants happen to send text
   simultaneously.  In such circumstances, the result may be
   unpleasantly jerky presentation of text from each sending
   participant.  It should be noted that it is only the number of users
   sending text within the same moment that causes jerkiness, not the
   total number of users with real-time text capability.

3.18.  Security for Session Control and Media



   Security mechanisms to provide confidentiality, integrity protection,
   and peer authentication SHOULD be applied when possible regarding the
   capabilities of the participating devices by using the Session
   Initiation Protocol (SIP) over TLS by default according to
   Section 3.1.3 of [RFC5630] on the session control level and by
   default using DTLS-SRTP [RFC5764] at the media level.  In
   applications where legacy endpoints without security are allowed, a
   negotiation SHOULD be performed to decide if encryption at the media
   level will be applied.  If no other security solution is mandated for
   the application, then the Opportunistic Secure Real-time Transport
   Protocol (OSRTP) [RFC8643] is a suitable method to be applied to
   negotiate SRTP media security with DTLS.  For simplicity, most SDP
   examples below are expressed without the security additions.  The
   principles (but not all details) for applying DTLS-SRTP security
   [RFC5764] are shown in a couple of the following examples.

   Further general security considerations are covered in Section 10.

   End-to-end encryption would require further work and could be based
   on WebRTC as specified in Section 1.2 or on double encryption as
   specified in [RFC8723].

3.19.  SDP Offer/Answer Examples



   This section shows some examples of SDP for session negotiation of
   the real-time text media in SIP sessions.  Audio is usually provided
   in the same session, and sometimes also video.  The examples only
   show the part of importance for the real-time text media.  The
   examples relate to the single RTP stream mixing for multiparty-aware
   endpoints and for multiparty-unaware endpoints.

      |  Note: Multiparty real-time text MAY also be provided through
      |  other methods, e.g., by a Selective Forwarding Middlebox (SFM).
      |  In that case, the SDP of the offer will include something
      |  specific for that method, e.g., an SDP attribute or another
      |  media format.  An answer selecting the use of that method would
      |  accept it via a corresponding acknowledgement included in the
      |  SDP.  The offer may also contain the "rtt-mixer" SDP media
      |  attribute for the main real-time text media when the offerer
      |  has this capability for both multiparty methods, while an
      |  answer, choosing to use SFM, will not include the "rtt-mixer"
      |  SDP media attribute.

   Offer example for the "text/red" format, multiparty support, and
   capability for 90 characters per second:

      m=text 11000 RTP/AVP 100 98
      a=rtpmap:98 t140/1000
      a=fmtp:98 cps=90
      a=rtpmap:100 red/1000
      a=fmtp:100 98/98/98
      a=rtt-mixer

   Answer example from a multiparty-aware device:

      m=text 14000 RTP/AVP 100 98
      a=rtpmap:98 t140/1000
      a=fmtp:98 cps=90
      a=rtpmap:100 red/1000
      a=fmtp:100 98/98/98
      a=rtt-mixer

   Offer example for the "text/red" format, including multiparty and
   security:

      a=fingerprint: (fingerprint1)
      m=text 11000 RTP/AVP 100 98
      a=rtpmap:98 t140/1000
      a=rtpmap:100 red/1000
      a=fmtp:100 98/98/98
      a=rtt-mixer

   The "fingerprint" is sufficient to offer DTLS-SRTP, with the media
   line still indicating RTP/AVP.

      |  Note: For brevity, the entire value of the SDP "fingerprint"
      |  attribute is not shown in this and the following example.

   Answer example from a multiparty-aware device with security:

      a=fingerprint: (fingerprint2)
      m=text 16000 RTP/AVP 100 98
      a=rtpmap:98 t140/1000
      a=rtpmap:100 red/1000
      a=fmtp:100 98/98/98
      a=rtt-mixer

   With the "fingerprint", the device acknowledges the use of DTLS-SRTP.

   Answer example from a multiparty-unaware device that also does not
   support security:

      m=text 12000 RTP/AVP 100 98
      a=rtpmap:98 t140/1000
      a=rtpmap:100 red/1000
      a=fmtp:100 98/98/98

3.20.  Packet Sequence Example from Interleaved Transmission



   This example shows a symbolic flow of packets from a mixer, including
   loss and recovery.  The sequence includes interleaved transmission of
   text from two real-time text sources: A and B.  P indicates primary
   data.  R1 is the first redundant generation of data, and R2 is the
   second redundant generation of data.  A1, B1, A2, etc. are text
   chunks (T140blocks) received from the respective sources and sent on
   to the receiver by the mixer.  X indicates a dropped packet between
   the mixer and a receiver.  The session is assumed to use the original
   and two redundant generations of real-time text.

     |-----------------------|
     |Seq no 101, Time=20400 |
     |CC=1                   |
     |CSRC list A            |
     |R2: A1, Offset=600     |
     |R1: A2, Offset=300     |
     |P:  A3                 |
     |-----------------------|

   Assuming that earlier packets (with text A1 and A2) were received in
   sequence, text A3 is received from packet 101 and assigned to
   reception buffer A.  The mixer is now assumed to have received
   initial text from source B 100 ms after packet 101 and will send that
   text.  Transmission of A2 and A3 as redundancy is planned for 330 ms
   after packet 101 if no new text from A is ready to be sent before
   that.

      |-----------------------|
      |Seq no 102, Time=20500 |
      |CC=1                   |
      |CSRC list B            |
      |R2  Empty, Offset=600  |
      |R1: Empty, Offset=300  |
      |P:  B1                 |
      |-----------------------|

      Packet 102 is received.

      B1 is retrieved from this packet.  Redundant transmission of B1 is
      planned 330 ms after packet 102.

      X------------------------|
      X Seq no 103, Timer=20730|
      X CC=1                   |
      X CSRC list A            |
      X R2: A2, Offset=630     |
      X R1: A3, Offset=330     |
      X P:  Empty              |
      X------------------------|

      Packet 103 is assumed to be lost due to network problems.

      It contains redundancy for A.  Sending A3 as second-level
      redundancy is planned for 330 ms after packet 103.

      X------------------------|
      X Seq no 104, Timer=20800|
      X CC=1                   |
      X CSRC list B            |
      X R2: Empty, Offset=600  |
      X R1: B1, Offset=300     |
      X P:  B2                 |
      X------------------------|

      Packet 104 contains text from B, including new B2 and redundant
      B1.  It is assumed dropped due to network problems.

      The mixer has A3 redundancy to send, but no new text appears from
      A, and therefore the redundancy is sent 330 ms after the previous
      packet with text from A.

      |------------------------|
      | Seq no 105, Timer=21060|
      | CC=1                   |
      | CSRC list A            |
      | R2: A3, Offset=660     |
      | R1: Empty, Offset=330  |
      | P:  Empty              |
      |------------------------|

      Packet 105 is received.

      A gap for lost packets 103 and 104 is detected.  Assume that no
      other loss was detected during the last second.  It can then be
      concluded that nothing was totally lost.

      R2 is checked.  Its original time was 21060-660=20400.  A packet
      with text from A was received with that timestamp, so nothing
      needs to be recovered.

      B1 and B2 still need to be transmitted as redundancy.  This is
      planned 330 ms after packet 104.  That would be at 21130.

      |-----------------------|
      |Seq no 106, Timer=21130|
      |CC=1                   |
      |CSRC list B            |
      | R2: B1, Offset=630    |
      | R1: B2, Offset=330    |
      | P:  Empty             |
      |-----------------------|

      Packet 106 is received.

      The second-level redundancy in packet 106 is B1 and has a
      timestamp offset of 630 ms.  The timestamp of packet 106 minus 630
      is 20500, which is the timestamp of packet 102 that was received.
      So, B1 does not need to be retrieved.  The first-level redundancy
      in packet 106 has an offset of 330.  The timestamp of packet 106
      minus 330 is 20800.  That is later than the latest received packet
      with source B.  Therefore, B2 is retrieved and assigned to the
      input buffer for source B.  No primary is available in packet 106.

      After this sequence, A3, B1, and B2 have been received.  In this
      case, no text was lost.

3.21.  Maximum Character Rate "cps" Setting



   The default maximum rate of reception of "text/t140" real-time text,
   as specified in [RFC4103], is 30 characters per second.  The actual
   rate is calculated without regard to any redundant text transmission
   and is, in the multiparty case, evaluated for all sources
   contributing to transmission to a receiver.  The value MAY be
   modified in the "cps" parameter of the "fmtp" attribute for the
   "text/t140" format of the "text" media section.

   A mixer combining real-time text from a number of sources may
   occasionally have a higher combined flow of text coming from the
   sources.  Endpoints SHOULD therefore include a suitable higher value
   for the "cps" parameter, corresponding to its real reception
   capability.  The default "cps" value 30 can be assumed to be
   sufficient for small meetings and well-managed larger conferences
   with users only making manual text entry.  A "cps" value of 90 can be
   assumed to be sufficient even for large unmanaged conferences and for
   cases when speech-to-text technologies are used for text entry.  This
   is also a reachable performance for receivers in modern technologies,
   and 90 is therefore the RECOMMENDED "cps" value.  See [RFC4103] for
   the format and use of the "cps" parameter.  The same rules apply for
   the multiparty case.

4.  Presentation-Level Considerations



   "Protocol for multimedia application text conversation" [T140]
   provides the presentation-level requirements for RTP transport as
   described in [RFC4103].  Functions for erasure and other formatting
   functions are specified in [T140], which has the following general
   statement for the presentation:

   |  The display of text from the members of the conversation should be
   |  arranged so that the text from each participant is clearly
   |  readable, and its source and the relative timing of entered text
   |  is visualized in the display.  Mechanisms for looking back in the
   |  contents from the current session should be provided.  The text
   |  should be displayed as soon as it is received.

   Strict application of [T140] is essential for the interoperability of
   real-time text implementations and to fulfill the intention that the
   session participants have the same information conveyed in the text
   contents of the conversation without necessarily having the exact
   same layout of the conversation.

   [T140] specifies a set of presentation control codes (Section 4.2.4)
   to include in the stream.  Some of them are optional.
   Implementations MUST ignore optional control codes that they do not
   support.

   There is no strict "message" concept in real-time text.  The Unicode
   Line Separator character SHALL be used as a separator allowing a part
   of received text to be grouped in a presentation.  The character
   combination "CRLF" may be used by other implementations as a
   replacement for the Line Separator.  The "CRLF" combination SHALL be
   erased by just one erasing action, the same as the Line Separator.
   Presentation functions are allowed to group text for presentation in
   smaller groups than the Line Separators imply and present such groups
   with a source indication together with text groups from other sources
   (see the following presentation examples).  Erasure has no specific
   limit by any delimiter in the text stream.

4.1.  Presentation by Multiparty-Aware Endpoints



   A multiparty-aware receiving party presenting real-time text MUST
   separate text from different sources and present them in separate
   presentation fields.  The receiving party MAY separate the
   presentation of parts of text from a source in readable groups based
   on criteria other than a Line Separator and merge these groups in the
   presentation area when it benefits the user to most easily find and
   read text from the different participants.  The criteria MAY, for
   example, be a received comma, a full stop, some other type of phrase
   delimiter, or a long pause.

   When text is received from multiple original sources, the
   presentation SHALL provide a view where text is added in multiple
   presentation fields.

   If the presentation presents text from different sources in one
   common area, the presenting endpoint SHOULD insert text from the
   local user, where the text ends at suitable points and is merged
   properly with received text to indicate the relative timing for when
   the text groups were completed.  In this presentation mode, the
   receiving endpoint SHALL present the source of the different groups
   of text.  This presentation style is called the "chat" style here and
   provides the possibility of following text arriving from multiple
   parties and the approximate relative time that text is received as
   related to text from the local user.

   A view of a three-party real-time text call in chat style is shown in
   this example.

             _________________________________________________
            |                                              |^|
            |[Alice] Hi, Alice here.                       |-|
            |                                              | |
            |[Bob] Bob as well.                            | |
            |                                              | |
            |[Eve] Hi, this is Eve, calling from Paris.    | |
            |      I thought you should be here.           | |
            |                                              | |
            |[Alice] I am coming on Thursday, my           | |
            |      performance is not until Friday morning.| |
            |                                              | |
            |[Bob] And I on Wednesday evening.             | |
            |                                              | |
            |[Alice] Can we meet on Thursday evening?      | |
            |                                              | |
            |[Eve] Yes, definitely. How about 7pm.         | |
            |     at the entrance of the restaurant        | |
            |     Le Lion Blanc?                           | |
            |[Eve] we can have dinner and then take a walk |-|
            |______________________________________________|v|
            | <Eve-typing> But I need to be back to        |^|
            |    the hotel by 11 because I need            |-|
            |                                              | |
            | <Bob-typing> I wou                           |-|
            |______________________________________________|v|
            | of course, I underst                           |
            |________________________________________________|

      Figure 1: Example of a Three-Party Real-Time Text Call Presented
             in Chat Style Seen at Participant Alice's Endpoint

   Presentation styles other than the chat style MAY be arranged.

   Figure 2 shows how a coordinated column view MAY be presented.

   _____________________________________________________________________
   |       Bob          |       Eve            |       Alice           |
   |____________________|______________________|_______________________|
   |                    |                      |I will arrive by TGV.  |
   |My flight is to Orly|                      |Convenient to the main |
   |                    |Hi all, can we plan   |station.               |
   |                    |for the seminar?      |                       |
   |Eve, will you do    |                      |                       |
   |your presentation on|                      |                       |
   |Friday?             |Yes, Friday at 10.    |                       |
   |Fine, wo            |                      |We need to meet befo   |
   |___________________________________________________________________|

           Figure 2: An Example of a Coordinated Column View of a
           Three-Party Session with Entries Ordered Vertically in
                           Approximate Time Order

4.2.  Multiparty Mixing for Multiparty-Unaware Endpoints



   When the mixer has indicated multiparty real-time text capability in
   an SDP negotiation but the multiparty capability negotiation fails
   with an endpoint, the agreed-upon "text/red" or "text/t140" format
   SHALL be used and the mixer SHOULD compose a best-effort presentation
   of multiparty real-time text in one stream intended to be presented
   by an endpoint with no multiparty awareness, when that is desired in
   the actual implementation.  The following specifies a procedure that
   MAY be applied in that situation.

   This presentation format has functional limitations and SHOULD be
   used only to enable participation in multiparty calls by legacy
   deployed endpoints implementing only RFC 4103 without any multiparty
   extensions specified in this document.

   The principles and procedures below do not specify any new protocol
   elements.  They are instead composed of information provided in
   [T140] and an ambition to provide a best-effort presentation on an
   endpoint that has functions originally intended only for two-party
   calls.

   The mixer performing the mixing for multiparty-unaware endpoints
   SHALL compose a simulated, limited multiparty real-time text view
   suitable for presentation in one presentation area.  The mixer SHALL
   group text in suitable groups and prepare them for presentation by
   inserting a Line Separator between them if the transmitted text did
   not already end with a new line (Line Separator or CRLF).  A
   presentable label SHALL be composed and sent for the source initially
   in the session and after each source switch.  With this procedure,
   the time for switching from transmission of text from one source to
   transmission of text from another source depends on the actions of
   the users.  In order to expedite source switching, a user can, for
   example, end its turn with a new line.

4.2.1.  Actions by the Mixer at Reception from the Call Participants



   When text is received by the mixer from the different participants,
   the mixer SHALL recover text from redundancy if any packets are lost.
   The marker for lost text [T140ad1] SHALL be inserted in the stream if
   unrecoverable loss appears.  Any Unicode BOM characters, possibly
   used for keep-alives, SHALL be deleted.  The time of creation of text
   (retrieved from the RTP timestamp) SHALL be stored together with the
   received text from each source in queues for transmission to the
   recipients in order to be able to evaluate text loss.

4.2.2.  Actions by the Mixer for Transmission to the Recipients



   The following procedure SHALL be applied for each multiparty-unaware
   recipient of multiparty text from the mixer.

   The text for transmission SHALL be formatted by the mixer for each
   receiving user for presentation in one single presentation area.
   Text received from a participant SHOULD NOT be included in
   transmissions to that participant, because it is usually presented
   locally at transmission time.  When there is text available for
   transmission from the mixer to a receiving party from more than one
   participant, the mixer SHALL switch between transmission of text from
   the different sources at suitable points in the transmitted stream.

   When switching the source, the mixer SHALL insert a Line Separator if
   the already-transmitted text did not end with a new line (Line
   Separator or CRLF).  A label SHALL be composed of information in the
   CNAME and NAME fields in RTCP reports from the participant to have
   its text transmitted, or from other session information for that
   user.  The label SHALL be delimited by suitable characters (e.g.,
   "[ ]") and transmitted.  The CSRC SHALL indicate the selected source.
   Then, text from that selected participant SHALL be transmitted until
   a new suitable point for switching the source is reached.

   Information available to the mixer for composing the label may
   contain sensitive personal information that SHOULD NOT be revealed in
   sessions not securely authenticated and confidentiality protected.
   Privacy considerations regarding how much personal information is
   included in the label SHOULD therefore be taken when composing the
   label.

   Seeking a suitable point for switching the source SHALL be done when
   there is older text waiting for transmission from any party than the
   age of the last transmitted text.  Suitable points for switching are:

   *  A completed phrase ending with a comma.

   *  A completed sentence.

   *  A new line (Line Separator or CRLF).

   *  A long pause (e.g., > 10 seconds) in received text from the
      currently transmitted source.

   *  If text from one participant has been transmitted with text from
      other sources waiting for transmission for a long time (e.g., > 1
      minute) and none of the other suitable points for switching has
      occurred, a source switch MAY be forced by the mixer at the next
      word delimiter, and also even if a word delimiter does not occur
      within some period of time (e.g., 15 seconds) after the scan for a
      word delimiter started.

   When switching the source, the source that has the oldest text in
   queue SHALL be selected to be transmitted.  A character display count
   SHALL be maintained for the currently transmitted source, starting at
   zero after the label is transmitted for the currently transmitted
   source.

   The status SHALL be maintained for the latest control code for Select
   Graphic Rendition (SGR) from each source.  If there is an SGR code
   stored as the status for the current source before the source switch
   is done, a reset of SGR SHALL be sent by the sequence SGR 0 [U+009B
   U+0000 U+006D] after the new line and before the new label during a
   source switch.  See Section 4.2.4 for an explanation.  This
   transmission does not influence the display count.

   If there is an SGR code stored for the new source after the source
   switch, that SGR code SHALL be transmitted to the recipient before
   the label.  This transmission does not influence the display count.

4.2.3.  Actions on Transmission of Text



   Text from a source sent to the recipient SHALL increase the display
   count by one per transmitted character.

4.2.4.  Actions on Transmission of Control Codes



   The following control codes, as specified by T.140 [T140], require
   specific actions.  They SHALL cause specific considerations in the
   mixer.  Note that the codes presented here are expressed in UTF-16,
   while transmission is made in the UTF-8 encoding of these codes.

   BEL (U+0007):  Bell.  Alert in session.  Provides for alerting during
      an active session.  The display count SHALL NOT be altered.

   NEW LINE (U+2028):  Line Separator.  Check and perform a source
      switch if appropriate.  Increase the display count by 1.

   CR LF (U+000D U+000A):  A supported, but not preferred, way of
      requesting a new line.  Check and perform a source switch if
      appropriate.  Increase the display count by 1.

   INT (ESC U+0061):  Interrupt (used to initiate the mode negotiation
      procedure).  The display count SHALL NOT be altered.

   SGR (U+009B Ps U+006D):  Select Graphic Rendition.  Ps represents the
      rendition parameters specified in [ISO6429].  (For freely
      available equivalent information, please see [ECMA-48].)  The
      display count SHALL NOT be altered.  The SGR code SHOULD be stored
      for the current source.

   SOS (U+0098):  Start of String.  Used as a general protocol element
      introducer, followed by a maximum 256-byte string and the ST.  The
      display count SHALL NOT be altered.

   ST (U+009C):  String Terminator.  End of SOS string.  The display
      count SHALL NOT be altered.

   ESC (U+001B):  Escape.  Used in control strings.  The display count
      SHALL NOT be altered for the complete escape code.

   Byte order mark (BOM) (U+FEFF):  "Zero width no-break space".  Used
      for synchronization and keep-alive.  It SHALL be deleted from
      incoming streams.  It SHALL also be sent first after session
      establishment to the recipient.  The display count SHALL NOT be
      altered.

   Missing text mark (U+FFFD):  "Replacement character".  Represented as
      a question mark in a rhombus, or, if that is not feasible,
      replaced by an apostrophe (').  It marks the place in the stream
      of possible text loss.  This mark SHALL be inserted by the
      reception procedure in the case of unrecoverable loss of packets.
      The display count SHALL be increased by one when sent as for any
      other character.

   SGR:  If a control code for SGR other than a reset of the graphic
      rendition (SGR 0) is sent to a recipient, that control code SHALL
      also be stored as the status for the source in the storage for SGR
      status.  If a reset graphic rendition (SGR 0) originating from a
      source is sent, then the SGR status storage for that source SHALL
      be cleared.  The display count SHALL NOT be increased.

   BS (U+0008):  "Back Space".  Intended to erase the last entered
      character by a source.  Erasure by backspace cannot always be
      performed as the erasing party intended.  If an erasing action
      erases all text up to the end of the leading label after a source
      switch, then the mixer MUST NOT transmit more backspaces.
      Instead, it is RECOMMENDED that a letter "X" be inserted in the
      text stream for each backspace as an indication of the intent to
      erase more.  A new line is usually coded by a Line Separator, but
      the character combination "CRLF" MAY be used instead.  Erasure of
      a new line is, in both cases, done by just one erasing action
      (backspace).  If the display count has a positive value, it SHALL
      be decreased by one when the BS is sent.  If the display count is
      at zero, it SHALL NOT be altered.

4.2.5.  Packet Transmission



   A mixer transmitting to a multiparty-unaware endpoint SHALL send
   primary data only from one source per packet.  The SSRC SHALL be the
   SSRC of the mixer.  The CSRC list MAY contain one member and be the
   SSRC of the source of the primary data.

4.2.6.  Functional Limitations



   When a multiparty-unaware endpoint presents a conversation in one
   display area in a chat style, it inserts source indications for
   remote text and local user text as they are merged in completed text
   groups.  When an endpoint using this layout receives and presents
   text mixed for multiparty-unaware endpoints, there will be two levels
   of source indicators for the received text: one generated by the
   mixer and inserted in a label after each source switch, and another
   generated by the receiving endpoint and inserted after each switch
   between the local source and the remote source in the presentation
   area.  This will waste display space and look inconsistent to the
   reader.

   New text can be presented from only one source at a time.  Switching
   the source to be presented takes place at suitable places in the
   text, such as the end of a phrase, the end of a sentence, or a Line
   Separator, or upon detecting inactivity.  Therefore, the time to
   switch to present waiting text from other sources may grow long, and
   it will vary and depend on the actions of the currently presented
   source.

   Erasure can only be done up to the latest source switch.  If a user
   tries to erase more text, the erasing actions will be presented as a
   letter "X" after the label.

   Text loss because of network errors may hit the label between entries
   from different parties, causing the risk of a misunderstanding
   regarding which source provided a piece of text.

   Because of these facts, it is strongly RECOMMENDED that multiparty
   awareness be implemented in real-time text endpoints.  The use of the
   mixing method for multiparty-unaware endpoints should be left for use
   with endpoints that are impossible to upgrade to become multiparty
   aware.

4.2.7.  Example Views of Presentation on Multiparty-Unaware Endpoints



   The following pictures are examples of the view on a participant's
   display for the multiparty-unaware case.

   Figure 3 shows how a coordinated column view MAY be presented on
   Alice's device in a view with two columns.  The mixer inserts labels
   to show how the sources alternate in the column with received text.
   The mixer alternates between the sources at suitable points in the
   text exchange so that text entries from each party can be
   conveniently read.

             ___________________________________________________
            |       Conference        |          Alice          |
            |_________________________|_________________________|
            |                         |I will arrive by TGV.    |
            |[Bob]: My flight is to   |Convenient to the main   |
            |Orly.                    |station.                 |
            |[Eve]: Hi all, can we    |                         |
            |plan for the seminar.    |                         |
            |                         |                         |
            |[Bob]: Eve, will you do  |                         |
            |your presentation on     |                         |
            |Friday?                  |                         |
            |[Eve]: Yes, Friday at 10.|                         |
            |[Bob]: Fine, wo          |We need to meet befo     |
            |_________________________|_________________________|

          Figure 3: Alice, Who Has a Conference-Unaware Client, Is
         Receiving the Multiparty Real-Time Text in a Single Stream

   In Figure 4, there is a tradition in receiving applications to
   include a label showing the source of the text, here shown with
   parentheses "()".  The mixer also inserts source labels for the
   multiparty call participants, here shown with brackets "[]".

              _________________________________________________
             |                                              |^|
             |(Alice) Hi, Alice here.                       |-|
             |                                              | |
             |(mix)[Bob] Bob as well.                       | |
             |                                              | |
             |[Eve] Hi, this is Eve, calling from Paris     | |
             |      I thought you should be here.           | |
             |                                              | |
             |(Alice) I am coming on Thursday, my           | |
             |      performance is not until Friday morning.| |
             |                                              | |
             |(mix)[Bob] And I on Wednesday evening.        | |
             |                                              | |
             |[Eve] we can have dinner and then walk        | |
             |                                              | |
             |[Eve] But I need to be back to                | |
             |    the hotel by 11 because I need            | |
             |                                              |-|
             |______________________________________________|v|
             | of course, I underst                           |
             |________________________________________________|

          Figure 4: An Example of a View of the Multiparty-Unaware
         Presentation in Chat Style, Where Alice Is the Local User

5.  Relationship to Conference Control



5.1.  Use with SIP Centralized Conferencing Framework



   The Session Initiation Protocol (SIP) conferencing framework, mainly
   specified in [RFC4353], [RFC4579], and [RFC4575], is suitable for
   coordinating sessions, including multiparty real-time text.  The
   real-time text stream between the mixer and a participant is one and
   the same during the conference.  Participants get announced by
   notifications when participants are joining or leaving, and further
   user information may be provided.  The SSRC of the text to expect
   from joined users MAY be included in a notification.  The
   notifications MAY be used for both security purposes and translation
   to a label for presentation to other users.

5.2.  Conference Control



   In managed conferences, control of the real-time text media SHOULD be
   provided in the same way as for other media, e.g., for muting and
   unmuting by the direction attributes in SDP [RFC8866].

   Note that floor control functions may be of value for real-time text
   users as well as for users of other media in a conference.

6.  Gateway Considerations



   Multiparty real-time text sessions may involve gateways of different
   kinds.  Gateways involved in setting up sessions SHALL correctly
   reflect the multiparty capability or unawareness of the combination
   of the gateway and the remote endpoint beyond the gateway.

6.1.  Gateway Considerations with Textphones



   One case that may occur is a gateway to the Public Switched Telephone
   Network (PSTN) for communication with textphones (e.g., TTYs).
   Textphones are limited devices with no multiparty awareness, and it
   SHOULD therefore be appropriate for the gateway to not indicate
   multiparty awareness for that case.  Another solution is that the
   gateway indicates multiparty capability towards the mixer and
   includes the multiparty mixer function for multiparty-unaware
   endpoints itself.  This solution makes it possible to adapt to the
   functional limitations of the textphone.

   More information on gateways to textphones is found in [RFC5194].

6.2.  Gateway Considerations with WebRTC



   Gateway operation between RTP-mixer-based multiparty real-time text
   and WebRTC-based real-time text may also be required.  Real-time text
   transport in WebRTC is specified in [RFC8865].

   A multiparty bridge may have functionality for communicating via
   real-time text in both (1) RTP streams with real-time text and (2)
   WebRTC T.140 data channels.  Other configurations may consist of a
   multiparty bridge with either technology for real-time text transport
   and a separate gateway for conversion of the text communication
   streams between RTP and T.140 data channels.

   In WebRTC, it is assumed that for a multiparty session, one T.140
   data channel is established for each source from a gateway or bridge
   to each participant.  Each participant also has a data channel with a
   two-way connection with the gateway or bridge.

   A T.140 data channel used for two-way communication is for text from
   the WebRTC user and from the bridge or gateway itself to the WebRTC
   user.  The label parameter of this T.140 data channel is used as the
   NAME field in RTCP to participants on the RTP side.  The other T.140
   data channels are only for text from other participants to the WebRTC
   user.

   When a new participant has entered the session with RTP transport of
   real-time text, a new T.140 data channel SHOULD be established to
   WebRTC users with the label parameter composed of information from
   the NAME field in RTCP on the RTP side.

   When a new participant has entered the multiparty session with real-
   time text transport in a WebRTC T.140 data channel, the new
   participant SHOULD be announced by a notification to RTP users.  The
   label parameter from the WebRTC side or other suitable information
   from the session or stream establishment procedure SHOULD be used to
   compose the NAME RTCP field on the RTP side.

   When a participant on the RTP side is disconnected from the
   multiparty session, the corresponding T.140 data channel(s) SHOULD be
   closed.

   When a WebRTC user of T.140 data channels disconnects from the mixer,
   the corresponding RTP streams or sources in an RTP-mixed stream
   SHOULD be closed.

   T.140 data channels MAY be opened and closed by negotiation or
   renegotiation of the session, or by any other valid means, as
   specified in Section 1 of [RFC8865].

7.  Updates to RFC 4103



   This document updates [RFC4103] by introducing an SDP media
   attribute, "rtt-mixer", for negotiation of multiparty-mixing
   capability with the format described in [RFC4103] and by specifying
   the rules for packets when multiparty capability is negotiated and in
   use.

8.  Congestion Considerations



   The congestion considerations and recommended actions provided in
   [RFC4103] are also valid in multiparty situations.

   The time values SHALL then be applied per source of text sent to a
   receiver.

   In the very unlikely event that many participants in a conference
   send text simultaneously for a long period of time, a delay may build
   up for the presentation of text at the receivers if the limitation in
   characters per second ("cps") to be transmitted to the participants
   is exceeded.  A delay of more than 15 seconds can cause confusion in
   the session.  It is therefore RECOMMENDED that an RTP mixer discard
   such text causing excessive delays and insert a general indication of
   possible text loss [T140ad1] in the session.  If the main text
   contributor is indicated in any way, the mixer MAY avoid deleting
   text from that participant.  It should, however, be noted that human
   creation of text normally contains pauses, when the transmission can
   catch up, so that transmission-overload situations are expected to be
   very rare.

9.  IANA Considerations



9.1.  Registration of the "rtt-mixer" SDP Media Attribute



   IANA has registered the new SDP attribute "rtt-mixer".

   Contact name:  IESG

   Contact email:  iesg@ietf.org

   Attribute name:  rtt-mixer

   Attribute semantics:  See RFC 9071, Section 2.3

   Attribute value:  none

   Usage level:  media

   Purpose:  To indicate mixer and endpoint support of multiparty mixing
      for real-time text transmission, using a common RTP stream for
      transmission of text from a number of sources mixed with one
      source at a time and where the source is indicated in a single
      CSRC-list member.

   Charset Dependent:  no

   O/A procedures:  See RFC 9071, Section 2.3

   Mux Category:  NORMAL

   Reference:  RFC 9071

10.  Security Considerations



   The RTP-mixer model requires the mixer to be allowed to decrypt,
   pack, and encrypt secured text from conference participants.
   Therefore, the mixer needs to be trusted to maintain confidentiality
   and integrity of the real-time text data.  This situation is similar
   to the situation for handling audio and video media in centralized
   mixers.

   The requirement to transfer information about the user in RTCP
   reports in SDES, CNAME, and NAME fields, and in conference
   notifications, may have privacy concerns, as already stated in RFC
   3550 [RFC3550], and may be restricted for privacy reasons.  When used
   for the creation of readable labels in the presentation, the
   receiving user will then get a more symbolic label for the source.

   The services available through the real-time text mixer may be of
   special interest to deaf and hard-of-hearing individuals.  Some users
   may want to refrain from revealing such characteristics broadly in
   conferences.  Conference systems where the mixer is included MAY need
   to be designed with the confidentiality of such characteristics in
   mind.

   Participants with malicious intentions may appear and, for example,
   disrupt the multiparty session by emitting a continuous flow of text.
   They may also send text that appears to originate from other
   participants.  Countermeasures should include requiring secure
   signaling, media, and authentication, and providing higher-layer
   conference functions, e.g., for blocking, muting, and expelling
   participants.

   Participants with malicious intentions may also try to disrupt the
   presentation by sending incomplete or malformed control codes.
   Handling of text from the different sources by the receivers MUST
   therefore be well separated so that the effects of such actions only
   affect text from the source causing the action.

   Care should be taken to avoid the possibility of attacks by
   unauthenticated call participants, and even eavesdropping and
   manipulation of content by non-participants, if the use of the mixer
   is permitted for users both with and without security procedures.

   As already stated in Section 3.18, security in media SHOULD be
   applied by using DTLS-SRTP [RFC5764] at the media level.

   Further security considerations specific to this application are
   specified in Section 3.18.

11.  References



11.1.  Normative References



   [ECMA-48]  Ecma International, "ECMA-48: Control functions for coded
              character sets", 5th edition, June 1991,
              <https://www.ecma-international.org/publications-and-
              standards/standards/ecma-48/>.

   [ISO6429]  ISO/IEC, "Information technology - Control functions for
              coded character sets", ISO/IEC ISO/IEC 6429:1992, December
              1992, <https://www.iso.org/obp/ui/#iso:std:iso-
              iec:6429:ed-3:v1:en>.

   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119,
              DOI 10.17487/RFC2119, March 1997,
              <https://www.rfc-editor.org/info/rfc2119>.

   [RFC3550]  Schulzrinne, H., Casner, S., Frederick, R., and V.
              Jacobson, "RTP: A Transport Protocol for Real-Time
              Applications", STD 64, RFC 3550, DOI 10.17487/RFC3550,
              July 2003, <https://www.rfc-editor.org/info/rfc3550>.

   [RFC4102]  Jones, P., "Registration of the text/red MIME Sub-Type",
              RFC 4102, DOI 10.17487/RFC4102, June 2005,
              <https://www.rfc-editor.org/info/rfc4102>.

   [RFC4103]  Hellstrom, G. and P. Jones, "RTP Payload for Text
              Conversation", RFC 4103, DOI 10.17487/RFC4103, June 2005,
              <https://www.rfc-editor.org/info/rfc4103>.

   [RFC5630]  Audet, F., "The Use of the SIPS URI Scheme in the Session
              Initiation Protocol (SIP)", RFC 5630,
              DOI 10.17487/RFC5630, October 2009,
              <https://www.rfc-editor.org/info/rfc5630>.

   [RFC5764]  McGrew, D. and E. Rescorla, "Datagram Transport Layer
              Security (DTLS) Extension to Establish Keys for the Secure
              Real-time Transport Protocol (SRTP)", RFC 5764,
              DOI 10.17487/RFC5764, May 2010,
              <https://www.rfc-editor.org/info/rfc5764>.

   [RFC6263]  Marjou, X. and A. Sollaud, "Application Mechanism for
              Keeping Alive the NAT Mappings Associated with RTP / RTP
              Control Protocol (RTCP) Flows", RFC 6263,
              DOI 10.17487/RFC6263, June 2011,
              <https://www.rfc-editor.org/info/rfc6263>.

   [RFC7675]  Perumal, M., Wing, D., Ravindranath, R., Reddy, T., and M.
              Thomson, "Session Traversal Utilities for NAT (STUN) Usage
              for Consent Freshness", RFC 7675, DOI 10.17487/RFC7675,
              October 2015, <https://www.rfc-editor.org/info/rfc7675>.

   [RFC8174]  Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC
              2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174,
              May 2017, <https://www.rfc-editor.org/info/rfc8174>.

   [RFC8865]  Holmberg, C. and G. Hellström, "T.140 Real-Time Text
              Conversation over WebRTC Data Channels", RFC 8865,
              DOI 10.17487/RFC8865, January 2021,
              <https://www.rfc-editor.org/info/rfc8865>.

   [RFC8866]  Begen, A., Kyzivat, P., Perkins, C., and M. Handley, "SDP:
              Session Description Protocol", RFC 8866,
              DOI 10.17487/RFC8866, January 2021,
              <https://www.rfc-editor.org/info/rfc8866>.

   [T140]     ITU-T, "Protocol for multimedia application text
              conversation", ITU-T Recommendation T.140, February 1998,
              <https://www.itu.int/rec/T-REC-T.140-199802-I/en>.

   [T140ad1]  ITU-T, "Recommendation T.140 Addendum", February 2000,
              <https://www.itu.int/rec/T-REC-T.140-200002-I!Add1/en>.

11.2.  Informative References



   [RFC4353]  Rosenberg, J., "A Framework for Conferencing with the
              Session Initiation Protocol (SIP)", RFC 4353,
              DOI 10.17487/RFC4353, February 2006,
              <https://www.rfc-editor.org/info/rfc4353>.

   [RFC4575]  Rosenberg, J., Schulzrinne, H., and O. Levin, Ed., "A
              Session Initiation Protocol (SIP) Event Package for
              Conference State", RFC 4575, DOI 10.17487/RFC4575, August
              2006, <https://www.rfc-editor.org/info/rfc4575>.

   [RFC4579]  Johnston, A. and O. Levin, "Session Initiation Protocol
              (SIP) Call Control - Conferencing for User Agents",
              BCP 119, RFC 4579, DOI 10.17487/RFC4579, August 2006,
              <https://www.rfc-editor.org/info/rfc4579>.

   [RFC5194]  van Wijk, A., Ed. and G. Gybels, Ed., "Framework for Real-
              Time Text over IP Using the Session Initiation Protocol
              (SIP)", RFC 5194, DOI 10.17487/RFC5194, June 2008,
              <https://www.rfc-editor.org/info/rfc5194>.

   [RFC7667]  Westerlund, M. and S. Wenger, "RTP Topologies", RFC 7667,
              DOI 10.17487/RFC7667, November 2015,
              <https://www.rfc-editor.org/info/rfc7667>.

   [RFC8643]  Johnston, A., Aboba, B., Hutton, A., Jesske, R., and T.
              Stach, "An Opportunistic Approach for Secure Real-time
              Transport Protocol (OSRTP)", RFC 8643,
              DOI 10.17487/RFC8643, August 2019,
              <https://www.rfc-editor.org/info/rfc8643>.

   [RFC8723]  Jennings, C., Jones, P., Barnes, R., and A.B. Roach,
              "Double Encryption Procedures for the Secure Real-Time
              Transport Protocol (SRTP)", RFC 8723,
              DOI 10.17487/RFC8723, April 2020,
              <https://www.rfc-editor.org/info/rfc8723>.

   [RFC8825]  Alvestrand, H., "Overview: Real-Time Protocols for
              Browser-Based Applications", RFC 8825,
              DOI 10.17487/RFC8825, January 2021,
              <https://www.rfc-editor.org/info/rfc8825>.

Acknowledgements



   The author wants to thank the following persons for support, reviews,
   and valuable comments: Bernard Aboba, Amanda Baber, Roman Danyliw,
   Spencer Dawkins, Martin Duke, Lars Eggert, James Hamlin, Benjamin
   Kaduk, Murray Kucherawy, Paul Kyzivat, Jonathan Lennox, Lorenzo
   Miniero, Dan Mongrain, Francesca Palombini, Colin Perkins, Brian
   Rosen, Rich Salz, Jürgen Schönwälder, Robert Wilton, Dale Worley,
   Yong Xin, and Peter Yee.

Author's Address



   Gunnar Hellström
   Gunnar Hellström Accessible Communication
   SE-13670 Vendelsö
   Sweden

   Email: gunnar.hellstrom@ghaccess.se