draft-ietf-mmusic-trickle-ice-01.xml

<?xml version="1.0" encoding="UTF-8"?>
<?xml-stylesheet type='text/xsl' href='rfc2629.xslt' ?>

<!DOCTYPE rfc SYSTEM "rfc2629.dtd">

<rfc category='std' ipr='trust200902'
     docName='draft-ietf-mmusic-trickle-ice-01'>

<?rfc toc='yes' ?>
<?rfc symrefs='yes' ?>
<?rfc sortrefs='yes'?>
<?rfc iprnotified='no' ?>
<?rfc strict='yes' ?>
<?rfc compact='yes' ?>
  <front>

    <title abbrev='Trickle ICE'>
      Trickle ICE: Incremental Provisioning of Candidates for the
      Interactive Connectivity Establishment (ICE) Protocol
    </title>
    <author initials='E.' surname='Ivov'
            fullname='Emil Ivov'>
      <organization abbrev='Jitsi'>Jitsi</organization>
      <address>
        <postal>
          <street></street>
          <city>Strasbourg</city>
          <code>67000</code>
          <country>France</country>
        </postal>
        <phone>+33 6 72 81 15 55</phone>
        <email>emcho@jitsi.org</email>
      </address>
    </author>
    <author fullname="Eric Rescorla" initials="E.K." surname="Rescorla">
      <organization>RTFM, Inc.</organization>
      <address>
        <postal>
          <street>2064 Edgewood Drive</street>
          <city>Palo Alto</city>
          <region>CA</region>
          <code>94303</code>
          <country>USA</country>
        </postal>
        <phone>+1 650 678 2350</phone>
        <email>ekr@rtfm.com</email>
      </address>
    </author>
    <author fullname="Justin Uberti" initials="J." surname="Uberti">
      <organization>Google</organization>
      <address>
        <postal>
          <street>747 6th St S</street>
          <city>Kirkland</city>
          <region>WA</region>
          <code>98033</code>
          <country>USA</country>
        </postal>
        <phone>+1 857 288 8888</phone>
        <email>justin@uberti.name</email>
      </address>
    </author>
    <date />
    <abstract>
      <t>
        This document describes an extension to the Interactive
        Connectivity Establishment (ICE) protocol that allows ICE agents
        to send and receive candidates incrementally rather than
        exchanging complete lists. With such incremental provisioning,
        ICE agents can begin connectivity checks while they are still
        gathering candidates and considerably shorten the time necessary
        for ICE processing to complete.
      </t>
      <t>
        The above mechanism is also referred to as "trickle ICE".
      </t>
    </abstract>
  </front>
  <middle>
    <section title='Introduction'>
      <t>
        The Interactive Connectivity Establishment (ICE) protocol
        <xref target="RFC5245"/> describes mechanisms for gathering,
        candidates, prioritizing them, choosing default ones, exchanging
        them with the remote party, pairing them and ordering them into
        check lists. Once all of the above have been completed, and only
        then, the participating agents can begin a phase of connectivity
        checks and eventually select the pair of candidates that will be
        used in the following session.
      </t>
      <t>
        While the above sequence has the advantage of being relatively
        straightforward to implement and debug once deployed, it may
        also prove to be rather lengthy. Gathering candidates or
        candidate harvesting would often involve things like querying
        <xref target="RFC5389">STUN</xref> servers, discovering UPnP
        devices, and allocating relayed candidates at
        <xref target="RFC5766">TURN</xref> servers. All of these can
        be delayed for a noticeable amount of time and while they can be
        run in parallel, they still need to respect the pacing
        requirements from <xref target="RFC5245"/>, which is likely to
        delay them even further. Some or all of the above would also
        have to be completed by the remote agent. Both agents would
        next perform connectivity checks and only then would they be
        ready to begin streaming media.
      </t>
      <t>
        All of the above could lead to relatively lengthy session
        establishment times and degraded user experience.
      </t>
      <t>
        The purpose of this document is to define an alternative mode of
        operation for ICE implementations, also known as "trickle ICE",
        where candidates can be exchanged incrementally. This would
        allow ICE agents to exchange host candidates as soon as a
        session has been initiated. Connectivity checks for a media
        stream would also start as soon as the first candidates for that
        stream have become available.
      </t>
      <t>
        Trickle ICE allows reducing session establishment times in cases
        where connectivity is confirmed for the first exchanged
        candidates (e.g. where the host candidates for one of the agents
        are directly reachable from the second agent). Even when this is
        not the case, running candidate harvesting for both agents and
        connectivity checks all in parallel allows to considerably
        reduce ICE processing times.
      </t>
      <t>
        It is worth pointing out that before being introduced to the
        IETF, trickle ICE had already been included in specifications
        such as <xref target="XEP-0176">XMPP Jingle</xref> and it has
        been in use in various implementations and deployments.
      </t>
      <t>
        In addition to the basics of trickle ICE, this document also
        describes how support for trickle ICE needs to be discovered,
        how regular ICE processing needs to be modified when
        building and updating check lists, and how trickle ICE
        implementations should interoperate with agents that only
        implement <xref target="RFC5245"/> processing.
      </t>
      <t>
        This specification does not define usage of trickle ICE with any
        specific signalling protocol, contrary to
        <xref target="RFC5245"/> which contains a usage for ICE with
        SIP. Such usages would have to be specified in separate
        documents such as for example
        <xref target="I-D.ivov-mmusic-trickle-ice-sip"/>.
      </t>
      <t>
        Trickle ICE does however reuse and build upon the SDP syntax
        defined by vanilla ICE.
      </t>
    </section>
    <section title="Terminology">
      <t>
        The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL
        NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and
        "OPTIONAL" in this document are to be interpreted as described
        in <xref target="RFC2119"/>.
      </t>
      <t>
        This specification makes use of all terminology defined by the
        protocol for Interactive Connectivity Establishment in
        <xref target="RFC5245"/>.
      </t>
      <t>
        <list style="hanging">
          <t hangText="Vanilla ICE:">
            The Interactive Connectivity Establishment protocol as
            defined in <xref target="RFC5245"/>.
          </t>
          <t hangText="Candidate Harvester:">
            A module used by an ICE agent to obtain local candidates.
            Candidate harvesters use different mechanisms for
            discovering local candidates. Some of them would typically
            make use of protocols such as STUN or TURN. Others may also
            employ techniques that are not referenced within
            <xref target="RFC5245"/>. UPnP based port allocation and
            XMPP Jingle Relay Nodes <xref target="XEP-0278"/> are among
            the possible examples.
          </t>
          <t hangText="Trickled Candidates:">
            Candidates that a trickle ICE agent is sending subsequently
            to but within the context defined by an offer or an answer.
            Trickled candidates can be sent in parallel with candidate
            harvesting and connectivity checks.
          </t>
          <t hangText="Trickling/Trickle (v.):">
            The act of sending trickled candidates.
          </t>
          <t hangText="Half Trickle:">
            A trickle ICE mode of operation where the offerer gathers
            its first generation of candidates strictly before creating
            and sending the offer. Once sent, that offer can be
            processed by vanilla ICE agents and does not require support
            for this specification. It also allows trickle ICE capable
            answerers to still gather candidates and perform
            connectivity checks in a non-blocking way, thus roughly
            offering "half" the advantages of trickle ICE. The mechanism
            is mostly meant for use in cases where support for trickle
            ICE cannot be confirmed prior to sending a first offer.
          </t>
          <t hangText="Full Trickle:">
            Regular mode of operation for trickle ICE agents, used in
            opposition to the half trickle mode of operation.
          </t>
        </list>
      </t>
    </section>
    <section title='Incompatibility with Standard ICE'
             anchor='incompat'>
      <t>
        The ICE protocol was designed to be fairly flexible so that it
        would work in and adapt to as many network environments as
        possible. It is hence important to point out at least some of
        the reasons why, despite its flexibility, the specification in
        <xref target="RFC5245"/> would not support trickle ICE.
      </t>
      <t>
        <xref target="RFC5245"/> describes the conditions required to
        update check lists and timer states while an ICE agent is in the
        Running state. These conditions are verified upon transaction
        completion and one of them stipulates that:
      </t>
      <t>
        <list style='empty'>
          <t>
            If there is not a pair in the valid list for each component
            of the media stream, the state of the check list is set to
            Failed.
          </t>
        </list>
      </t>
      <t>
        This could be a problem and cause ICE processing to fail
        prematurely in a number of scenarios. Consider the following
        case:
      </t>
      <t>
        <list style='symbols'>
          <t>
            Alice and Bob are both located in different networks with
            Network Address Translation (NAT). Alice and Bob themselves
            have different address but both networks use the same
            <xref target="RFC1918"/> block.
          </t>
          <t>
            Alice sends Bob the candidate 10.0.0.10 which also happens
            to correspond to an existing host on Bob's network.
          </t>
          <t>
            Bob creates a check list consisting solely of 10.0.0.10 and
            starts checks.
          </t>
          <t>
            These checks reach the host at 10.0.0.10 in Bob's network,
            which responds with an ICMP "port unreachable" error and per
            <xref target="RFC5245"/> Bob marks the transaction as
            Failed.
          </t>
        </list>
        At this point the check list only contains Failed candidates and
        the valid list is empty. This causes the media stream and
        potentially all ICE processing to Fail.
      </t>
      <t>
        A similar race condition would occur if the initial offer from
        Alice only contains candidates that can be determined as
        unreachable (per
        <xref target="I-D.keranen-mmusic-ice-address-selection"/>) from
        any of the candidates that Bob has gathered. This would be the
        case if Bob's candidates only contain IPv4 addresses and the
        first candidate that he receives from Alice is an IPv6 one.
      </t>
      <t>
        Another potential problem could arise when a non-trickle
        ICE implementation sends an offer to a trickle one. Consider the
        following case:
        <list style='symbols'>
          <t>
            Alice's client has a non-trickle ICE implementation
          </t>
          <t>
            Bob's client has support for trickle ICE.
          </t>
          <t>
            Alice and Bob are behind NATs with address-dependent
            filtering <xref target="RFC4787"/>.
          </t>
          <t>
            Bob has two STUN servers but one of them is currently
            unreachable
          </t>
        </list>
      </t>
      <t>
        After Bob's agent receives Alice's offer it would immediately
        start connectivity checks. It would also start gathering
        candidates, which would take long because of the unreachable
        STUN server. By the time Bob's answer is ready and sent to
        Alice, Bob's connectivity checks may well have failed: until
        Alice gets Bob's answer, she won't be able to start connectivity
        checks and punch holes in her NAT. The NAT would hence be
        filtering Bob's checks as originating from an unknown endpoint.
      </t>
    </section>
    <section title='Determining Support for Trickle ICE' anchor="disco">
      <t>
        According to <xref target="RFC5245"/> every time an agent
        supporting trickle ICE generates an offer or an answer, it MUST
        include the "trickle" token in the ice-options attribute.
        Syntax for this token is defined in <xref target="sdp.offer"/>.
      </t>
      <t>
        Additionally, in order to avoid interoperability problems such
        as those described in <xref target="incompat"/>, it is important
        that trickle ICE negotiation is only attempted in cases where
        the remote party actually supports this specification. Agents
        that receive offers or answers can verify support by examining
        them for the "trickle" ice-options token. However, agents
        that are about to send a first offer, have no immediate way of
        doing this. This means that usages of trickle for specific
        protocols would need to either:
      </t>
      <t>
        <list style='symbols'>
          <t>
            Provide a way for agents to verify support of trickle ICE
            prior to initiating a session. XMPP's
            <xref target="XEP-0030"> Service discovery</xref> is an
            example for one such mechanism;
          </t>
          <t>
            Make support for trickle ICE mandatory so that support could
            be assumed the agents.
          </t>
        </list>
      </t>
      <t>
        Alternately, for cases where a protocol provides neither of the
        above, agents may either rely on provisioning/configuration, or
        use the half trickle procedure described in
        <xref target="half-trickle"/>.
      </t>
      <t>
        Note that out-of-band discovery semantics and half trickle are
        only necessary prior to session initiation, or in other words,
        when sending the initial offer. Once a session is established
        and trickle ICE support is confirmed for both parties, either
        agent can use full trickle for subsequent offers.
      </t>
      <section title='Unilateral Use of Trickle ICE (Half Trickle)'
               anchor="half-trickle">
        <t>
          The idea of using half trickle is about having the caller
          send a regular, vanilla ICE offer, with a complete set of
          candidates. This offer still indicates support for
          trickle ice, so the answerer is able to respond with an
          incomplete set of candidates and continue trickling the rest.
          Half trickle offers will typically contain an
          end-of-candidates indication.
        </t>
        <t>
          The mechanism can be used in cases where there is no way for
          an agent to verify in advance whether a remote party supports
          trickle ice. Because it contains a full set of candidates, its
          first offer can thus be handled by a regular vanilla ICE
          agent, while still allowing a trickle one to use the
          optimisation defined in this specification. This prevents
          negotiation from failing in the former case while still giving
          roughly half the trickle ICE benefits in the latter (hence the
          name of the mechanism).
        </t>
        <t>
          Use of half trickle is only necessary during an initial
          offer/answer exchange. Once both parties have received a
          session description from their peer, they can each reliably
          determine trickle ICE support and use it for all subsequent
          offer/answer exchanges.
        </t>
        <t>
          It is worth pointing out that using half trickle may actually
          bring more than just half the improvement in terms of user
          experience. This can happen in cases where an agent
          starts gathering candidates upon user interface cues that a
          call is pending, such as activity on a keypad or the phone
          going off hook. This would mean a part or all candidate
          harvesting could have completed before the agent actually
          needs to send the offer. Given that the answerer will be able
          to trickle candidates, both agents will be able to start
          connectivity checks and complete ICE processing earlier than
          with vanilla ICE and potentially even as early as with full
          trickle.
        </t>
        <t>
          However, such anticipation is not not always possible. For
          example, a multipurpose user agent or a WebRTC web page where
          communication is a non-central feature (e.g. calling a support
          line in case of a problem with the main features) would not
          necessarily have a way of distinguishing between call
          intentions and other user activity. Still, even in these
          cases, using half trickle would be an improvement over vanilla
          ICE as it would optimize performance for answerers.
        </t>
      </section>
    </section>
    <section title='Sending the Initial Offer' anchor="initial-offer">
      <t>
        An agent starts gathering candidates as soon as it has an
        indication that communication is imminent (e.g. a user interface
        cue or an explicit request to initiate a session). Contrary to
        vanilla ICE, implementations of trickle ICE do not need to
        gather candidates in a blocking manner. Therefore, unless half
        trickle is being used, agents SHOULD generate and transmit their
        initial offer as early as possible, in order to allow the remote
        party to start gathering and trickling candidates.
      </t>
      <t>
        Trickle ICE agents MAY include any set of candidates in an
        offer. This includes the possibility of generating one with no
        candidates, or one that contains all the candidates that the
        agent is planning on using in the following session.
      </t>
      <t>
        For optimal performance, it is RECOMMENDED that an initial offer
        contains host candidates only. This would allow both agents to
        start gathering server reflexive, relayed and other non-host
        candidates simultaneously, and it would also enable them to
        begin connectivity checks.
      </t>
      <t>
        If the privacy implications of revealing host addresses are a
        concern, agents MAY generate an offer that contains no
        candidates and then only trickle candidates that do not reveal
        host addresses (e.g. relayed candidates).
      </t>
      <t>
        Prior to actually sending an initial offer, agents MAY verify if
        the remote party supports trickle ICE, where such mechanisms
        actually exist. If absence of such support is confirmed agents
        MUST fall back to using vanilla ICE or abandon the entire
        session.
      </t>
      <t>
        All trickle ICE offers and answers MUST indicate support of this
        specification, as explained in <xref target="sdp.offer"/>.
      </t>
      <t>
        Calculating priorities and foundations, as well as determining
        redundancy of candidates work the same way they do with vanilla
        ICE.
      </t>
      <section title='Encoding the SDP' anchor="sdp.offer">
        <t>
          The process of encoding the SDP <xref target="RFC4566"/> is
          mostly the same as the one used by vanilla ICE. Still, trickle
          ICE does require a few differences described here.
        </t>
        <t>
          Agents MUST indicate support for Trickle ICE by including the
          "trickle" token for the "a=ice-options" attribute:
          <figure>
            <artwork>
<![CDATA[
    a=ice-options:trickle
]]>
            </artwork>
          </figure>
        </t>
        <t>
          As mentioned earlier in this section, Offers and Answers can
          contain any set of candidates, which means that a trickle ICE
          session description MAY contain no candidates at all. In such
          cases the agent would still need to place an address in the
          "c=" line(s). If the use of a host address there is
          undesirable (e.g. for privacy reasons), the agent MAY set the
          connection address to IP6 ::. In this case it MUST also
          set the port number to 9 (Discard). There is no need to
          include a fictitious candidate for the IP6 :: address when
          doing so.
        </t>
        <t>
          It is worth noting that the use of IP6 :: has been selected
          over IP4 0.0.0.0, even though <xref target="RFC3264"/> already
          gives the latter semantics appropriate for such use. The
          reason for this choice is the historic use of 0.0.0.0 as a
          means of putting a stream on hold <xref target="RFC2543"/> and
          the ambiguity that this may cause with legacy libraries and
          applications.
        </t>
        <t>
          It is also worth mentioning that use of IP6 :: here does not
          constitute any kind of indication as to the actual use of
          IPv6 candidates in a session and it can very well appear in
          a negotiation that only involves IPv4 candidates.
        </t>
      </section>
    </section>
    <section title='Receiving the Initial Offer' >
      <t>
        When an agent receives an initial offer, it will first check if
        it indicates support for trickle ICE as explained in
        <xref target="disco"/>. If this is not the case, the agent MUST
        process the offer according to the <xref target="RFC5245"/>
        procedures or standard <xref target="RFC3264"/> processing in
        case no ICE support is detected at all.
      </t>
      <t>
        It is worth pointing out that in case support for trickle ICE is
        confirmed, an agent will automatically assume support for
        vanilla ICE as well even if the support verification procedure
        in <xref target="RFC5245"/> indicates otherwise. Specifically,
        such verification would indicate lack of support when the offer
        contains no candidates. The IP6 :: address present in the c=
        line in that case would not "appear in a candidate attribute".
        Obviously, a fallback to <xref target="RFC3264"/> is not
        required when this happens.
      </t>
      <t>
        If, the offer does indicate support for trickle ICE, the agent
        will determine its role, start gathering and prioritizing
        candidates and, while doing so it will also respond by sending
        its own answer, so that both agents can start forming check
        lists and begin connectivity checks.
      </t>
      <section title="Sending the Initial Answer">
        <t>
          An agent can respond to an initial offer at any point while
          gathering candidates. The answer can again contain any set of
          candidates including none or all of them. Unless it is
          protecting host addresses for privacy reasons, the agent would
          typically construct this initial answer including only them,
          thus allowing the remote party to also start forming
          checklists and performing connectivity checks.
        </t>
        <t>
          The answer MUST indicate support for trickle ICE as described
          by <xref target="disco"/>.
        </t>
      </section>
      <section title="Forming check lists and beginning connectivity
                      checks" anchor="check.lists">
        <t>
          After exchanging offer and answer, and as soon as they have
          obtained local and remote candidates, agents will begin
          forming candidate pairs, computing their priorities and
          creating check lists according to the vanilla ICE procedures
          described in <xref target="RFC5245"/>. Obviously in order for
          candidate pairing to be possible, it would be necessary that
          both the offer and the answer contained candidates. If this
          was not the case agents will still create the check lists (so
          that their Active/Frozen state could be monitored and updated)
          but they will only populate them once they actually have the
          candidate pairs.
        </t>
        <t>
          Initially, all check lists will have their Active/Frozen state
          set to Frozen.
        </t>
        <t>
          Trickle ICE agents will then inspect the first check list and
          attempt to unfreeze all candidates belonging to the first
          component on the first media stream (i.e. the first media
          stream that was reported to the ICE implementation from the
          using application). If this checklist is still empty however,
          agents will hold off further processing until this is no
          longer the case.
        </t>
        <t>
          Respecting the order in which lists have been reported to an
          ICE implementation, or in other words, the order in which
          they appear in SDP, is crucial to the frozen candidates
          algorithm and important when making sure that connectivity
          checks are performed simultaneously by both agents.
        </t>
      </section>
      <section title='Encoding the SDP' anchor="sdp.answer">
        <t>
          The process for encoding the SDP at the answerer is identical
          to the process followed by the offerer for both full and lite
          implementations, as described in <xref target="sdp.offer"/>.
        </t>
      </section>
    </section>
    <section title="Receiving the Initial Answer">
      <t>
        When receiving an answer, agents will follow vanilla ICE
        procedures to determine their role and they would then
        form check lists (as described in <xref target="check.lists"/>)
        and begin connectivity checks .
      </t>
    </section>
    <section title='Performing Connectivity Checks' >
      <t>
        For the most part, trickle ICE agents perform connectivity
        checks following vanilla ICE procedures. Of course, the
        asynchronous nature of candidate harvesting in trickle ICE would
        impose a number of changes described here.
      </t>
      <section title="Check List and Timer State Updates"
               anchor="state-updates">
        <t>
          The vanilla ICE specification requires that agents update
          check lists and timer states upon completing a connectivity
          check transaction. During such an update vanilla ICE agents
          would set the state of a check list to Failed if the following
          two conditions are satisfied:
        </t>
        <t>
          <list style="symbols">
            <t>
              all of the pairs in the check list are either in the
              Failed or Succeeded state;
            </t>
            <t>
              if at least one of the components of the media stream
              has no pairs in its valid list.
            </t>
          </list>
        </t>
        <t>
          With trickle ICE, the above situation would often occur when
          candidate harvesting and trickling are still in progress and
          it is perfectly possible that future checks will succeed. For
          this reason trickle ICE agents add the following conditions to
          the above list:
        </t>
        <t>
          <list style="symbols">
            <t>
              all candidate harvesters have completed and the agent
              is not expecting to discover any new local candidates;
            </t>
            <t>
              the remote agent has sent an end-of-candidates indication
              for that check list as described in
              <xref target="end-of-candidates"/>.
            </t>
          </list>
        </t>
        <t>
          Vanilla ICE requires that agents then update all other check
          lists, placing one pair in each of them into the Waiting
          state, effectively unfreezing all remaining check lists. Given
          that with trickle ICE, other check lists may still be empty at
          that point, a trickle ICE agent SHOULD also maintain an
          explicit Active/Frozen state for every check list, rather than
          deducing it from the state of the pairs it contains. This
          state should be set to Active when unfreezing the first pair
          in a list or when that couldn't happen because a list was
          empty.
        </t>
      </section>
    </section>
    <section title='Discovering and Sending Additional Local Candidates'
             anchor="send-trickling">
      <t>
        After an offer or an answer have been sent, agents will most
        likely continue discovering new local candidates as STUN, TURN
        and other non-host candidate harvesting mechanisms begin to
        yield results. Whenever an agent discovers such a new candidate
        it will compute its priority, type, foundation and component id
        according to normal vanilla ICE procedures.
      </t>
      <t>
        The new candidate is then checked for redundancy against the
        existing list of local candidates. If its transport address and
        base match those of an existing candidate, it will be considered
        redundant and will be ignored. This would often happen for
        server reflexive candidates that match the host addresses they
        were obtained from (e.g. when the latter are public IPv4
        addresses). Contrary to vanilla ICE, trickle ICE agents will
        consider the new candidate redundant regardless of its priority.
      </t>
      <t>
        Next the client sends (i.e. trickles) the newly learnt
        candidate(s) to the remote agent. The actual delivery of the new
        candidates will be specified by using protocols such as SIP.
        Trickle ICE imposes no restrictions on the way this is done or
        whether it is done at all. For example, some applications may
        choose not to send trickle updates for server reflexive
        candidates and rely on the discovery of peer reflexive ones
        instead.
      </t>
      <t>
        When trickle updates are sent however, each candidate MUST be
        delivered to the receiving Trickle ICE implementation not more
        than once and in the same order that they were sent. In other
        words, if there are any candidate retransmissions, they must
        be hidden from the ICE implementation.
      </t>
      <t>
        Also, candidate trickling needs to be correlated to a specific
        ICE negotiation session, so that if there is an ICE restart, any
        delayed updates for a previous session can be recognized as such
        and ignored by the receiving party.
      </t>
      <t>
        One important aspect of Vanilla ICE is that connectivity checks
        for a specific foundation and component be attempted
        simultaneously by both agents, so that any firewalls or NATs
        fronting the agents would whitelist both endpoints and allow
        all except for the first (suicide) packets to go through. This
        is also crucial to unfreezing candidates in the right time.
      </t>
      <t>
        In order to preserve this feature here, when trickling
        candidates agents MUST respect the order of the components as
        they appear (implicitly or explicitly) in the Offer/Answer
        descriptions. Therefore a candidate for a specific component
        MUST NOT be sent prior to candidates for other components within
        the same foundation.
      </t>
      <t>
        For example, the following session description contains two
        components (RTP and RTCP), and two foundations (host and  the
        server reflexive):
        <figure>
          <artwork>
<![CDATA[
  v=0
  o=jdoe 2890844526 2890842807 IN IP4 10.0.1.1
  s=
  c=IN IP4 10.0.1.1
  t=0 0
  a=ice-pwd:asd88fgpdd777uzjYhagZg
  a=ice-ufrag:8hhY
  m=audio 5000 RTP/AVP 0
  a=rtpmap:0 PCMU/8000
  a=candidate:1 1 UDP 2130706431 10.0.1.1 5000 typ host
  a=candidate:1 2 UDP 2130706431 10.0.1.1 5001 typ host
  a=candidate:2 1 UDP 1694498815 192.0.2.3 5000 typ srflx
      raddr 10.0.1.1 rport 8998
  a=candidate:2 2 UDP 1694498815 192.0.2.3 5001 typ srflx
      raddr 10.0.1.1 rport 8998
]]>
          </artwork>
        </figure>
        For this description the RTCP host candidate MUST NOT be sent
        prior to the RTP host candidate. Similarly the RTP server
        reflexive candidate MUST be sent together with or prior to the
        RTCP server reflexive candidate.
      </t>
      <t>
        Note that the order restriction only applies among candidates
        that belong to the same foundation.
      </t>
      <t>
        It is also equally important to preserve this order across media
        streams and this is covered by the requirement to always start
        unfreezing candidates starting from the first media stream
        <xref target="check.lists"/>.
      </t>
      <t>
        Once the candidate has been sent to the remote party, the agent
        checks if any remote candidates are currently known for this
        same stream. If this is not the case the new candidate will
        simply be added to the list of local candidates.
      </t>
      <t>
        Otherwise, if the agent has already learned of one or more
        remote candidates for this stream and component, it will begin
        pairing the new local candidates with them and adding the pairs
        to the existing check lists according to their priority.
      </t>
      <section title='Pairing newly learned candidates and updating
                      check lists'
                 anchor="cand-pairing">
        <t>
          Forming candidate pairs will work the way it is described by
          the vanilla ICE specification. Actually adding the new pair to
          a check list however, will happen according to the rules
          described below.
        </t>
        <t>
          If the check list where the pair is to be added already
          contains the maximum number of candidate pairs (100 by default
          as per <xref target="RFC5245"/>), the new pair is discarded.
        </t>
        <t>
          If the new pair's local candidate is server reflexive, the
          server reflexive candidate MUST be replaced by its base before
          adding the pair to the list. Once this is done, the agent
          examines the check list looking for another pair that would be
          redundant with the new one. If such a pair exists, the newly
          formed pair is ignored.
        </t>
        <t>
          For all other pairs, including those with a server reflexive
          local candidate that were not found to be redundant:
          <list style="symbols">
            <t>
              if this check list is Frozen then the new pair will
              also be assigned a Frozen state.
            </t>
            <t>
              else if the check list is Active and it is either empty or
              contains only candidates in the Succeeded and Failed
              states, then the new pair's state is set to Waiting.
            </t>
            <t>
              else if the check list is non-empty and Active, then the
              new pair state will be set to
              <list style="hanging">
                <t hangText="Frozen: ">
                  if there is at least one pair in the list whose
                  foundation matches the one in the new pair and whose
                  state is neither Succeeded nor Failed (eventually the
                  new pair will get unfrozen after the the on-going
                  check for the existing pair concludes);
                </t>
                <t hangText="Waiting: ">
                  if the list contains no pairs with the same foundation
                  as the new one, or, in case such pairs exist but they
                  are all in either the Succeeded or Failed states.
                </t>
              </list>
            </t>
          </list>
        </t>
      </section>
      <section title="Encoding the SDP for Additional Candidates">
        <t>
          To facilitate interoperability an ICE agent will encode
          additional candidates using the vanilla ICE SDP syntax. For
          example:
          <figure>
            <artwork>
<![CDATA[
    a=candidate:2 1 UDP 1658497328 198.51.100.33 5000 typ host
]]>
            </artwork>
          </figure>
          Given that such lines do not provide a relationship between
          the candidate and the m line that it relates to, signalling
          protocols using trickle ICE MUST establish that relation
          themselves using an <xref target="RFC3388">MID</xref>. Such
          MIDs use "media stream identification", as defined in
          <xref target="RFC3388"/>, to identify a corresponding m-line.
          When creating candidate lines usages of trickle ICE MUST use
          the MID if possible, or the m-line index if not. Obviously,
          agents MUST NOT send individual candidates prior to generating
          the corresponding SDP session description.
        </t>
        <t>
          The exact means of transporting additional candidates to a
          remote agent is left to the protocols using trickle ICE. It is
          important to note, however, that these candidate exchanges are
          not part of the offer/answer model.
        </t>
      </section>
      <section title='Announcing End of Candidates'
             anchor="end-of-candidates">
      <t>
        Once all candidate harvesters for a specific media stream
        complete, or expire, the agents will generate an
        "end-of-candidates" indication for that stream and send it to
        the remote agent via the signalling channel. Such indications
        are sent in the form of a media-level attribute that has the
        following form: end-of-candidates.
        <figure>
          <artwork>
<![CDATA[
            a=end-of-candidates
]]>
          </artwork>
        </figure>
        The end-of-candidates indications can be sent as part of an
        offer, which would typically be the case with half trickle
        initial offers, they can accompany the last candidate an agent
        can send for a stream, and they can also be sent alone (e.g.
        after STUN Binding requests or TURN Allocate requests to a
        server timeout and the agent has no other active harvesters).
      </t>
      <t>
        Controlled trickle ICE agents SHOULD always send
        end-of-candidates indications once harvesting for a media stream
        has completed unless ICE processing terminates before they've
        had a chance to do so. Sending the indication is necessary in
        order to avoid ambiguities and speed up ICE conclusion. This is
        necessary in order to avoid ambiguities and speed up ICE
        conclusion. Controlling agents on the other hand MAY sometimes
        conclude ICE processing prior to sending end-of-candidates
        notifications for all streams. This would typically be the case
        with aggressive nomination. Yet it is RECOMMENDED that
        controlling agents do send such indications whenever possible
        for the sake of consistency and keeping middle boxes and
        controlled agents up-to-date on the state of ICE processing.
      </t>
      <t>
        When sending end-of-candidates during trickling, rather than as
        a part of an offer or an answer, it is the responsibility of the
        using protocol to define means that can be used to relate the
        indication to one or more specific m-lines.
      </t>
      <t>
        Receiving an end-of-candidates notification allows an agent to
        update check list states and, in case valid pairs do not exist
        for every component in every media stream, determine that ICE
        processing has failed. It also allows agents to speed ICE
        conclusion in cases where a candidate pair has been validates
        but it involves the use of lower-preference transports such as
        TURN. In such situations some implementations may choose to wait
        in case higher-priority candidates are received and
        end-of-candidates provides an indication that this is not going
        to happen.
      </t>
      <t>
        An agent MAY also choose to generate an end-of-candidates
        event before candidate harvesting has actually completed, if the
        agent determines that harvesting has continued for more than an
        acceptable period of time. However, an agent MUST NOT send any
        more candidates after it has send an end-of-candidates
        notification.
      </t>
      <t>
        When performing half trickle agents SHOULD send
        end-of-candidates together with their initial offer unless they
        are planning on potentially sending additional candidates in
        case the remote party turns out to actually support trickle ICE.
      </t>
      <t>
        When end-of-candidates is sent as part of an offer or an answer
        it can appear as a session-level attribute, which would be
        equivalent to having it appear in all m-lines.
      </t>
      <t>
        Once an agent sends the end-of-candidates event, it will
        update the state of the corresponding check list as explained
        in section <xref target="state-updates"/>. Past that point
        agents MUST NOT send any new candidates. Once an agent has
        received an end-of-candidates indication, it MUST also ignore
        any newly received candidates for that media stream. Adding new
        candidates to the negotiation is hence only possible through an
        ICE restart.
      </t>
      <t>
        It is important to note that This specification does not
        override vanilla ICE semantics for concluding ICE processing.
        This means that even if end-of-candidates indications are sent
        agents will still have to go through pair nomination. Also, if
        pairs have been nominated for components and media streams, ICE
        processing will still conclude even if end-of-candidate
        indications have not been received for all streams.
      </t>
      </section>
    </section>
    <section title='Receiving Additional Remote Candidates'
             anchor="recv-trickling">
      <t>
        At any point of ICE processing, a trickle ICE agent may receive
        new candidates from the remote agent. When this happens and no
        local candidates are currently known for this same stream, the
        new remote candidates are simply added to the list of remote
        candidates.
      </t>
      <t>
        Otherwise, the new candidates are used for forming candidate
        pairs with the pool of local candidates and they are added to
        the local check lists as described in
        <xref target="cand-pairing"/>.
      </t>
      <t>
        Once the remote agent has completed candidate harvesting, it
        will send an end-of-candidates event. Upon receiving such an
        event, the local agent MUST update check list states as per
        <xref target="state-updates"/>. This may lead to some check
        lists being marked as Failed.
      </t>
    </section>
    <section title='Receiving an End Of Candidates Notification'
             anchor="end-of-candidates.recv">
      <t>
        When an agent receives an end-of-candidates notification
        for a specific check list, they will update its state as per
        <xref target="state-updates"/>. In case the list is still in
        the Active state after the update, the agent will persist the
        the fact that an end-of-candidates notification has been
        received for and take it into account in future list updates.
      </t>
    </section>
    <section title="Trickle ICE and Peer Reflexive Candidates">
      <t>
        Even though Trickle ICE does not explicitly modify the
        procedures for handling peer reflexive candidates, their
        processing could be impacted in implementations. With Trickle
        ICE, it is possible that server reflexive candidates be
        discovered as peer reflexive in cases where incoming
        connectivity checks are received from these candidates before
        the trickle updates that carry them.
      </t>
      <t>
        While this would certainly increase the number of cases where
        ICE processing nominates and selects candidates discovered as
        peer-reflexive it does not require any change in processing.
      </t>
      <t>
        It is also likely that, some applications would prefer not to
        trickle server reflexive candidates to entities that are known
        to be publicly accessible and where sending a direct STUN
        binding request is likely to reach the destination faster than
        the trickle update that travels through the signalling path.
      </t>
    </section>
    <section title='Concluding ICE Processing'
                 anchor="concluding.ice">
      <t>
        This specification does not directly modify the procedures
        ending ICE processing described in Section 8 of
        <xref target="RFC5245"/>, and trickle ICE implementations will
        follow the same rules.
      </t>
      <t>

      </t>
    </section>
    <section title='Subsequent Offer/Answer Exchanges'
             anchor="subsequent.oa">
      <t>
        Either agent MAY generate a subsequent offer at any time allowed
        by <xref target="RFC3264"/>. When this happens agents will use
        <xref target="RFC5245"/> semantics to determine whether or not
        the new offer requires an ICE restart. If this is the case then
        agents would perform trickle ICE as they would in an initial
        offer/answer exchange.
      </t>
      <t>
        The only differences between an ICE restart and a brand new
        media session are that:
      </t>
      <t>
        <list style='symbols'>
          <t>
            during the restart, media can continue to be sent to the
            previously validated pair.
          </t>
          <t>
            both agents are already aware whether or not their peer
            supports trickle ICE, and there is no longer need for
            performing half trickle or confirming support with other
            mechanisms.
          </t>
        </list>
      </t>
    </section>
    <section title='Interaction with ICE Lite'>
      <t>
        Behaviour of Trickle ICE capable ICE lite agents does not
        require any particular rules other than those already defined
        in this specification and <xref target="RFC5245"/>. This section
        is hence added with an informational purpose only.
      </t>
      <t>
        A Trickle ICE capable ICE Lite agent would generate offers or
        answers as per <xref target="RFC5245"/>. Both will indicate
        support for trickle ICE (<xref target="sdp.offer"/>) and given
        that they will contain a complete set of candidates (the agent's
        host candidates) these offers and answers would also be
        accompanied with an end-of-candidates notification.
      </t>
      <t>
        When performing full trickle, a full ICE implementation could
        send an offer or an answer with no candidates and an IP6 ::
        connection line address. After receiving an answer that
        identifies the remote agent as an ICE lite implementation, the
        offerer may very well choose to not send any additional
        candidates. The same is also true in the case when the ICE lite
        agent is making the offer and the full ICE one is answering. In
        these cases the connectivity checks would be enough for the ICE
        lite implementation to discover all potentially useful
        candidates as peer reflexive. The following example illustrates
        one such ICE session:
      </t>
      <figure title="Example " anchor="fig-ice-lite">
        <artwork>
<![CDATA[
        ICE Lite                                          Bob
         Agent
           |   Offer (a=ice-lite a=ice-options:trickle)    |
           |---------------------------------------------->|
           |                                               |no cand
           |         Answer (a=ice-options:trickle)        |trickling
           |<----------------------------------------------|
           |              Connectivity Checks              |
           |<--------------------------------------------->|
  peer rflx|                                               |
 cand disco|                                               |
           |                                               |
           |<=============== MEDIA FLOWS =================>|

]]>
        </artwork>
      </figure>
      <t>
        In addition to reducing signaling traffic this approach also
        removes the need to discover STUN bindings, or to make TURN or
        UPnP allocations which may considerably lighten ICE processing.
      </t>
    </section>
    <section title='Example Flow'>
      <t>
        A typical successful trickle ICE exchange with an Offer/Answer
        protocol would look this way:
      </t>
      <figure title="Example " anchor="fig-example">
        <artwork>
<![CDATA[
        Alice                                            Bob
          |                     Offer                     |
          |---------------------------------------------->|
          |            Additional Candidates              |
          |---------------------------------------------->|
          |                                               |
          |                     Answer                    |
          |<----------------------------------------------|
          |            Additional Candidates              |
          |<----------------------------------------------|
          |                                               |
          | Additional Candidates and Connectivity Checks |
          |<--------------------------------------------->|
          |                                               |
          |<=============== MEDIA FLOWS =================>|

]]>
        </artwork>
      </figure>
    </section>
    <section title='Security Considerations'>
      <t>
        This specification inherits most of its semantics from
        <xref target="RFC5245"/> and as a result all security
        considerations described there remain the same.
      </t>
    </section>
    <section title='Acknowledgements'>
      <t>
        The authors would like to thank Bernard Aboba, Christer
        Holmberg, Dale R. Worley, Enrico Marocco, Flemming Andreasen,
        Jonathan Lennox and Martin Thomson for their reviews and
        suggestions on improving this document.
      </t>
    </section>
  </middle>
  <back>
    <references title='Normative References'>
      <?rfc include="reference.RFC.2119"?>
      <?rfc include="reference.RFC.3264"?>
      <?rfc include="reference.RFC.4566"?>
      <?rfc include="reference.RFC.5245"?>

    </references>
    <references title='Informative References'>
      <?rfc include="reference.RFC.1918"?>
      <?rfc include="reference.RFC.2543"?>
      <?rfc include="reference.RFC.3261"?>
      <?rfc include="reference.RFC.3388"?>
      <?rfc include="reference.RFC.3840"?>
      <?rfc include="reference.RFC.4787"?>
      <?rfc include="reference.RFC.5389"?>
      <?rfc include="reference.RFC.5766"?>
      <?rfc include="reference.I-D.keranen-mmusic-ice-address-selection"?>
      <?rfc include="reference.I-D.ivov-mmusic-trickle-ice-sip"?>
      <reference anchor="XEP-0176">
        <front>
          <title>XEP-0176: Jingle ICE-UDP Transport Method</title>
          <author initials='J.' surname='Beda' fullname='Joe Beda'>
                  <organization abbrev='Google'>Google</organization>
          </author>
          <author initials='S.' surname='Ludwig'
                  fullname='Scott Ludwig'>
            <organization abbrev='Google'>Google</organization>
          </author>
          <author initials='P.' surname='Saint-Andre'
                  fullname='Peter Saint-Andre'>
            <organization abbrev='Cisco'>Cisco</organization>
          </author>
          <author initials='J.' surname='Hildebrand'
                  fullname='Joe Hildebrand'>
            <organization abbrev='Cisco'>Cisco</organization>
          </author>
          <author initials='S.' surname='Egan' fullname='Sean Egan'>
            <organization abbrev='Google'>Google </organization>
          </author>
          <author initials='R.' surname='McQueen'
                      fullname='Robert McQueen'>
            <organization abbrev='Collabora'>Collabora</organization>
          </author>
          <date month="June" year="2009" />
        </front>
        <seriesInfo name="XEP" value="XEP-0176" />
      </reference>
      <reference anchor="XEP-0030">
        <front>
          <title>XEP-0030: Service Discovery</title>
          <author initials='J.' surname='Hildebrand'
                  fullname='Joe Hildebrand'>
            <organization abbrev='Cisco'>Cisco</organization>
          </author>
          <author initials='P.' surname='Millard'
                  fullname='Peter Millard'>
          </author>
          <author initials='R.' surname='Eatmon'
                  fullname='Ryan Eatmon'>
          </author>
          <author initials='P.' surname='Saint-Andre'
                  fullname='Peter Saint-Andre'>
            <organization abbrev='Cisco'>Cisco</organization>
          </author>
          <date month="June" year="2008" />
        </front>
        <seriesInfo name="XEP" value="XEP-0030" />
      </reference>
      <reference anchor="XEP-0115">
        <front>
          <title>XEP-0115: Entity Capabilities</title>
          <author initials='J.' surname='Hildebrand'
                  fullname='Joe Hildebrand'>
            <organization abbrev='Cisco'>Cisco</organization>
          </author>
          <author initials='P.' surname='Saint-Andre'
                  fullname='Peter Saint-Andre'>
            <organization abbrev='Cisco'>Cisco</organization>
          </author>
          <author initials='R.' surname='Tronçon'
                  fullname='Remko Tronçon'>
            <organization abbrev='Synopsys'>Synopsys</organization>
          </author>
          <author initials='J.' surname='Konieczny'
                  fullname='Jacek Konieczny'>
          </author>
          <date month="February" year="2008" />
        </front>
        <seriesInfo name="XEP" value="XEP-0115" />
      </reference>
      <reference anchor="XEP-0278">
        <front>
          <title>XEP-0278: Jingle Relay Nodes</title>
          <author initials='T.' surname='Camargo'
                  fullname='T. Camargo'>
          </author>
          <date month="June" year="2011" />
        </front>
        <seriesInfo name="XEP" value="XEP-0278" />
      </reference>
    </references>
    <section title='Open issues'>
      <t>
        At the time of writing of this document the authors have no
        clear view on how and if the following list of issues should
        be addressed.
      </t>
      <section title="MID/Stream Indices in SDP">
        <t>
          This specification does not currently define syntax for
          candidate-to-stream bindings although it says that they should
          be implemented with MID or a stream index. Yet, it is
          reasonable to assume that most usages would need to do this
          within the SDP and it may make sense to agree on the format.
          Here's one possible way to do this:
          <figure>
            <artwork>
<![CDATA[
      a=mid:1
      a=candidate:1 1 UDP 1658497328 192.168.100.33 5000 typ host
      a=candidate:2 1 UDP 1658497328 96.1.2.3 5000 typ srflx
      a=mid:2
      a=candidate:2 1 UDP 1658497328 96.1.2.3 5002 typ srflx
      a=end-of-candidates
]]>
            </artwork>
          </figure>
        </t>
      </section>
      <section title='Starting checks'>
        <t>
          Normally Vanilla ICE implementations would first activate a
          check list, validate at least one pair in every component
          and only then unfreeze all other checklists. With trickle ICE
          this would be suboptimal since, candidates can arrive randomly
          and we would be wasting time waiting for a checklist to fill
          (almost as if we were doing vanilla ICE). We need to decide if
          unfreezing everything solely based on foundation is good
          enough.
        </t>
      </section>
    </section>
    <section title='Changes From Earlier Versions'>
      <t>
        Note to the RFC-Editor: please remove this section prior to
        publication as an RFC.
      </t>
      <section title='Changes From draft-ivov-01 and draft-mmusic-00'>
        <t>
          <list style='symbols'>
            <t>
              Added a requirement to trickle candidates by order of
              components to avoid deadlocks in the unfreezing algorithm.
            </t>
            <t>
              Added an informative note on peer-reflexive candidates
              explaining that nothing changes for them semantically but
              they do become a more likely occurrence for Trickle ICE.
            </t>
            <t>
              Limit the number of pairs to 100 to comply with 5245.
            </t>
            <t>
              Added clarifications on the non-importance of how newly
              discovered candidates are trickled/sent to the remote
              party or if this is done at all.
            </t>
            <t>
              Added transport expectations for trickled candidates
              as per Dale Worley's recommendation.
            </t>
          </list>
        </t>
      </section>
      <section title='Changes From draft-ivov-00'>
        <t>
          <list style='symbols'>
            <t>
              Specified that end-of-candidates is a media level
              attribute which can of course appear as session level,
              which is equivalent to having it appear in all m-lines.
              Also made end-of-candidates optional for cases such as
              aggressive nomination for controlled agents.
            </t>
            <t>
              Added an example for ICE lite and trickle ICE to
              illustrate how, when talking to an ICE lite agent doesn't
              need to send or even discover any candidates.
            </t>
            <t>
              Added an example for ICE lite and trickle ICE to
              illustrate how, when talking to an ICE lite agent doesn't
              need to send or even discover any candidates.
            </t>
            <t>
              Added wording that explicitly states ICE lite agents
              have to be prepared to receive no candidates over
              signalling and that they should not freak out if this
              happens. (Closed the corresponding open issue).
            </t>
            <t>
              It is now mandatory to use MID when trickling candidates
              and using m-line indexes is no longer allowed.
            </t>
            <t>
              Replaced use of 0.0.0.0 to IP6 :: in order to avoid
              potential issues with RFC2543 SDP libraries that interpret
              0.0.0.0 as an on-hold operation. Also changed the port
              number here from 1 to 9 since it already has a more
              appropriate meaning. (Port change suggested by Jonathan
              Lennox).
            </t>
            <t>
              Closed the Open Issue about use about what to do with
              cands received after end-of-cands. Solution: ignore, do
              an ice restart if you want to add something.
            </t>
            <t>
              Added more terminology, including trickling, trickled
              candidates, half trickle, full trickle,
            </t>
            <t>
              Added a reference to the SIP usage for trickle ICE as
              requested at the Boston interim.
            </t>
          </list>
        </t>
      </section>
      <section title='Changes From draft-rescorla-01'>
        <t>
          <list style='symbols'>
            <t>
              Brought back explicit use of Offer/Answer. There are no
              more attempts to try to do this in an O/A independent way.
              Also removed the use of ICE Descriptions.
            </t>
            <t>
              Added SDP specification for trickled candidates, the
              trickle option and 0.0.0.0 addresses in m-lines, and
              end-of-candidates.
            </t>
            <t>
              Support and Discovery. Changed that section to be less
              abstract. As discussed in IETF85, the draft now says
              implementations and usages need to either determine
              support in advance and directly use trickle, or do
              half trickle. Removed suggestion about use of discovery in
              SIP or about letting implementing protocols do what they
              want.
            </t>
            <t>
              Defined Half Trickle. Added a section that says how it
              works. Mentioned that it only needs to happen in the first
              o/a (not necessary in updates), and added Jonathan's
              comment about how it could, in some cases, offer more than
              half the improvement if you can pre-gather part or all of
              your candidates before the user actually presses the call
              button.
            </t>
            <t>
              Added a short section about subsequent offer/answer
              exchanges.
            </t>
            <t>
              Added a short section about interactions with ICE Lite
              implementations.
            </t>
            <t>
              Added two new entries to the open issues section.
            </t>
          </list>
        </t>
      </section>
      <section title='Changes From draft-rescorla-00'>
        <t>
          <list style='symbols'>
            <t>
              Relaxed requirements about verifying support following
              a discussion on MMUSIC.
            </t>
            <t>
              Introduced ICE descriptions in order to remove ambiguous
              use of 3264 language and inappropriate references to
              offers and answers.
            </t>
            <t>
              Removed inappropriate assumption of adoption by RTCWEB
              pointed out by Martin Thomson.
            </t>
          </list>
        </t>
      </section>
    </section>
  </back>
</rfc>