Jingle ICE-UDP Transport Method

%ents; ICE-CORE"> ]>

Jingle ICE-UDP Transport Method This document defines a Jingle transport method that results in sending data between two XMPP entities via the User Datagram Protocol (UDP) as negotiated using the Interactive Connectivity Establishment (ICE) methodology. &LEGALNOTICE; 0176 Proposed Standards Track Standards Council XMPP Core XEP-0166 TO BE ASSIGNED &joebeda; &scottlu; &stpeter; &hildjj; &seanegan; 0.12 2007-11-28 psa

Moved ice-tcp definition to a separate specification.

0.11 2007-11-27 psa

Further editorial review; also added sections on modification of existing candidates and exchange of subsequent candidates.

0.10 2007-11-15 psa

Editorial review and consistency check.

0.9 2007-06-28 psa

Updated to track ICE-16.

0.8 2007-04-17 psa

Separately defined ice-tcp and ice-udp transport methods to enable clearer definition of transport methods and reuse by application types; specified Jingle conformance, including definition of ice-udp as lossy and ice-tcp as reliable.

0.7 2007-03-23 psa

Updated to track ICE-14 and ICE-TCP-03; moved text on discovery of STUN servers to separate specification.

0.6 2006-12-21 psa

Modified spec to use provisional namespace before advancement to Draft (per XEP-0053).

0.5 2006-10-31 psa

Updated to track ICE-12; corrected service discovery process; completed editorial review; removed mention of DTMF, which is for audio only.

0.4 2006-09-13 psa

Updated to track ICE-10; added section on service discovery.

0.3 2006-07-12 se/psa

Specified that DTMF must use in-band signalling (XEP-0181).

0.2 2006-03-24 psa

Recommended use of RTP-native methods for DTMF.

0.1 2006-03-01 psa/jb

Initial version (split from XEP-0166).

&xep0166; defines a framework for negotiating and managing out-of-band data sessions over XMPP. In order to provide a flexible framework, the base Jingle specification defines neither data transport methods nor application formats, leaving that up to separate specifications.

The current document defines a transport method for establishing and managing data exchanges between XMPP entities over the User Datagram Protocol (see &rfc0768;), using the ICE methodology developed within the IETF and specified in &ice; (hereafter referred to as &icecore;). Use of the ice-udp method results in a lossy transport suitable for media applications where some packet loss is tolerable (e.g., audio and video).

Note: &icecore; has been approved for publication as an RFC but has not yet been published as an RFC. While every effort has been made to keep this document synchronized with &icecore;, the interested reader is referred to &icecore; for a detailed description of the ICE methodology.

The process for ICE negotiation is largely the same in Jingle as it is in ICE. There are several differences:

Instead of using SIP as the signalling channel, Jingle uses XMPP as the signalling channel.
In Jingle, each candidate transport is sent in a separate IQ exchange (rather than sending all candidates at once as in &icecore;). This approach takes advantage of the request-response semantics of the XMPP &IQ; stanza type and enables the parties to send higher-priority candidates earlier in the negotiation, but implies that Jingle is not exactly an offer-answer protocol as specified in RFC 3264.
Syntax from the Session Description Protocol (see &rfc4566;) is mapped to an XML syntax suitable for sending over the XMPP signalling channel.
ICE candidates can be upgraded during a session (e.g., to change an IP address).
Either party can continue to send ICE candidates throughout a session and renegotiate which candidate will be used.

The reader is referred to &icecore; for a description of various terms used in the context of ICE. Those terms are not reproduced here.

The Jingle transport method defined herein are designed to meet the following requirements:

Make it possible to establish and manage out-of-band connections between two XMPP entities, even if they are behind Network Address Translators (NATs) or firewalls.
Enable use of UDP as the transport protocol itself.
Make it relatively easy to implement support in standard Jabber/XMPP clients.
Where communication with non-XMPP entities is needed, push as much complexity as possible onto server-side gateways between the XMPP network and the non-XMPP network.

In accordance with Section 8 of XEP-0166, this document specifies the following information related to the Jingle ice-udp transport method:

The transport negotiation process is defined in the Protocol Description section of this document.
The semantics of the &TRANSPORT; element are defined in the ICE Negotiation section of this document.
Successful negotiation of the ice-udp method results in use of a lossy transport that is suitable for applications where some packet loss is tolerable, such as audio and video.
If multiple components are to be communicated over the transport in the context of the Real-time Transport Protocol (RTP; see &rfc3550;), the component numbered "1" shall be associated with RTP and the component numbered "2" shall be associated with the Real Time Control Protocol (RTCP).

In order for the initiator in a Jingle exchange to start the negotiation, it MUST send a Jingle "session-initiate" stanza as described in XEP-0166. A content type MUST include one transport method. If the initiator wishes to negotiate the ice-udp transport method for an application format, it MUST include an empty &TRANSPORT; child element qualified by the 'http://www.xmpp.org/extensions/xep-0176.html#ns' namespace &NSNOTE;.

[ ... ] ]]>

As described in XEP-0166, to acknowledge receipt of the session initiation request, the responder returns an IQ-result:

]]>

Once the responder acknowledges receipt of the session initiation request as shown above, both initiator and responder MUST immediately negotiate connectivity over the ICE transport by exchanging XML-formatted candidate transports for the channel. This negotiation proceeds immediately in order to maximize the possibility that media can be exchanged as quickly as possible. Concurrent with negotiation of the ICE candidates, it is possible for the initiator and responder to negotiate which content types the session will include, which transport methods will be tried for each content type, etc. Those negotiation flows are shown in other specifications, such as XEP-0166. This document specifies only negotiation of the ICE transport method.

Note: In order to expedite session establishment, the initiator MAY send transport candidates immediately after sending the "session-initiate" message and before receiving acknowledgement from the responder (i.e., the initiator MUST consider the session to be live even before receiving acknowledgement). Given in-order delivery, the responder should receive such "transport-info" messages after receiving the "session-initiate" message; if not, it is appropriate for the responder to return <unknown-session/> errors since it according to its state machine the session does not exist. If either party receives an <unknown-session/> from the other party, it MUST terminate the negotiation and the session.

The candidate syntax and negotiation flow are described below.

The following is an example of the candidate format:

]]>

The attributes of the <candidate/> element are described in the following table:

Name	Description	SDP Syntax	Example
component	A Component ID as defined in &icecore;.	Component ID value in a=candidate line	1
foundation	A Foundation as defined in &icecore;.	Foundation value in a=candidate line	1
generation	An index, starting at 0, that enables the parties to keep track of updates to the candidate throughout the life of the session.	N/A	0
ip	The Internet Protocol (IP) address for the candidate transport mechanism; this may be either an IPv4 address or an IPv6 address.	IP Address value in a=candidate line	10.0.1.1
network	An index, starting at 0, referencing which network this candidate is on for a given peer (used for diagnostic purposes if the calling hardware has more than one Network Interface Card).	N/A	0
port	The port at the candidate IP address.	Port value in a=candidate line	8998
priority	A Priority as defined in &icecore; In accordance with the rules specified in Section 4.1.1 of &icecore;, the priority values shown in the examples within this document have been calculated as follows. The "type preference" for host candidates is stipulated to be "126" and for server reflexive candidates "100". The "local preference" for network 0 is stipulated to be "4096", for network 1 "2048", and for network 2 "1024".	Priority value in a=candidate line	1678246398
protocol	The protocol to be used. The only allowable value is "udp".	Transport protocol field in a=candidate line	udp
pwd	A Password as defined in &icecore;.	a=ice-pwd line	asd88fgpdd777uzjYhagZg
type	A Candidate Type as defined in &icecore;. The allowable values are "host" for host candidates, "prflx" for peer reflexive candidates, "relay" for relayed candidates, and "srflx" for server reflexive candidates.	Typ field in a=candidate line	srflx
ufrag	A User Fragment as defined in &icecore;.	a=ice-ufrag line	8hhy

The first step in negotiating connectivity is for both parties to immediately begin sending transport candidates methods to the other client. The fact that both parties send candidates means that Jingle requires each party to be a full implementation of ICE, not a "lite" implementation as specified in &icecore;. These candidates SHOULD be gathered by following the procedure specified in Section 4.1.1 of &icecore; and prioritized by following the procedure specified in Section 4.1.2 of &icecore;. Each candidate MUST be sent in a &JINGLE; element with an action of "transport-info".

If the responder receives and can successfully process a given candidate, it returns an IQ-result (if not, for example because the candidate data is improperly formatted, it returns an error). Note: The responder is only indicating receipt of the candidate, not telling the initiator that the candidate will be used.

The initiator keeps sending candidates, one after the other (without stopping to receive an acknowledgement of receipt from the responder for each candidate) until it has exhausted its supply of possible or desirable candidate transports. (Because certain candidates may be more "expensive" in terms of bandwidth or processing power, the initiator may not want to advertise their existence unless necessary.) For each candidate, the responder acknowledges receipt.

At the same time (i.e., immediately after acknowledging receipt of the session-initiate request, not waiting for the initiator to begin or finish sending candidates), the responder also begins sending potential candidates, in order of desirability according to the responder. As above, the initiator acknowledges receipt of the candidates.

As the initiator and responder receive candidates, they probe the various candidate transports for connectivity. In performing these connectivity checks, a client SHOULD follow the procedure specified in Section 7 of &icecore;.

]]> ]]> ]]>

For each candidate received, the other party MUST acknowledge receipt or return an error:

]]>

If, based on STUN connectivity checks (see &rfc3489; and &rfc3489bis;), the responder determines that it will be able to establish a connection using a given candidate, it sends a &JINGLE; element with an action of 'content-accept' (or 'session-accept') to the initiator, specifying the candidate that succeeded:

[ ... ] ]]>

The &JINGLE; element in the content-accept or session-accept stanza SHOULD possess a 'responder' attribute that explicitly specifies the full JID of the responding entity. If the 'responder' attribute is provided, all future commmunications SHOULD be sent to the JID provided in the 'responder' attribute.

If the initiator can also send data over that candidate, then it acknowledges the responder's acceptance:

]]>

Now the initiator and responder can begin sending data over the negotiated connection.

If a candidate succeeded for the responder but the initiator cannot send data over that candidate, it MUST return a ¬acceptable; error in response to the responder's acceptance of the successful candidate:

]]>

If the responder cannot find a suitable candidate transport or it receives a ¬acceptable; error from the initiator in response to its acceptance of a suitable transport, it SHOULD terminate the session as described in Section 6.8 of XEP-0166.

The creator of a content type MAY modify an existing, in-use candidate at any time during the session, for example to change the IP address or port. This is done by sending a content-modify action with the changed candidate information, where the value of the 'generation' is incremented to specify that the candidate information is a modification to an existing candidate.

An example follows (change to IP address and port).

]]>

The recipient then acknowledges receipt.

]]>

If the modification is acceptable, the recipient then sends a content-accept action.

[ ... ] ]]>

The initiator then acknowledges the responder's acceptance:

]]>

The parties then use the modified candidate in subsequent communications.

Even after content acceptance or session acceptance, either party MAY continue to send additional candidates to the other party (e.g., because the user agent has become aware of a new media proxy or NIC). As above, such candidates are shared by sending a transport-info action.

]]>

The receiving party MUST acknowledge receipt of the candidate.

]]>

The receiving party SHOULD check the newly-offered candidate for connectivity, as above. If the candidate is acceptable, the receiving party shall send a content-accept action.

[ ... ] ]]>

The other party then acknowledges the content-accept.

]]>

The parties then use the new candidate in subsequent communications.

If an entity supports the Jingle ice-udp transport, it MUST return a feature of "http://www.xmpp.org/extensions/xep-0176.html#ns" &NSNOTE; in response to &xep0030; information requests.

]]> ... ... ]]>

Naturally, support MAY also be determined via the dynamic, presence-based profile of Service Discovery defined in &xep0115;.

In order to speed the negotiation process so that media can flow as quickly as possible, the initiatior should gather and prioritize candidates in advance or as soon as the principal begins the process of initiating a session.

This specification applies exclusively to Jabber/XMPP clients and places no additional requirements on Jabber/XMPP servers. However, service administrators may wish to deploy a STUN server in order to ease the client-to-client negotiation process. See &xep0215; for related information.

In order to secure the data stream that is negotiated via the Jingle ICE transport, implementations SHOULD use encryption methods appropriate to the transport method and media being exchanged (for details regarding audio and video exchanges via RTP, refer to XEP-0167 and XEP-0180).

This document requires no interaction with &IANA;.

Until this specification advances to a status of Draft, its associated namespaces shall be:

http://www.xmpp.org/extensions/xep-0176.html#ns

Upon advancement of this specification, the ®ISTRAR; shall issue permanent namespaces in accordance with the process defined in Section 4 of &xep0053;.

The following namespaces are requested, and are thought to be unique per the XMPP Registrar's requirements:

urn:xmpp:jingle:transport:ice-udp

The XMPP Registrar shall include "ice-udp" in its registry of Jingle transport methods. The registry submission is as follows:


  ice-udp
  
    A method for negotiation of out-of-band UDP connections with built-in NAT
    and firewall traversal, equivalent to the IETF's Interactive Connectivity
    Establishment (ICE) methodology when resulting in the use of UDP as the
    transport protocol.
  
  lossy
  XEP-0176

    ]]>

]]>