Session Initiation Protocol (SIP) Recording Metadata

This section describes each class of the metadata model and the attributes of each class. This section also describes how different classes are linked and the XML element for each of them.

Each instance of an RS class, namely the RS object, represents a SIP session created between an SRC and SRS for the purpose of recording a CS. The RS object is represented in the XML schema using the 'recording' element, which in turn relies on the SIP/SDP session with which the XML document is associated to provide the attributes of the RS element.

An RS class has the following attributes: start-time - Represents the start time of an RS object. end-time - Represents the end time of an RS object. 'start-time' and 'end-time' attribute values are derivable from the Date header (if present in the SIP message) in the RS. In cases where the Date header is not present, 'start-time' is derivable from the time at which the SRS receives the notification of the SIP message to set up the RS, and 'end-time' is derivable from the time at which the SRS receives a disconnect on the RS SIP dialog.

Each instance of an RS has: Zero or more instances of CS-Groups. Zero or more instances of CS objects. Zero or more instances of MediaStream objects. Zero instances of CSs and CS-Groups in a 'recording' element are allowed to accommodate persistent recording scenarios. A persistent RS is a SIP dialog that is set up between the SRC and the SRS, even before any CS is set up. The metadata sent from the SRC to the SRS when the persistent RS SIP dialog is set up may not have any CS (and the related CS-Group) elements in the XML, as there may not be a session that is associated to the RS yet. For example, a phone acting as an SRC can set up an RS with the SRS, possibly even before the phone is part of a CS. Once the phone joins a CS, the same RS would be used to convey the CS metadata.

One instance of a CS-Group class, namely the CS-Group object, provides association or grouping of all related CSs. For example, in a contact center flow, a call can get transferred to multiple agents. Each of these can trigger the setup of a new CS. In cases where the SRC knows the related CSs, it can group them using the CS-Group element. The CS-Group object is represented in the XML schema using the 'group' element.

A CS-Group has the following attributes: group_id - This attribute groups different CSs that are related. The SRC (or the SRS) is responsible for ensuring the uniqueness of 'group_id' in cases where multiple SRCs interact with the same SRS. The mechanism by which the SRC groups the CS is outside the scope of this document. associate-time - This is the time when a grouping is formed. The rules that determine how a grouping of different CS objects is done by the SRC are outside the scope of this document. disassociate-time - 'disassociate-time' for the CS-Group is calculated by the SRC as the time when the grouping ends.

The linkages between a CS-Group class and other classes are associations. A CS-Group is associated with the RS and CS in the following manner: There are one or more RS objects per CS-Group. Each CS-Group object has to be associated with one or more RSs. Here, each RS can be set up by the potentially different SRCs. There are one or more CSs per CS-Group (for example, in cases where the call is transferred). A CS cannot be associated with more than one CS-Group.

A CS class and its object in the metadata model represent the CS and its properties as seen by the SRC. The CS object is represented in the XML schema using the 'session' element.

A CS class has the following attributes: session_id - This attribute is used to uniquely identify an instance of a CS object, namely the 'session' XML element within the metadata XML document. 'session_id' is generated using the rules mentioned in . reason - This represents the reason why a CS was terminated. The value for this attribute is derived from the SIP Reason header of the CS. There MAY be multiple instances of the 'reason' XML element inside a 'session' element. The 'reason' XML element has 'protocol' as an attribute, which indicates the protocol from which the reason string is derived. The default value for the 'protocol' attribute is "SIP". The 'reason' element can be derived from a SIP Reason header in the CS. sipSessionID - This attribute carries a SIP Session-ID as defined in . Each CS object can have zero or more 'sipSessionID' elements. More than one 'sipSessionID' attribute may be present in a CS. For example, if three participants -- A, B, and C -- are in a conference that has a focus acting as an SRC, the metadata sent from the SRC to the SRS will likely have three 'sipSessionID' elements that correspond to the SIP dialogs that the focus has with each of the three participants. group-ref - A 'group-ref' attribute MAY be present to indicate the group (identified by 'group_id') to which the enclosing session belongs. start-time - This optional attribute represents the start time of the CS as seen by the SRC. stop-time - This optional attribute represents the stop time of the CS as seen by the SRC. This document does not specify attributes relating to what should happen to a recording of a CS after it has been delivered to the SRS (e.g., how long to retain the recording, what access controls to apply). The SRS is assumed to behave in accordance with its local policy. The ability of the SRC to influence this policy is outside the scope of this document. However, if there are implementations where the SRC desires to specify its own policy preferences, this information could be sent as extension data attached to the CS.

A CS is linked to the CS-Group, participant, MediaStream (MS), and RS classes by using the association relationship. The association between the CS and the participant allows the following: A CS will have zero or more participants. A participant is associated with zero or more CSs. This includes participants who are not directly part of any CS. An example of such a case is participants in a pre&nbhy;mixed media stream. The SRC may have knowledge of such participants but not have any signaling relationship with them. This might arise if one participant in a CS is a conference focus. To summarize, even if the SRC does not have direct signaling relationships with all participants in a CS, it should nevertheless create a participant object for each participant that it knows about. The model also allows participants in a CS that are not participants in the media. An example is the identity of a Third Party Call Control (3pcc) that has initiated a CS to two or more participants in the CS. Another example is the identity of a conference focus. Of course, a focus is probably in the media, but since it may only be there as a mixer, it may not report itself as a participant in any of the media streams. The association between the CS and the media stream allows the following: A CS will have zero or more streams. A stream can be associated with at most one CS. A stream in a persistent RS is not required to be associated with any CS before the CS is created, and hence the zero association is allowed. The association between the CS and the RS allows the following: Each instance of an RS has zero or more instances of CS objects. Each CS has to be associated with one or more RSs. Each RS can be potentially set up by different SRCs.

The CS-RS Association class describes the association of a CS to an RS for a period of time. A single CS may be associated with different RSs (perhaps by different SRCs) and may be associated and dissociated several times. The CS-RS Association class is represented in XML using the 'sessionrecordingassoc' XML element.

The CS-RS Association class has the following attributes: associate-time - associate-time is calculated by the SRC as the time it sees a CS associated to an RS. disassociate-time - disassociate-time is calculated by the SRC as the time it sees a CS disassociate from an RS. session_id - Each instance of this class MUST have a 'session_id' attribute that identifies the CS to which this association belongs.

The CS-RS Association class is linked to the CS and RS classes.

A participant class and its objects have information about a device that is part of a CS and/or contributes/consumes media stream(s) belonging to a CS. The participant object is represented in the XML schema using the 'participant' element.

A participant class has two attributes: nameID - This attribute is a list of (Name, AoR) tuples. An AoR (Section 6 of ) can be either a SIP/SIPS/tel URI ("SIPS" means "SIP Secure"; the tel URI is discussed in ), a Fully Qualified Domain Name (FQDN), or an IP address. For example, the AoR may be drawn from the From header field or the P&nbhy;Asserted-Identity header field. The SRC's local policy is used to decide where to draw the AoR from. The Name parameter represents the participant name (SIP display name) or dialed number (when known). Multiple tuples are allowed for cases where a participant has more than one AoR. For example, a P&nbhy;Asserted&nbhy;Identity header can have both SIP and tel URIs. participant_id - This attribute is used to identify the 'participant' XML element within the XML document. It is generated using the rules mentioned in . This attribute MUST be used for all references to a participant within a CS-Group, and MAY be used to reference the same participant more globally. This document does not specify other attributes relating to participants (e.g., participant role, participant type). An SRC that has information regarding these attributes can provide this information as part of extension data to the 'participant' XML element from the SRC to the SRS.

The participant class is linked to the MS and CS classes by using an association relationship. The association between the participant and the MS allows the following: A participant will receive zero or more media streams. A participant will send zero or more media streams. (The same participant provides multiple streams, e.g., audio and video.) A media stream will be received by zero or more participants. It is possible, though perhaps unlikely, that a stream is generated but sent only to the SRC and SRS, not to any participant -- for example, in conferencing where all participants are on hold and the SRC is collocated with the focus. Also, a media stream may be received by multiple participants (e.g., "whisper" calls, side conversations). A media stream will be sent by one or more participants (pre-mixed streams). An example of a case where a participant receives zero or more streams is where a supervisor may have a side conversation with an agent while the agent converses with a customer.

The Participant-CS Association class describes the association of a participant to a CS for a period of time. A participant may be associated to and dissociated from a CS several times (for example, connecting to a conference, then disconnecting, then connecting again). The Participant-CS Association object is represented in the XML schema using the 'participantsessionassoc' element.

The Participant-CS Association class has the following attributes: associate-time - associate-time is calculated by the SRC as the time it sees a participant associated to a CS. disassociate-time - disassociate-time is calculated by the SRC as the time it sees a participant disassociate from a CS. It is possible that a given participant can have multiple associate times / disassociate times within a given communication session. param - The capabilities here are those that are indicated in the Contact header as defined in Section 9 of . For example, in a CS (which can be a conference), you can have participants who are playing the role of "focus". These participants do not contribute to media in the CS; however, they switch the media received from one participant to every other participant in the CS. Indicating the capabilities of the participants (here, "focus") would be useful for the recorder to learn about these kinds of participants. The capabilities are represented using the 'param' XML element in the metadata. The 'param' XML element encoding defined in is used to represent the capability attributes in metadata. Each participant may have zero or more capabilities. A participant may use different capabilities, depending on the role it plays at a particular instance -- for example, if a participant moves across different CSs (e.g., due to transfer) or is simultaneously present in different CSs with different roles. participant_id - This attribute identifies the participant to which this association belongs. session_id - This attribute identifies the session to which this association belongs.

The Participant-CS Association class is linked to the participant and CS classes.

A MS class (and its objects) has the properties of media as seen by the SRC and sent to the SRS. Different snapshots of MS objects may be sent whenever there is a change in media (e.g., a direction change, like pause/resume, codec change, and/or participant change). The MS object is represented in the XML schema using the 'stream' element.

A MS class has the following attributes: label - The 'label' attribute within the 'stream' XML element references an SDP "a=label" attribute that identifies an m-line within the RS SDP. That m-line carries the media stream from the SRC to the SRS. content-type - The content of a MS element will be described in terms of the "a=content" attribute defined in Section 5 of . If the SRC wishes to convey the content-type to the SRS, it does so by including an "a=content" attribute with the m-line in the RS SDP. stream_id - Each 'stream' element has a unique 'stream_id' attribute that helps to uniquely identify the stream. This identifier is generated using the rules mentioned in . session_id - This attribute associates the stream with a specific 'session' element. The metadata model can include media streams that are not being delivered to the SRS. For example, an SRC offers audio and video towards an SRS that accepts only audio in response. The metadata snapshots sent from the SRC to the SRS can continue to indicate the changes to the video stream as well.

A MS class is linked to the participant and CS classes by using the association relationship. Details regarding associations with the participant are described in . Details regarding associations with the CS are mentioned in .

A Participant-Stream Association class describes the association of a participant to a MS for a period of time, as a sender or as a receiver, or both. This class is represented in XML using the 'participantstreamassoc' element.

A Participant-Stream Association class has the following attributes: associate-time - This attribute indicates the time a participant started contributing to a MS. disassociate-time - This attribute indicates the time a participant stopped contributing to a MS. send - This attribute indicates whether a participant is contributing to a stream or not. This attribute has a value that points to a stream represented by its unique_id. The presence of this attribute indicates that a participant is contributing to a stream. If a participant stops contributing to a stream due to changes in a CS, a snapshot MUST be sent from the SRC to the SRS with no 'send' element for that stream. recv - This attribute indicates whether a participant is receiving a media stream or not. This attribute has a value that points to a stream represented by its unique_id. The presence of this attribute indicates that a participant is receiving a stream. If the participant stops receiving a stream due to changes in a CS (like hold), a snapshot MUST be sent from the SRC to the SRS with no 'recv' element for that stream. participant_id - This attribute points to the participant with which a 'stream' element is associated. The 'participantstreamassoc' XML element is used to represent a participant association with a stream. The 'send' and 'recv' XML elements MUST be used to indicate whether a participant is contributing to a stream or receiving a stream. There MAY be multiple instances of the 'send' and 'recv' XML elements inside a 'participantstreamassoc' element. If a metadata snapshot is sent with a 'participantstreamassoc' element that does not have any 'send' and 'recv' elements, it means that the participant is neither contributing to any streams nor receiving any streams.

The Participant-Stream Association class is linked to the participant and MS classes.

The XML elements 'associate-time', 'disassociate-time', 'start-time', and 'stop-time' contain strings representing the date and time. The value of these elements MUST follow the Instant Messaging and Presence Protocol (IMPP) date-time format . Timestamps that contain "T" or "Z" MUST use the capitalized forms. As a security measure, the 'timestamp' element MUST be included in all tuples, unless the exact time of the status change cannot be determined.

A unique_id is generated in two steps: The Universally Unique Identifier (UUID) is created using any of the procedures mentioned in Sections 4.3, 4.4, and 4.5 of . The algorithm MUST ensure that it does not use any potentially personally identifying information to generate the UUIDs. If implementations are using a Name&nbhy;Based UUID as defined in Section 4.3 of , a namespace ID generated using the guidance in Section 4.2 or 4.5 of might be a good choice. The UUID is encoded using base64 as defined in . The above-mentioned unique_id mechanism SHOULD be used for each metadata element. Multiple SRCs can refer to the same element/UUID (how each SRC learns the UUID here is beyond the scope of this document). If two SRCs use the same UUID, they MUST retain the UUID&wj;/element mapping. If the SRS detects that a UUID is mapped to more than one element at any point in time, it MUST treat this as an error. For example, the SRS may choose to reject or ignore the portions of metadata where it detects that the same UUID is mapped to an element that is different than the expected element (the SRS learns the mapped UUID when it sees an element for the first time in a metadata instance).

The Metadata version is defined to help the SRC and SRS know the version of metadata XML schema used. SRCs and SRSs that support this specification MUST use version 1 in the namespace (urn:ietf:params:xml:ns:recording:1) in all the XML documents. Implementations may not interoperate if the version implemented by the sender is not known by the receiver. No negotiation of versions is provided. The version number has no significance, although documents that update or obsolete this document (possibly including drafts of such documents) should include a higher version number if the metadata XML schema changes.