]> National Institute of Information and Communications Technology

4-2-1 Nukui-Kitamachi Koganei Tokyo 184-8795 Japan asaeda@nict.go.jp

Huawei Technologies Co., Ltd.

101 Software Avenue, Yuhua District Nanjing Jiangsu 210012 China bill.wu@huawei.com

Huawei Technologies Co., Ltd.

101 Software Avenue, Yuhua District Nanjing Jiangsu 210012 China Rachel@huawei.com

This document defines two RTP Control Protocol (RTCP) Extended Report (XR) Blocks that allow the reporting of synchronization delay and offset metrics for use in a range of RTP applications.

This document defines two new block types to augment those defined in , for use in a range of RTP applications. The first new block type supports reporting of Initial Synchronization Delay to establish multimedia session. Information is recorded about time difference between the start of RTP sessions and the time the RTP receiver acquires all components of RTP sessions in the multimedia session . The second new block type supports reporting of the relative synchronization offset time of two arbitrary streams (e.g., between audio and video streams), with the same RTCP CNAME included in RTCP SDES packets . These metrics belong to the class of transport level metrics defined in .

The use of RTCP for reporting is defined in . defined an extensible structure for reporting using an RTCP Extended Report (XR). This document defines a new Extended Report block for use with and .

The RTP Monitoring Architectures provides guideline for reporting block format using RTCP XR. The new report block described in this memo is in compliance with the monitoring architecture specified in .

When joining each session in layered video sessions or the multimedia session, a receiver may not synchronize playout across the multimedia session or layered video session until RTCP SR packets have been received on all components of RTP sessions. The component RTP session are referred to as each RTP session for each media type in multimedia session or separate RTP session for each layer in the layered video session. For multicast session, the initial synchronization delay metric varies with the session bandwidth, the number of members, and the number of senders in the session. The RTP flow Initial synchronization delay block defined in this document can be used to report such metric, i.e., the initial synchronization delay to receive all the RTP streams belonging to the same multimedia session or layered video session. In the absence of packet loss, the initial synchronization delay equals to the average time taken to receive the first RTCP packet in the RTP session with the longest RTCP reporting interval. In the presence of packet loss, the media synchronization should rely on the in-band mapping of RTP and NTP-format timestamps or wait until the reporting interval has passed, and the next RTCP SR packet is sent. Receivers of the RTP flow initial synchronization delay block could use this metric to compare with targets (i.e., Service Level Agreement or thresholds of the system) to help ensure the quality of real-time application performance. In an RTP multimedia session, there can be an arbitrary number of streams carried in different RTP sessions, with the same RTCP CNAME. These streams may be not synchronized with each other. For example, one audio stream and one video stream belong to the same session, and the audio stream is transmitted lagging behind video stream for multiple tens of milliseconds . The RTP Flows Synchronization Offset block can be used to report such synchronization offset between video stream and audio stream. This block is also applied to the case where an RTP session can contain media streams with media from multiple media types. The metrics defined in the RTP flows synchronization Offset block can be used by the network manager for trouble shooting and dealing with user experience issues.

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119. In addition, the following terms are defined: A multimedia session comprises a set of concurrent RTP sessions among a common group of participants, using one RTP session for each media type. The initial synchronization Delay is the average time for receiver to synchronize all components of a multimedia session . Synchronization between two media streams must be maintained to ensure satisfactory QoE. Two media streams can be of the same or different media type belonging to one RTP session or in different media types belonging to one multimedia session. The Synchronization Offset is the relative time difference of the two media streams that need to be synchronized.

This block is sent by RTP receivers and reports Initial synchronization delay beyond the information carried in the standard RTCP packet format. Information is recorded about time difference between the start of multimedia session and the time when the RTP receiver acquires all components of RTP sessions measured at the receiving end of RTP stream. This block needs only be exchanged occasionally, for example sent once at the start of RTP session.

The RTP Flows Initial Synchronization Delay Report Block has the following format:

0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | BT=RFISD | Reserved | Block length=2 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | SSRC of Source | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Initial Synchronization Delay | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Figure 1: Report Block Structure

The RTP Flows Initial Synchronization Delay Report Block is identified by the constant <RFISD>. [Note to RFC Editor: please replace RFISD with the IANA provided RTCP XR block type for this block.] This field is reserved for future definition. In the absence of such a definition, the bits in this field MUST be set to zero and ignored by the receiver. The constant 2, in accordance with the definition of this field in Section 3 of RFC 3611. The SSRC of the media source SHALL be set to the value of the SSRC identifier carried in any arbitrary component of RTP sessions belonging to the same multimedia session. The average delay, expressed in units of 1/65536 seconds, from the beginning of multimedia session to the time when RTCP packets are received on all of the components RTP sessions. It is recommended that the beginning of multimedia session is chosen as the time when the receiver has joined the first RTP session of the multimedia session. The value of the initial synchronization delay is calculated based on received RTCP SR packets or the RTP header extension containing in-band mapping of RTP and NTP-format timestamps . If there is no packet loss, the initial synchronization delay is expected to be equal to the average time taken to receive the first RTCP packet in the RTP session with the longest RTCP reporting interval or the average time taken to receive the first RTP header extension containing in-band mapping of RTP and NTP-format timestamps. If the measurement is unavailable, the value of this field with all bits set to 1 MUST be reported.

In the RTP multimedia sessions or one RTP session, there can be an arbitrary number of Media streams and each media stream (e.g., audio stream or video stream) is sent in a separate RTP stream. In case of one RTP session, each media stream or each medium uses different SSRC. The receiver associates RTP streams to be synchronized by means of RTCP CNAME contained in the RTCP Source Description (SDES) packets . This block is sent by RTP receivers and reports synchronization offset of two arbitrary RTP streams that needs to be synchronized in the RTP multimedia session. Information is recorded about the relative average time difference between two arbitrary RTP streams (one is reporting stream, the other is reference stream) with the same CNAME and measured at the receiving end of RTP stream. In order to tell what the offset of reporting stream is relative to, the block for reference stream with synchronization offset of zero should be reported. Instances of this Block refer by Synchronization source (SSRC) to the separate auxiliary Measurement Information block which describes measurement periods in use (see section 4.2). This metrics block relies on the measurement period in the Measurement Information block indicating the span of the report and SHOULD be sent in the same compound RTCP packet as the measurement information block. If the measurement period is not received in the same compound RTCP packet as this Block, this Block MUST be discarded.

The RTP Flow General Synchronization Offset Report Block has the following format:

0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | BT=RFSO | I | Reserved | Block length=3 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | SSRC of source | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Synchronization Offset, most significant word | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Synchronization Offset, least significant word | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Figure 2: Report Block Structure

The RTP Flow General Synchronization Offset Report Block is identified by the constant <RFSO>. [Note to RFC Editor: please replace RFSO with the IANA provided RTCP XR block type for this block.] This field is used to indicate whether the Burst/Gap Discard Summary Statistics metrics are Sampled, Interval or Cumulative metrics: I=10: Interval Duration - the reported value applies to the most recent measurement interval duration between successive metrics reports. I=11: Cumulative Duration - the reported value applies to the accumulation period characteristic of cumulative measurements. I=01: Sampled Value - the reported value is a sampled instantaneous value. In this document, the value I=00 is the reserved value and MUST NOT be used. If the value I=00 is received, then the XR block MUST be ignored by the receiver. This field is reserved for future definition. In the absence of such a definition, the bits in this field MUST be set to zero and MUST be ignored by the receiver. The constant 3, in accordance with the definition of this field in Section 3 of RFC 3611. The SSRC of the media source SHALL be set to the value of the SSRC identifier of the reporting RTP stream to which the XR relates. The synchronization offset of the reporting RTP stream relative to the reference stream with the same CNAME. The calculation of Synchronization Offset is similar to Difference D calculation in the RFC3550. That is to say, if Si is the NTP timestamp from the reporting RTP packet i, and Ri is the time of arrival in NTP timestamp units for reporting RTP packet i, Sj is the NTP timestamp from the reference RTP packet j, and Rj is the time of arrival in NTP timestamp units for reference RTP packet j, then the value of the synchronization offset D may be expressed as D(i,j) = (Rj - Ri) - (Sj - Si) = (Rj - Sj) - (Ri - Si) If in-band delivery of NTP-format timestamps is supported , Si and Sj should be obtained directly from the RTP packets where NTP timestamps are available. If not, Si and Sj should be calculated from their corresponding RTP timestamps. The value of the synchronization offset is represented using a 64-bit signed NTP-format timestamp as defined in , which is 64-bit signed fixed-point number with the integer part in the first 32 bits and the fractional part in the last 32 bits. A positive value of the synchronization offset means that the reporting stream leads before the reference stream, while a negative one means the reporting stream lags behind the reference stream. The synchronization offset of zero means the stream is the reference stream.If the measurement is unavailable, the value of this field with all bits set to 1 MUST be reported.

defines the use of SDP (Session Description Protocol) for signaling the use of XR blocks. XR blocks MAY be used without prior signaling.

Two new parameters are defined for the two report blocks defined in this document to be used with Session Description Protocol (SDP) using the Augmented Backus-Naur Form (ABNF) . They have the following syntax within the "rtcp-xr" attribute :

xr-format = xr-rfisd-block / xr-rfso-block xr-rfisd-block = "rtp-flow-init-syn-delay" xr-rfso-block = "rtp-flow-syn-offset"

Refer to Section 5.1 of RFC 3611 for a detailed description and the full syntax of the "rtcp-xr" attribute.

When SDP is used in offer-answer context, the SDP Offer/Answer usage defined in applies.

New report block types for RTCP XR are subject to IANA registration. For general guidelines on IANA allocations for RTCP XR, refer to Section 6.2 of. This document assigns two new block type values in the RTCP XR Block Type Registry: RFISD RTP Flows Initial Synchronization Delay <RFISD> RFSO RTP Flows Synchronization Offset Metrics Block <RFSO> This document also registers two new SDP parameters for the "rtcp-xr" attribute in the RTCP XR SDP Parameters Registry: "rtp-flow-init-syn-delay " "rtp-flow-syn-offset" The contact information for the registrations is: Qin Wu sunseawq@huawei.com 101 Software Avenue, Yuhua District Nanjing, Jiangsu 210012, China

The new RTCP XR report blocks proposed in this document introduces no new security considerations beyond those described in .

The authors would like to thank Bill Ver Steeg, David R Oran, Ali Begen, Colin Perkins, Roni Even, Kevin Gross, Jing Zhao, Fernando Boronat Seguí, Mario Montagud Climent, Youqing Yang, Wenxiao Yu and Yinliang Hu,Jonathan Lennox for their valuable comments and suggestions on this document.

This memo describes an RTP payload format for Scalable Video Coding (SVC) as defined in Annex G of ITU-T Recommendation H.264, which is technically identical to Amendment 3 of ISO/IEC International Standard 14496-10. The RTP payload format allows for packetization of one or more Network Abstraction Layer (NAL) units in each RTP packet payload, as well as fragmentation of a NAL unit in multiple RTP packets. Furthermore, it supports transmission of an SVC stream over a single as well as multiple RTP sessions. The payload format defines a new media subtype name "H264-SVC", but is still backwards compatible to [I-D.ietf-avt-rtp-rfc3984bis] since the base layer, when encapsulated in its own RTP stream, must use the H.264 media subtype name ("H264") and the packetization method specified in [I- D.ietf-avt-rtp-rfc3984bis]. The payload format has wide applicability in videoconferencing, Internet video streaming, and high bit-rate entertainment-quality video, among others.Table of Contents &rfc3611; &rfc2119; &rfc4566; &rfc5234; &rfc3550; This memo outlines how RTP sessions are synchronized, and discusses how rapidly such synchronisation can occur. We show that most RTP sessions can be synchronized immediately, but that the use of video switching multipoint conference units (MCUs) or large source specific multicast (SSM) groups can greatly increase the synchronisation delay. This increase in delay can be unacceptable to some applications that use layered and/or multi-description codecs. This memo introduces three mechanisms to reduce the synchronisation delay for such sessions. First, it updates the RTP Control Protocol (RTCP) timing rules to reduce the initial synchronisation delay for SSM sessions. Second, a new feedback packet is defined for use with the Extended RTP Profile for RTCP-based Feedback (RTP/AVPF), allowing video switching MCUs to rapidly request resynchronisation. Finally, new RTP header extensions are defined to allow rapid synchronisation of late joiners, and guarantee correct timestamp based decoding order recovery for layered codecs in the presence of clock skew. Huawei ITU-T BBF Forum

RFC EDITOR NOTE: please change XXXX in [RFCXXXX] by the new RFC number, when assigned. Initial Synchronization Delay Metric Metric Name: RTP Initial Synchronization Delay Metric Description: See Section 2.1,Initial Synchronization Delay term [RFCXXXX]. Method of Measurement or Calculation: See section 3.2, Initial Synchronization Delay definition [RFCXXXX]. Units of Measurement: See section 3.2, Initial Synchronization Delay definition [RFCXXXX]. Measurement Point(s) with Potential Measurement Domain: See section 3, 1st paragraph [RFCXXXX]. Measurement Timing: See section 3, 2nd paragraph [RFCXXXX] for measurement timing. Use and applications: See section 1.4 [RFCXXXX]. Reporting model: See RFC3611. Synchronization Offset Metric Metric Name: RTP Synchronization Offset Delay Metric Description: See Section 2.1, Synchronization Offset term [RFCXXXX]. Method of Measurement or Calculation: See section 4.2, Initial Synchronization Delay definition [RFCXXXX]. Units of Measurement: See section 4.2, Initial Synchronization Delay definition [RFCXXXX]. Measurement Point(s) with Potential Measurement Domain: See section 4, 2nd paragraph [RFCXXXX]. Measurement Timing: See section 4, 3rd paragraph [RFCXXXX] for measurement timing and section 4.2 [RFCXXXX] for Interval Metric flag. Use and applications: See section 1.4 [RFCXXXX]. Reporting model: See RFC3611.

Note to the RFC-Editor: please remove this section prior to publication as an RFC.

The following are the major changes compared to previous version: Some Editorial changes based on Gen-Art Reviewer comments.

The following are the major changes compared to previous version: Minor Editorial changes.

The following are the major changes compared to previous version: Some Editorial changes.

The following are the major changes compared to previous version: Editorial changes and typo fixed.

The following are the major changes compared to previous version: Additional text to clarify on how to distinguish report stream from reference stream. Other Editorial changes.

The following are the major changes compared to previous version: Remove the need to signal the reference source in the synchronisation offset metrics RTCP XR report. Apply RFC6390 template to metrics in the appendix. Other editorial changes to get inline with other XRBLOCK drafts.

The following are the major changes compared to previous version: Editorial change based on comments raised on the list and in the IETF85 meeting