< draft-zheng-xrblock-effective-loss-index-00.txt   draft-zheng-xrblock-effective-loss-index-01.txt >
Network Working Group H. Zheng Network Working Group H. Zheng
Internet-Draft R. Even Internet-Draft R. Even
Intended status: Informational Q. Wu Intended status: Informational Q. Wu
Expires: May 3, 2018 Huawei Expires: August 2, 2018 Huawei
R. Gu R. Gu
China Mobile China Mobile
October 30, 2017 R. Huang
Huawei
January 29, 2018
RTP Control Protocol (RTCP) Extended Report (XR) Block for Effective RTP Control Protocol (RTCP) Extended Report (XR) Block for Effective
Loss Index Reporting Loss Index Reporting
draft-zheng-xrblock-effective-loss-index-00 draft-zheng-xrblock-effective-loss-index-01
Abstract Abstract
This document defines a new metric for RTP applications to measure This document defines a new metric for RTP monitors to estimate the
the effectiveness of stream repair means, and an RTP Control Protocol effectiveness of stream repair means, and an RTP Control Protocol
(RTCP) Extended Report (XR) Block to report the metric. (RTCP) Extended Report (XR) Block to report the metric.
Status of This Memo Status of This Memo
This Internet-Draft is submitted in full conformance with the This Internet-Draft is submitted in full conformance with the
provisions of BCP 78 and BCP 79. provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF). Note that other groups may also distribute Task Force (IETF). Note that other groups may also distribute
working documents as Internet-Drafts. The list of current Internet- working documents as Internet-Drafts. The list of current Internet-
skipping to change at page 1, line 40 skipping to change at page 1, line 42
Internet-Drafts are draft documents valid for a maximum of six months Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any and may be updated, replaced, or obsoleted by other documents at any
time. It is inappropriate to use Internet-Drafts as reference time. It is inappropriate to use Internet-Drafts as reference
material or to cite them other than as "work in progress." material or to cite them other than as "work in progress."
This Internet-Draft will expire on May 3, 2018. This Internet-Draft will expire on May 3, 2018.
Copyright Notice Copyright Notice
Copyright (c) 2017 IETF Trust and the persons identified as the Copyright (c) 2018 IETF Trust and the persons identified as the
document authors. All rights reserved. document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents Provisions Relating to IETF Documents
(http://trustee.ietf.org/license-info) in effect on the date of (http://trustee.ietf.org/license-info) in effect on the date of
publication of this document. Please review these documents publication of this document. Please review these documents
carefully, as they describe your rights and restrictions with respect carefully, as they describe your rights and restrictions with respect
to this document. Code Components extracted from this document must to this document. Code Components extracted from this document must
include Simplified BSD License text as described in Section 4.e of include Simplified BSD License text as described in Section 4.e of
the Trust Legal Provisions and are provided without warranty as the Trust Legal Provisions and are provided without warranty as
described in the Simplified BSD License. described in the Simplified BSD License.
Table of Contents Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.1. Effective Loss Index . . . . . . . . . . . . . . . . . . 3 1.1. Effective Loss Index . . . . . . . . . . . . . . . . . . . 3
1.2. Applicability . . . . . . . . . . . . . . . . . . . . . . 4 1.2. Applicability . . . . . . . . . . . . . . . . . . . . . . 4
1.3. RTCP and RTCP XR Reports . . . . . . . . . . . . . . . . 5 1.3. RTCP and RTCP XR Reports . . . . . . . . . . . . . . . . . 5
1.4. Performance Metrics Framework . . . . . . . . . . . . . . 5 1.4. Performance Metrics Framework . . . . . . . . . . . . . . 5
2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 5 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 5
3. Effective Loss Index Report Block . . . . . . . . . . . . . . 5 3. Effective Loss Index Report Block . . . . . . . . . . . . . . 5
4. SDP Signaling . . . . . . . . . . . . . . . . . . . . . . . . 6 4. SDP Signaling . . . . . . . . . . . . . . . . . . . . . . . . 6
4.1. SDP rtcp-xr-attrib Attribute Extension . . . . . . . . . 6 4.1. SDP rtcp-xr-attrib Attribute Extension . . . . . . . . . . 6
4.2. Offer/Answer Usage . . . . . . . . . . . . . . . . . . . 7 4.2. Offer/Answer Usage . . . . . . . . . . . . . . . . . . . . 8
5. Security Considerations . . . . . . . . . . . . . . . . . . . 7 5. Security Considerations . . . . . . . . . . . . . . . . . . . 8
6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 7 6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 8
6.1. New RTCP XR Block Type Value . . . . . . . . . . . . . . 7 6.1. New RTCP XR Block Type Value . . . . . . . . . . . . . . . 8
6.2. New RTCP XR SDP Parameter . . . . . . . . . . . . . . . . 8 6.2. New RTCP XR SDP Parameter . . . . . . . . . . . . . . . . 9
6.3. Contact Information for Registrations . . . . . . . . . . 8 6.3. Contact Information for Registrations . . . . . . . . . . 9
7. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 8 7. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 9
8. References . . . . . . . . . . . . . . . . . . . . . . . . . 8 8. References . . . . . . . . . . . . . . . . . . . . . . . . . . 9
8.1. Normative References . . . . . . . . . . . . . . . . . . 8 8.1. Normative References . . . . . . . . . . . . . . . . . . . 9
8.2. Informative References . . . . . . . . . . . . . . . . . 8 8.2. Informative References . . . . . . . . . . . . . . . . . . 9
Appendix A. Metric Represented Using the Template from RFC 6390 10 Appendix A. Metric Represented Using the Template from RFC 6390 . 11
A.1. Effective Loss Index . . . . . . . . . . . . . . . . . . 10 A.1. Effective Loss Index . . . . . . . . . . . . . . . . . . . 11
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 10 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 11
1. Introduction 1. Introduction
RTP applications often use stream repair means, e.g. FEC (Forward RTP applications often use stream repair means, e.g. FEC (Forward
Error Correction) [RFC5109] and/or retransmission [RFC4588] to Error Correction) [RFC5109] and/or retransmission [RFC4588] to
improve the robustness of media streams. With the presence of those improve the robustness of media streams. With the presence of those
stream repair means, a degree of packet loss can be recovered for a stream repair means, a degree of packet loss can be recovered for a
media stream. In the past, some RTCP Extend Reports (XRs) were media stream. In the past, some RTCP Extend Reports (XRs) were
defined to reflect the situation of post-repair loss. For example, defined to reflect the situation of post-repair loss. For example,
[RFC5725] defines an XR block using Run Length Encoding (RLE) to [RFC5725] defines an XR block using Run Length Encoding (RLE) to
report post-repair loss; [RFC7509] defines count metrics for post- report post-repair loss; [RFC7509] defines count metrics for post-
repair loss. repair loss.
This document proposes a new metric Effective Loss Index (ELI) to This document proposes a new metric Effective Loss Index (ELI) to
measure the effectiveness of stream repair means. The new metric estimate the effectiveness of stream repair means. The new metric
provides a simpler view on the post-repair loss than the mechanisms provides a simpler view on the post-repair loss than the mechanisms
documented in [RFC5725] and [RFC7509]. EFI is an index, so the documented in [RFC5725] and [RFC7509]. EFI is an index, so the
values reported from different RTP sources can be compared directly, values reported from the monitors deployed in the different places in
which makes it easier to rank the effectiveness of loss repair means. the network can be compared directly, which makes it easier to
An example use case is to find endpoints whose ELI values are at diagnose the network problem when delivering the RTP streams. A use
bottom 10%. For those endpoints, more informative XR reports such as case is to compare the ELI value reported by a monitor with a certain
those in [RFC5725] and [RFC7509] can then be used to discover more reasonable threshold to see if there are any problems during an RTP
details about the loss situations. streaming session. For those endpoints, more informative XR reports
such as those in [RFC5725] and [RFC7509] can then be used to discover
more details about the loss situations.
This document also defines an XR block to report the metric, which This document also defines in section 3, an XR block to report the
can be found out in Section 3. metric.
1.1. Effective Loss Index 1.1. Effective Loss Index
Effective Loss Index (ELI) uses a simple model to measure the Effective Loss Index (ELI) uses a simple model to measure the loss
effectiveness of loss repair. The model assumes that repair means impact after applying loss repairsof loss repair. It is useful
are applied onto packets by batches of equal size. Lower ELI means especially in the middleboxes which usually are passive observer and
that the repair was more successful. Specifically, a batch is do not have the ability to recover the loss data.
identified by a range of RTP sequence numbers. The size of a batch
is number of packets. An application can agree upon a default batch
size, or use the SDP signaling defined in Section 4.1 to communicate
one.
An RTP endpoint is thought to process received packets and apply The model assumes that repair means are applied onto packets by
batches of equal size. Lower ELI means that loss impact is minimal.
Specifically, a batch is identified by a range of RTP sequence
numbers. The size of a batch is number of packets. An application
can agree upon a default batch size, or use the SDP signaling defined
in Section 4.1 to communicate one if the middlebox can see the SDP,
or just configure it.
An RTP endpoint is assumed to process received packets and apply
repair means batch by batch. For each batch, if there is still some repair means batch by batch. For each batch, if there is still some
unrecoverable loss after having applied the repair means, then the unrecoverable loss after having applied the repair means, then the
repair means are deemed as ineffective. The ineffectiveness is repair means are deemed as ineffective. The ineffectiveness is
denoted by Effective Loss Factor (ELF), along with a parameter denoted by Effective Loss Factor (ELF), along with a parameter Loss
Effective Loss Threshold, showing below: Repair Threshold, showing below:
if Post-Repair Loss > Effective Loss Threshold if Loss Packets Number > Loss Repair Threshold
Effective Loss Factor = 1 Effective Loss Factor = 1
else else
Effective Loss Factor = 0 Effective Loss Factor = 0
endif endif
Figure 1: Calculation of Effective Loss Factor Figure 1: Calculation of Effective Loss Factor
The parameters in Figure 1 are explained below: The parameters in Figure 1 are explained below:
o Post-Repair Loss is the number of packet lost after repair in the o Loss Packets Number is the number of packet lost in the batch.
batch.
o Effective Loss Threshold is in number of packets. o Loss Repair Threshold indicates the maximum loss packets number
that can be recovered.
The minimum value of Effective Loss Threshold is zero. This document The minimum value of Loss Repair Threshold is zero, which means there
does not mandate any value for Effective Loss Threshold. is no loss repair. This document does not mandate any value for Loss
Applications can prescribe a value for themselves without signaling. Repair Threshold. Applications can prescribe a value for themselves
without signaling. For example, it can be calculated by the batch
size multiplied by the fixed redundancy ratio of the FEC algorithm.
On the other hand, SDP signaling defined in Section 4.1 can be used On the other hand, SDP signaling defined in Section 4.1 can be used
to communicate the value. Determining an Effective Loss Threshold to communicate the value.
value for use can be empirical, applications may have to try out and
change the value from time to time, depending on their needs.
Effective Loss Index is an integer derived by calculating the average Effective Loss Index is an integer derived by calculating the average
Effective Loss Factor across a sequence of consecutive batches of RTP Effective Loss Factor across a sequence of consecutive batches of RTP
packets. Let ELF(i) be the Effective Loss Factor calculated for i-th packets. Let ELF(i) be the Effective Loss Factor calculated for i-th
batch, and N as number of batches in the sequence, then Effective batch, and N as number of batches in the sequence, then Effective
Loss Index is calculated as: Loss Index is calculated as:
ELF(1)+ELF(2)+ ...+ELF(N) ELF(1)+ELF(2)+ ...+ELF(N)
Effective Loss Index = ------------------------- x 10000 Effective Loss Index = -------------------------
N N
Figure 2: Calculation of Effective Loss Index Figure 2: Calculation of Effective Loss Index
The following is an example of how to calculate Effective Loss Index. The following is an example of how to calculate Effective Loss Index.
For simplicity and demonstration purpose, the size of batches is For simplicity and demonstration purpose, the size of a batch is
assumed to be 3, and the Effective Loss Threshold is assumed to be 1. assumed to be 3, and the Loss Repair Threshold is assumed to be 1.
The example processes a sequence of 9 RTP packets in 3 batches. The example processes a sequence of 9 RTP packets (x means lost) in 7
batches.
Batch Post-Repair Effective 1xx4x6x89
Loss Loss Factor
Batch Loss Effective Loss Factor
| 1 2 3 | 2, 3 1 | 1 2 3 | 2, 3 1
| 2 3 4 | 2, 3 1
| 3 4 5 | 3 0
| 4 5 6 | 5 0 | 4 5 6 | 5 0
| 5 6 7 | 5, 7 1
| 6 7 8 | 7 0
| 7 8 9 | 7 0 | 7 8 9 | 7 0
1 + 0 + 0 1+1+0+0+1+0+0
Effective Loss Index = ----------- x 10000 = 3333 Effective Loss Index = --------------- = 0.4285
3 7
1.2. Applicability This example provides fine grained estimation for loss recovery. It
can detect the loss burst happening over batches. Implementations can
also do coarse grained estimation by simply dividing total packets
into several batches.
1.2. Applicability
The metric defined by this document is applicable to a range of RTP The metric defined by this document is applicable to a range of RTP
applications that send packets in batches of equal length, probably applications that send packets with stream repair means (e.g.,
with stream repair means (e.g., Forward Error Correction (FEC) Forward Error Correction (FEC) [RFC5109] and/or retransmission
[RFC5109] and/or retransmission [RFC4588]) applied on the batches. [RFC4588]) applied on them. Note that this metric is only valuable
Note that in order to not interfere with the batches being protected, for FECs where he redundant data are sent in a different RTP stream
any additional packets generated by the stream repair means SHOULD be from the original media stream.
in a different RTP stream.
The number of batches among which ELI is calculated should not be too The number of batches among which ELI is calculated should not be too
few, otherwise the result may be too biased. However, specifying a few, otherwise the result may be biased. It is suggested to calculate
minimal number of batches seems unrealistic, due to the stream repair it based on the total number of RTP packets during the measurement
means used by applications can be quite different. This document interval, as in the section 1.1 example:
leaves it to applications to choose a suitable minimal value for the
number of batches. The number of batches = (The total number of RTP packets - the size
of a batch) + 1.
1.3. RTCP and RTCP XR Reports 1.3. RTCP and RTCP XR Reports
The use of RTCP for reporting is defined in [RFC3550]. [RFC3611] The use of RTCP for reporting is defined in [RFC3550]. [RFC3611]
defines an extensible structure for reporting by using an RTCP defines an extensible structure for reporting by using an RTCP
Extended Report (XR). This document defines a new Extended Report Extended Report (XR). This document defines a new Extended Report
block for use with [RFC3550] and [RFC3611]. block for use with [RFC3550] and [RFC3611].
1.4. Performance Metrics Framework 1.4. Performance Metrics Framework
skipping to change at page 6, line 4 skipping to change at page 6, line 18
identified by the constant 'TBD'. identified by the constant 'TBD'.
[[Editor Note: should replace 'TBD' with assigned value]] [[Editor Note: should replace 'TBD' with assigned value]]
Reserved: 8 bits: These bits are reserved for future use. They MUST Reserved: 8 bits: These bits are reserved for future use. They MUST
be set to zero by senders and ignored by receivers (see be set to zero by senders and ignored by receivers (see
Section 4.2 of [RFC6709]). Section 4.2 of [RFC6709]).
Block length: 16 bits: This field is in accordance with the Block length: 16 bits: This field is in accordance with the
definition in [RFC3611]. In this report block, it MUST be set to definition in [RFC3611]. In this report block, it MUST be set to
3. The block MUST be discarded if the block length is set to a 3. The block MUST be discarded if the block length is set to a
different value. different value.
SSRC of source: 32 bits: As defined in Section 4.1 of [RFC3611]. SSRC of source: 32 bits: The SSRC of the RTP data packet source
being reported upon by this report block, as defined in Section
4.1 of [RFC3611].
Effective Loss Index: 16 bits: The value of this field SHOULD be set Effective Loss Index: 16 bits: The value of Effective Loss Index,
to the calculated result of Effective Loss Index (as in Figure 2). equivalent to taking the integer part after multiplying the the
calculated result of Effective Loss Index (as in Figure 2) by
65535.
Padding: 16 bits: These bits MUST be set to zero by senders and Padding: 16 bits: These bits MUST be set to zero by senders and
ignored by receivers. ignored by receivers.
4. SDP Signaling 4. SDP Signaling
[RFC3611] defines the use of SDP (Session Description Protocol) for [RFC3611] defines the use of SDP (Session Description Protocol) for
signaling the use of RTCP XR blocks. However, XR blocks MAY be used signaling the use of RTCP XR blocks. However, XR blocks MAY be used
without prior signaling (see Section 5 of [RFC3611]). without prior signaling (see Section 5 of [RFC3611]).
skipping to change at page 8, line 23 skipping to change at page 9, line 23
6.3. Contact Information for Registrations 6.3. Contact Information for Registrations
The contact information for the registrations is: The contact information for the registrations is:
RAI Area Directors <rai-ads@ietf.org> RAI Area Directors <rai-ads@ietf.org>
7. Acknowledgements 7. Acknowledgements
This document has benefited greatly from the comments of various This document has benefited greatly from the comments of various
people. The following individuals have contributed to this document: people. The following individuals have contributed to this document:
Rachel Huang, Colin Perkins, Yanfang Zhang, Lingyan Wu. Colin Perkins, Yanfang Zhang.
8. References 8. References
8.1. Normative References 8.1. Normative References
[RFC2119] Bradner, S., "Key words for use in RFCs to Indicate [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate
Requirement Levels", BCP 14, RFC 2119, Requirement Levels", BCP 14, RFC 2119,
DOI 10.17487/RFC2119, March 1997, <https://www.rfc- DOI 10.17487/RFC2119, March 1997, <https://www.rfc-
editor.org/info/rfc2119>. editor.org/info/rfc2119>.
skipping to change at page 10, line 35 skipping to change at page 11, line 35
unsigned integer value representing the effectiveness of stream unsigned integer value representing the effectiveness of stream
repair means. repair means.
o Measurement Point(s) with Potential Measurement Domain: It is o Measurement Point(s) with Potential Measurement Domain: It is
measured at the receiving end of the RTP stream. measured at the receiving end of the RTP stream.
o Measurement Timing: This metric relies on the sequence number o Measurement Timing: This metric relies on the sequence number
interval to determine measurement timing. interval to determine measurement timing.
o Use and Applications: These metrics are applicable to any RTP o Use and Applications: These metrics are applicable to any RTP
application, especially those that use loss-repair mechanisms. applications, especially those that use loss-repair mechanisms.
See Section 1 for details. See Section 1 for details.
o Reporting Model: See RFC 3611. o Reporting Model: See RFC 3611.
Authors' Addresses Authors' Addresses
Hui Zheng (Marvin) Hui Zheng (Marvin)
Huawei Individual
Email: marvin.zhenghui@huawei.com Email: zh4ui@huawei.comoutlook.com
Roni Even Roni Even
Huawei Huawei
Email: roni.even@huawei.com Email: roni.even@huawei.com
Qin Wu Qin Wu
Huawei Huawei
Email: bill.wu@huawei.com Email: bill.wu@huawei.com
Rong Gu Rong Gu
China Mobile China Mobile
Email: gurong_cmcc@outlook.com Email: gurong_cmcc@outlook.com
Rachel Huang
Huawei
Email: rachel.huang@huawei.com
 End of changes. 35 change blocks. 
86 lines changed or deleted 109 lines changed or added

This html diff was produced by rfcdiff 1.48. The latest version is available from http://tools.ietf.org/tools/rfcdiff/