draft-ietf-ippm-delay-var-as-02.txt   rfc5481.txt 
Network Working Group A. Morton Network Working Group A. Morton
Internet-Draft AT&T Labs Request for Comments: 5481 AT&T Labs
Intended status: Informational B. Claise Category: Informational B. Claise
Expires: July 12, 2009 Cisco Systems, Inc. Cisco Systems, Inc.
January 8, 2009
Packet Delay Variation Applicability Statement Packet Delay Variation Applicability Statement
draft-ietf-ippm-delay-var-as-02
Status of this Memo
This Internet-Draft is submitted to IETF in full conformance with the
provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF), its areas, and its working groups. Note that
other groups may also distribute working documents as Internet-
Drafts.
Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any
time. It is inappropriate to use Internet-Drafts as reference
material or to cite them other than as "work in progress."
The list of current Internet-Drafts can be accessed at
http://www.ietf.org/ietf/1id-abstracts.txt.
The list of Internet-Draft Shadow Directories can be accessed at Status of This Memo
http://www.ietf.org/shadow.html.
This Internet-Draft will expire on July 12, 2009. This memo provides information for the Internet community. It does
not specify an Internet standard of any kind. Distribution of this
memo is unlimited.
Copyright Notice Copyright Notice
Copyright (c) 2009 IETF Trust and the persons identified as the Copyright (c) 2009 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 in effect on the date of
(http://trustee.ietf.org/license-info) in effect on the date of publication of this document (http://trustee.ietf.org/license-info).
publication of this document. Please review these documents Please review these documents carefully, as they describe your rights
carefully, as they describe your rights and restrictions with respect and restrictions with respect to this document.
to this document.
This document may contain material from IETF Documents or IETF
Contributions published or made publicly available before November
10, 2008. The person(s) controlling the copyright in some of this
material may not have granted the IETF Trust the right to allow
modifications of such material outside the IETF Standards Process.
Without obtaining an adequate license from the person(s) controlling
the copyright in such materials, this document may not be modified
outside the IETF Standards Process, and derivative works of it may
not be created outside the IETF Standards Process, except to format
it for publication as an RFC or to translate it into languages other
than English.
Abstract Abstract
Packet delay variation metrics appear in many different standards Packet delay variation metrics appear in many different standards
documents. The metric definition in RFC 3393 has considerable documents. The metric definition in RFC 3393 has considerable
flexibility, and it allows multiple formulations of delay variation flexibility, and it allows multiple formulations of delay variation
through the specification of different packet selection functions. through the specification of different packet selection functions.
Although flexibility provides wide coverage and room for new ideas, Although flexibility provides wide coverage and room for new ideas,
it can make comparisons of independent implementations more it can make comparisons of independent implementations more
difficult. Two different formulations of delay variation have come difficult. Two different formulations of delay variation have come
into wide use in the context of active measurements. This memo into wide use in the context of active measurements. This memo
examines a range of circumstances for active measurements of delay examines a range of circumstances for active measurements of delay
variation and their uses, and recommends which of the two forms is variation and their uses, and recommends which of the two forms is
best matched to particular conditions and tasks. best matched to particular conditions and tasks.
Requirements Language
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 [RFC2119].
Table of Contents Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 5 1. Introduction ....................................................4
1.1. Background Literature in IPPM and Elsewhere . . . . . . . 6 1.1. Requirements Language ......................................5
1.2. Organization of the Memo . . . . . . . . . . . . . . . . . 7 1.2. Background Literature in IPPM and Elsewhere ................5
2. Purpose and Scope . . . . . . . . . . . . . . . . . . . . . . 7 1.3. Organization of the Memo ...................................6
3. Brief Descriptions of Delay Variation Uses . . . . . . . . . . 8 2. Purpose and Scope ...............................................7
3.1. Inferring Queue Occupation on a Path . . . . . . . . . . . 8 3. Brief Descriptions of Delay Variation Uses ......................7
3.2. Determining De-jitter Buffer Size . . . . . . . . . . . . 9 3.1. Inferring Queue Occupation on a Path .......................7
3.3. Spatial Composition . . . . . . . . . . . . . . . . . . . 10 3.2. Determining De-Jitter Buffer Size ..........................8
3.4. Service Level Comparison . . . . . . . . . . . . . . . . . 11 3.3. Spatial Composition .......................................10
3.5. Application-Layer FEC Design . . . . . . . . . . . . . . . 11 3.4. Service-Level Comparison ..................................10
4. Formulations of IPDV and PDV . . . . . . . . . . . . . . . . . 11 3.5. Application-Layer FEC Design ..............................10
4.1. IPDV: Inter-Packet Delay Variation . . . . . . . . . . . . 11 4. Formulations of IPDV and PDV ...................................10
4.2. PDV: Packet Delay Variation . . . . . . . . . . . . . . . 12 4.1. IPDV: Inter-Packet Delay Variation ........................11
4.3. A "Point" about Measurement Points . . . . . . . . . . . . 12 4.2. PDV: Packet Delay Variation ...............................11
4.4. Examples and Initial Comparisons . . . . . . . . . . . . . 13 4.3. A "Point" about Measurement Points ........................12
5. Survey of Earlier Comparisons . . . . . . . . . . . . . . . . 14 4.4. Examples and Initial Comparisons ..........................12
5.1. Demichelis' Comparison . . . . . . . . . . . . . . . . . . 14 5. Survey of Earlier Comparisons ..................................13
5.2. Ciavattone et al. . . . . . . . . . . . . . . . . . . . . 15 5.1. Demichelis' Comparison ....................................13
5.3. IPPM List Discussion from 2000 . . . . . . . . . . . . . . 16 5.2. Ciavattone et al. .........................................15
5.4. Y.1540 Appendix II . . . . . . . . . . . . . . . . . . . . 18 5.3. IPPM List Discussion from 2000 ............................16
5.5. Clark's ITU-T SG 12 Contribution . . . . . . . . . . . . . 18 5.4. Y.1540 Appendix II ........................................18
6. Additional Properties and Comparisons . . . . . . . . . . . . 18 5.5. Clark's ITU-T SG 12 Contribution ..........................18
6.1. Packet Loss . . . . . . . . . . . . . . . . . . . . . . . 18 6. Additional Properties and Comparisons ..........................18
6.2. Path Changes . . . . . . . . . . . . . . . . . . . . . . . 19 6.1. Packet Loss ...............................................18
6.2.1. Lossless Path Change . . . . . . . . . . . . . . . . . 20 6.2. Path Changes ..............................................19
6.2.2. Path Change with Loss . . . . . . . . . . . . . . . . 21 6.2.1. Lossless Path Change ...............................20
6.3. Clock Stability and Error . . . . . . . . . . . . . . . . 22 6.2.2. Path Change with Loss ..............................21
6.4. Spatial Composition . . . . . . . . . . . . . . . . . . . 24 6.3. Clock Stability and Error .................................22
6.5. Reporting a Single Number (SLA) . . . . . . . . . . . . . 24 6.4. Spatial Composition .......................................24
6.6. Jitter in RTCP Reports . . . . . . . . . . . . . . . . . . 25 6.5. Reporting a Single Number (SLA) ...........................24
6.7. MAPDV2 . . . . . . . . . . . . . . . . . . . . . . . . . . 25 6.6. Jitter in RTCP Reports ....................................25
6.8. Load Balancing . . . . . . . . . . . . . . . . . . . . . . 26 6.7. MAPDV2 ....................................................25
6.8. Load Balancing ............................................26
7. Applicability of the Delay Variation Forms and 7. Applicability of the Delay Variation Forms and
Recommendations . . . . . . . . . . . . . . . . . . . . . . . 27 Recommendations ................................................27
7.1. Uses . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 7.1. Uses ......................................................27
7.1.1. Inferring Queue Occupancy . . . . . . . . . . . . . . 27 7.1.1. Inferring Queue Occupancy ..........................27
7.1.2. Determining De-jitter Buffer Size (and FEC Design) . . 27 7.1.2. Determining De-Jitter Buffer Size (and FEC
7.1.3. Spatial Composition . . . . . . . . . . . . . . . . . 28 Design) ............................................27
7.1.4. Service Level Specification: Reporting a Single 7.1.3. Spatial Composition ................................28
Number . . . . . . . . . . . . . . . . . . . . . . . . 28 7.1.4. Service-Level Specification: Reporting a
7.2. Challenging Circumstances . . . . . . . . . . . . . . . . 28 Single Number ......................................28
7.2.1. Clock and Storage Issues . . . . . . . . . . . . . . . 28 7.2. Challenging Circumstances .................................28
7.2.2. Frequent Path Changes . . . . . . . . . . . . . . . . 29 7.2.1. Clock and Storage Issues ...........................28
7.2.3. Frequent Loss . . . . . . . . . . . . . . . . . . . . 29 7.2.2. Frequent Path Changes ..............................29
7.2.4. Load Balancing . . . . . . . . . . . . . . . . . . . . 29 7.2.3. Frequent Loss ......................................29
7.3. Summary . . . . . . . . . . . . . . . . . . . . . . . . . 29 7.2.4. Load Balancing .....................................29
8. Measurement Considerations . . . . . . . . . . . . . . . . . . 30 7.3. Summary ...................................................30
8.1. Measurement Stream Characteristics . . . . . . . . . . . . 31 8. Measurement Considerations .....................................31
8.2. Measurement Devices . . . . . . . . . . . . . . . . . . . 32 8.1. Measurement Stream Characteristics ........................31
8.3. Units of Measurement . . . . . . . . . . . . . . . . . . . 32 8.2. Measurement Devices .......................................32
8.4. Test Duration . . . . . . . . . . . . . . . . . . . . . . 33 8.3. Units of Measurement ......................................33
8.5. Clock Sync Options . . . . . . . . . . . . . . . . . . . . 33 8.4. Test Duration .............................................33
8.6. Distinguishing Long Delay from Loss . . . . . . . . . . . 33 8.5. Clock Sync Options ........................................33
8.7. Accounting for Packet Reordering . . . . . . . . . . . . . 34 8.6. Distinguishing Long Delay from Loss .......................34
8.8. Results Representation and Reporting . . . . . . . . . . . 34 8.7. Accounting for Packet Reordering ..........................34
9. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 35 8.8. Results Representation and Reporting ......................35
10. Security Considerations . . . . . . . . . . . . . . . . . . . 35 9. Security Considerations ........................................35
11. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 35 10. Acknowledgments ...............................................35
12. Appendix on Calculating the D(min) in PDV . . . . . . . . . . 35 11. Appendix on Calculating the D(min) in PDV .....................35
13. References . . . . . . . . . . . . . . . . . . . . . . . . . . 36 12. References ....................................................36
13.1. Normative References . . . . . . . . . . . . . . . . . . . 36 12.1. Normative References .....................................36
13.2. Informative References . . . . . . . . . . . . . . . . . . 37 12.2. Informative References ...................................37
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 38
1. Introduction 1. Introduction
There are many ways to formulate packet delay variation metrics for There are many ways to formulate packet delay variation metrics for
the Internet and other packet-based networks. The IETF itself has the Internet and other packet-based networks. The IETF itself has
several specifications for delay variation [RFC3393], sometimes several specifications for delay variation [RFC3393], sometimes
called jitter [RFC3550] or even inter-arrival jitter [RFC3550], and called jitter [RFC3550] or even inter-arrival jitter [RFC3550], and
these have achieved wide adoption. The International these have achieved wide adoption. The International
Telecommunication Union - Telecommunication Standardization Sector Telecommunication Union - Telecommunication Standardization Sector
(ITU-T) has also recommended several delay variation metrics (called (ITU-T) has also recommended several delay variation metrics (called
skipping to change at page 5, line 35 skipping to change at page 4, line 35
broader than packet transfer performance, with "unwanted signal broader than packet transfer performance, with "unwanted signal
variation" as a general definition. "Jitter" has been used to variation" as a general definition. "Jitter" has been used to
describe frequency or phase variations, such as data stream rate describe frequency or phase variations, such as data stream rate
variations or carrier signal phase noise. The phrase "delay variations or carrier signal phase noise. The phrase "delay
variation" is almost self-defining and more precise, so it is variation" is almost self-defining and more precise, so it is
preferred in this memo. preferred in this memo.
Most (if not all) delay variation metrics are derived metrics, in Most (if not all) delay variation metrics are derived metrics, in
that their definitions rely on another fundamental metric. In this that their definitions rely on another fundamental metric. In this
case, the fundamental metric is one-way delay, and variation is case, the fundamental metric is one-way delay, and variation is
assessed by computing the difference between two individual one-way assessed by computing the difference between two individual one-way-
delay measurements, or a pair of singletons. One of the delay delay measurements, or a pair of singletons. One of the delay
singletons is taken as a reference, and the result is the variation singletons is taken as a reference, and the result is the variation
with respect to the reference. The variation is usually summarized with respect to the reference. The variation is usually summarized
for all packets in a stream using statistics. for all packets in a stream using statistics.
The industry has predominantly implemented two specific formulations The industry has predominantly implemented two specific formulations
of delay variation (for one survey of the situation, of delay variation (for one survey of the situation, see
see[Krzanowski]): [Krzanowski]):
1. Inter-Packet Delay Variation, IPDV, where the reference is the 1. Inter-Packet Delay Variation, IPDV, where the reference is the
previous packet in the stream (according to sending sequence), previous packet in the stream (according to sending sequence),
and the reference changes for each packet in the stream. and the reference changes for each packet in the stream.
Properties of variation are coupled with packet sequence in this Properties of variation are coupled with packet sequence in this
formulation. This form was called Instantaneous Packet Delay formulation. This form was called Instantaneous Packet Delay
Variation in early IETF contributions, and is similar to the Variation in early IETF contributions, and is similar to the
packet spacing difference metric used for interarrival jitter packet spacing difference metric used for interarrival jitter
calculations in [RFC3550]. calculations in [RFC3550].
skipping to change at page 6, line 22 skipping to change at page 5, line 22
It is important to note that the authors of relevant standards for It is important to note that the authors of relevant standards for
delay variation recognized there are many different users with delay variation recognized there are many different users with
varying needs, and allowed sufficient flexibility to formulate varying needs, and allowed sufficient flexibility to formulate
several metrics with different properties. Therefore, the comparison several metrics with different properties. Therefore, the comparison
is not so much between standards bodies or their specifications as it is not so much between standards bodies or their specifications as it
is between specific formulations of delay variation. Both Inter- is between specific formulations of delay variation. Both Inter-
Packet Delay Variation and Packet Delay Variation are compliant with Packet Delay Variation and Packet Delay Variation are compliant with
[RFC3393], because different packet selection functions will produce [RFC3393], because different packet selection functions will produce
either form. either form.
1.1. Background Literature in IPPM and Elsewhere 1.1. Requirements Language
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 [RFC2119].
1.2. Background Literature in IPPM and Elsewhere
With more people joining the measurement community every day, it is With more people joining the measurement community every day, it is
possible this memo is the first from the IP Performance Metrics possible this memo is the first from the IP Performance Metrics
(IPPM) Working Group that the reader has consulted. This section (IPPM) Working Group that the reader has consulted. This section
provides a brief roadmap and background on the IPPM literature, and provides a brief roadmap and background on the IPPM literature, and
the published specifications of other relevant standards the published specifications of other relevant standards
organizations. organizations.
The IPPM framework [RFC2330] provides a background for this memo and The IPPM framework [RFC2330] provides a background for this memo and
other IPPM RFCs. Key terms such as singleton, sample, and statistic other IPPM RFCs. Key terms such as singleton, sample, and statistic
are defined there, along with methods of collecting samples (Poisson are defined there, along with methods of collecting samples (Poisson
streams), time related issues, and the "packet of Type-P" convention. streams), time-related issues, and the "packet of Type-P" convention.
There are two fundamental and related metrics that can be applied to There are two fundamental and related metrics that can be applied to
every packet transfer attempt: one-way loss [RFC2680] and one-way every packet transfer attempt: one-way loss [RFC2680] and one-way
delay [RFC2679]. The metrics use a waiting time threshold to delay [RFC2679]. The metrics use a waiting time threshold to
distinguish between lost and delayed packets. Packets that arrive at distinguish between lost and delayed packets. Packets that arrive at
the measurement destination within their waiting time have finite the measurement destination within their waiting time have finite
delay and are not lost. Otherwise, packets are designated lost and delay and are not lost. Otherwise, packets are designated lost and
their delay is undefined. Guidance on setting the waiting time their delay is undefined. Guidance on setting the waiting time
threshold may be found in [RFC2680] and threshold may be found in [RFC2680] and [IPPM-Reporting].
[I-D.morton-ippm-reporting-metrics].
Another fundamental metric is packet reordering as specified in Another fundamental metric is packet reordering as specified in
[RFC4737]. The reordering metric was defined to be "orthogonal" to [RFC4737]. The reordering metric was defined to be "orthogonal" to
packet loss. In other words, the gap in a packet sequence caused by packet loss. In other words, the gap in a packet sequence caused by
loss does not result in reordered packets, but a re-arrangement of loss does not result in reordered packets, but a rearrangement of
packet arrivals from their sending order constitutes reordering. packet arrivals from their sending order constitutes reordering.
Derived metrics are based on the fundamental metrics. The metric of Derived metrics are based on the fundamental metrics. The metric of
primary interest here is delay variation [RFC3393], a metric which is primary interest here is delay variation [RFC3393], a metric that is
derived from one-way delay [RFC2680]. Another derived metric is the derived from one-way delay [RFC2680]. Another derived metric is the
loss patterns metric [RFC3357], which is derived from loss. loss patterns metric [RFC3357], which is derived from loss.
The measured values of all metrics (both fundamental and derived) The measured values of all metrics (both fundamental and derived)
depend to great extent on the stream characteristics used to collect depend to great extent on the stream characteristics used to collect
them. Both Poisson streams [RFC3393] and Periodic streams [RFC3432] them. Both Poisson streams [RFC3393] and Periodic streams [RFC3432]
have been used with the IPDV and PDV metrics. The choice of stream have been used with the IPDV and PDV metrics. The choice of stream
specifications for active measurement will depend on the purpose of specification for active measurement will depend on the purpose of
the characterization and the constraints of the testing environment. the characterization and the constraints of the testing environment.
Periodic streams are frequently chosen for use with IPDV and PDV, Periodic streams are frequently chosen for use with IPDV and PDV,
because the application streams that are most sensitive to delay because the application streams that are most sensitive to delay
variation exhibit periodicity. Additional details that are method- variation exhibit periodicity. Additional details that are method-
specific are discussed the section on Measurement Considerations. specific are discussed in Section 8 on "Measurement Considerations".
In the ITU-T, the framework, fundamental metrics and derived metrics In the ITU-T, the framework, fundamental metrics, and derived metrics
for IP performance are specified in Recommendation Y.1540 [Y.1540]. for IP performance are specified in Recommendation Y.1540 [Y.1540].
[G.1020] defines additional delay variation metrics, analyses the [G.1020] defines additional delay variation metrics, analyzes the
operation of fixed and adaptive de-jitter buffers, and describes an operation of fixed and adaptive de-jitter buffers, and describes an
example adaptive de-jitter buffer emulator. Appendix II of [G.1050] example adaptive de-jitter buffer emulator. Appendix II of [G.1050]
describes the models for network impairments (including delay describes the models for network impairments (including delay
variation) that are part of standardized IP network emulator which variation) that are part of standardized IP network emulator that may
may be useful when evaluating measurement techniques. be useful when evaluating measurement techniques.
1.2. Organization of the Memo 1.3. Organization of the Memo
The Purpose and Scope follows in Section 2. We then give a summary The Purpose and Scope follows in Section 2. We then give a summary
of the main tasks for delay variation metrics in section 3. Section of the main tasks for delay variation metrics in Section 3.
4 defines the two primary forms of delay variation, and section 5 Section 4 defines the two primary forms of delay variation, and
presents summaries of four earlier comparisons. Section 6 adds new Section 5 presents summaries of four earlier comparisons. Section 6
comparisons to the analysis, and section 7 reviews the applicability adds new comparisons to the analysis, and Section 7 reviews the
and recommendations for each form of delay variation. Section 8 then applicability and recommendations for each form of delay variation.
looks at many important delay variation measurement considerations. Section 8 then looks at many important delay variation measurement
Following the IANA and Security Considerations, there is an Appendix considerations. Following the Security Considerations, there is an
on the calculation of the minimum delay for the PDV form. appendix on the calculation of the minimum delay for the PDV form.
2. Purpose and Scope 2. Purpose and Scope
The IPDV and PDV formulations have certain features that make them The IPDV and PDV formulations have certain features that make them
more suitable for one circumstance and less so for another. The more suitable for one circumstance and less so for another. The
purpose of this memo is to compare two forms of delay variation, so purpose of this memo is to compare two forms of delay variation, so
that it will be evident which of the two is better suited for each of that it will be evident which of the two is better suited for each of
many possible uses and their related circumstances. many possible uses and their related circumstances.
The scope of this memo is limited to the two forms of delay variation The scope of this memo is limited to the two forms of delay variation
skipping to change at page 8, line 48 skipping to change at page 8, line 11
each queue varies, depending on the number of packets in the queue each queue varies, depending on the number of packets in the queue
when a particular packet arrives. If one assumes that at least one when a particular packet arrives. If one assumes that at least one
of the packets in a test stream encounters virtually empty queues all of the packets in a test stream encounters virtually empty queues all
along the path (and the path is stable), then the additional delay along the path (and the path is stable), then the additional delay
observed on other packets can be attributed to the time spent in one observed on other packets can be attributed to the time spent in one
or more queues. Otherwise, the delay variation observed is the or more queues. Otherwise, the delay variation observed is the
variation in queue time experienced by the test stream. variation in queue time experienced by the test stream.
It is worth noting that delay variation can occur beyond IP router It is worth noting that delay variation can occur beyond IP router
queues, in other communication components. Examples include media queues, in other communication components. Examples include media
contention: DOCSIS, IEEE 802.11 and some mobile radio technologies. contention: DOCSIS, IEEE 802.11, and some mobile radio technologies.
However, delay variation from all sources at the IP layer and below However, delay variation from all sources at the IP layer and below
will be quantified using the two formulations discussed here. will be quantified using the two formulations discussed here.
3.2. Determining De-jitter Buffer Size 3.2. Determining De-Jitter Buffer Size
Note - while this memo and other IPPM literature prefer the term Note -- while this memo and other IPPM literature prefer the term
delay variation, the terms "jitter buffer" and the more accurate "de- "delay variation", the terms "jitter buffer" and the more accurate
jitter buffer" are widely adopted names for a component of packet "de-jitter buffer" are widely adopted names for a component of packet
communication systems, and they will be used here to designate that communication systems, and they will be used here to designate that
system component. system component.
Most Isochronous applications (a.k.a. real-time applications) employ Most isochronous applications (a.k.a. real-time applications) employ
a buffer to smooth out delay variation encountered on the path from a buffer to smooth out delay variation encountered on the path from
source to destination. The buffer must be big enough to accommodate source to destination. The buffer must be big enough to accommodate
the expected variation of delay, or packet loss will result. the expected variation of delay, or packet loss will result.
However, if the buffer is too large, then some of the desired However, if the buffer is too large, then some of the desired
spontaneity of communication will be lost and conversational dynamics spontaneity of communication will be lost and conversational dynamics
will be affected. Therefore, application designers need to know the will be affected. Therefore, application designers need to know the
range of delay variation they must accommodate, whether they are range of delay variation they must accommodate, whether they are
designing fixed or adaptive buffer systems. designing fixed or adaptive buffer systems.
Network service providers also attempt to constrain delay variation Network service providers also attempt to constrain delay variation
skipping to change at page 9, line 40 skipping to change at page 8, line 51
stored. It is relatively simple to convert between octets and time stored. It is relatively simple to convert between octets and time
when the buffer read rate (in octets per second) is constant: when the buffer read rate (in octets per second) is constant:
read_rate * storage_time = storage_octets read_rate * storage_time = storage_octets
Units of time are used in the discussion below. Units of time are used in the discussion below.
The objective of a de-jitter buffer is to compensate for all prior The objective of a de-jitter buffer is to compensate for all prior
sources of delay variation and produce a packet stream with constant sources of delay variation and produce a packet stream with constant
delay. Thus, a packet experiencing the minimum transit delay from delay. Thus, a packet experiencing the minimum transit delay from
source to destination, D_min, should spend the maximum time in a de- source to destination, D_min, should spend the maximum time in a
jitter buffer, B_max. The sum of D_min and B_max should equal the de-jitter buffer, B_max. The sum of D_min and B_max should equal the
sum of the maximum transit delay (D_max) and the minimum buffer time sum of the maximum transit delay (D_max) and the minimum buffer time
(B_min). We have (B_min). We have
Constant = D_min + B_max = D_max + B_min, Constant = D_min + B_max = D_max + B_min,
after rearranging terms, after rearranging terms,
B_max - B_min = D_max - D_min = range(B) = range(D) B_max - B_min = D_max - D_min = range(B) = range(D)
where range(B) is the range of packet buffering times, and range(D) where range(B) is the range of packet buffering times, and range(D)
is the range of packet transit delays from source to destination. is the range of packet transit delays from source to destination.
Packets with transit delay between the max and min spend a Packets with transit delay between the max and min spend a
complimentary time in the buffer and also see the constant delay. complementary time in the buffer and also see the constant delay.
In practice, the minimum buffer time, B_min, may not be zero, and the In practice, the minimum buffer time, B_min, may not be zero, and the
maximum transit delay, D_max may be a high percentile (99.9%-ile) maximum transit delay, D_max, may be a high percentile (99.9th
instead of the maximum. percentile) instead of the maximum.
Note that B_max - B_min = range(B) is the range of buffering times Note that B_max - B_min = range(B) is the range of buffering times
needed to compensate for delay variation. The actual size of the needed to compensate for delay variation. The actual size of the
buffer may be larger (where B_min > 0) or smaller than range(B). buffer may be larger (where B_min > 0) or smaller than range(B).
There must be a process to align the de-jitter buffer time with There must be a process to align the de-jitter buffer time with
packet transit delay. This is a process to identify the packets with packet transit delay. This is a process to identify the packets with
minimum delay and schedule their play-out time so that they spend the minimum delay and schedule their play-out time so that they spend the
maximum time in the buffer. The error in the alignment process can maximum time in the buffer. The error in the alignment process can
be accounted for by a variable, A. In the equation below, the range be accounted for by a variable, A. In the equation below, the range
skipping to change at page 10, line 34 skipping to change at page 9, line 46
range(b) = b_max - b_min = D_max - D_min + A range(b) = b_max - b_min = D_max - D_min + A
where variable b represents the *available* buffer in a system with a where variable b represents the *available* buffer in a system with a
specific alignment, A, and b_max and b_min represent the limits of specific alignment, A, and b_max and b_min represent the limits of
the available buffer. the available buffer.
When A is positive, the de-jitter buffer applies more delay than When A is positive, the de-jitter buffer applies more delay than
necessary (where Constant = D_max+b_min+A represents one possible necessary (where Constant = D_max+b_min+A represents one possible
alignment). When A is negative, there is insufficient buffer time alignment). When A is negative, there is insufficient buffer time
available to compensate for range(D) because of mis-alignment. available to compensate for range(D) because of misalignment.
Packets with D_min may be arriving too early and encountering a full Packets with D_min may be arriving too early and encountering a full
buffer, or packets with D_max may be arriving too late, and in either buffer, or packets with D_max may be arriving too late, and in either
case the packets would be discarded. case, the packets would be discarded.
In summary, the range of transit delay variation is a critical factor In summary, the range of transit delay variation is a critical factor
in the determination of de-jitter buffer size. in the determination of de-jitter buffer size.
3.3. Spatial Composition 3.3. Spatial Composition
In Spatial Composition, the tasks are similar to those described In Spatial Composition, the tasks are similar to those described
above, but with the additional complexity of a multiple network path above, but with the additional complexity of a multiple network path
where several sub-paths are measured separately and no source to where several sub-paths are measured separately and no source-to-
destination measurements are available. In this case, the source to destination measurements are available. In this case, the source-to-
destination performance must be estimated, using Composed Metrics as destination performance must be estimated, using Composed Metrics as
described in [I-D.ietf-ippm-framework-compagg] and [Y.1541]. Note described in [IPPM-Framework] and [Y.1541]. Note that determining
that determining the composite delay variation is not trivial: simply the composite delay variation is not trivial: simply summing the sub-
summing the sub-path variations is not accurate. path variations is not accurate.
3.4. Service Level Comparison 3.4. Service-Level Comparison
IP performance measurements are often used as the basis for IP performance measurements are often used as the basis for
agreements (or contracts) between service providers and their agreements (or contracts) between service providers and their
customers. The measurement results must compare favorably with the customers. The measurement results must compare favorably with the
performance levels specified in the agreement. performance levels specified in the agreement.
Packet delay variation is usually one of the metrics specified in Packet delay variation is usually one of the metrics specified in
these agreements. In principle, any formulation could be specified these agreements. In principle, any formulation could be specified
in the Service Level Agreement (SLA). However, the SLA is most in the Service Level Agreement (SLA). However, the SLA is most
useful when the measured quantities can be related to ways in which useful when the measured quantities can be related to ways in which
the communication service will be utilized by the customer, and this the communication service will be utilized by the customer, and this
can usually be derived from one of the tasks described above. can usually be derived from one of the tasks described above.
3.5. Application-Layer FEC Design 3.5. Application-Layer FEC Design
The design of application-layer Forward Error Correction (FEC) The design of application-layer Forward Error Correction (FEC)
components is closely related to the design of a de-jitter buffer in components is closely related to the design of a de-jitter buffer in
several ways. The FEC designer must choose a protection interval several ways. The FEC designer must choose a protection interval
(time to send/receive a block of packets in a constant packet rate (time to send/receive a block of packets in a constant packet rate
system) consistent with the packet loss characteristics, but also system) consistent with the packet-loss characteristics, but also
mindful of the extent of delay variation expected. Further, the mindful of the extent of delay variation expected. Further, the
system designer must decide how long to wait for "late" packets to system designer must decide how long to wait for "late" packets to
arrive. Again, the range of delay variation is the relevant arrive. Again, the range of delay variation is the relevant
expression delay variation for these tasks. expression delay variation for these tasks.
4. Formulations of IPDV and PDV 4. Formulations of IPDV and PDV
This section presents the formulations of IPDV and PDV, and provides This section presents the formulations of IPDV and PDV, and provides
some illustrative examples. We use the basic singleton definition in some illustrative examples. We use the basic singleton definition in
[RFC3393] (which itself is based on [RFC2679]): [RFC3393] (which itself is based on [RFC2679]):
"Type-P-One-way-ipdv is defined for two packets from Src to Dst "Type-P-One-way-ipdv is defined for two packets from Src to Dst
selected by the selection function F, as the difference between the selected by the selection function F, as the difference between the
value of the Type-P-One-way-delay from Src to Dst at T2 and the value value of the Type-P-One-way-delay from Src to Dst at T2 and the value
of the Type-P-One-Way-Delay from Src to Dst at T1." of the Type-P-One-Way-Delay from Src to Dst at T1".
4.1. IPDV: Inter-Packet Delay Variation 4.1. IPDV: Inter-Packet Delay Variation
If we have packets in a stream consecutively numbered i = 1,2,3,... If we have packets in a stream consecutively numbered i = 1,2,3,...
falling within the test interval, then IPDV(i) = D(i)-D(i-1) where falling within the test interval, then IPDV(i) = D(i)-D(i-1) where
D(i) denotes the one-way-delay of the ith packet of a stream. D(i) denotes the one-way delay of the ith packet of a stream.
One-way delays are the difference between timestamps applied at the One-way delays are the difference between timestamps applied at the
ends of the path, or the receiver time minus the transmission time. ends of the path, or the receiver time minus the transmission time.
So D(2) = R2-T2. With this timestamp notation, it can be shown that So D(2) = R2-T2. With this timestamp notation, it can be shown that
IPDV also represents the change in inter-packet spacing between IPDV also represents the change in inter-packet spacing between
transmission and reception: transmission and reception:
IPDV(2) = D(2) - D(1) = (R2-T2) - (R1-T1) = (R2-R1) - (T2-T1) IPDV(2) = D(2) - D(1) = (R2-T2) - (R1-T1) = (R2-R1) - (T2-T1)
An example selection function given in [RFC3393] is "Consecutive An example selection function given in [RFC3393] is "Consecutive
Type-P packets within the specified interval." This is exactly the Type-P packets within the specified interval". This is exactly the
function needed for IPDV. The reference packet in the pair is always function needed for IPDV. The reference packet in the pair is the
the previous packet in the sending sequence. previous packet in the sending sequence.
Note that IPDV can take on positive and negative values (and zero). Note that IPDV can take on positive and negative values (and zero).
One way to analyze the IPDV results is to concentrate on the positive One way to analyze the IPDV results is to concentrate on the positive
excursions. However, this approach has limitations that are excursions. However, this approach has limitations that are
discussed in more detail below (see section 5.3). discussed in more detail below (see Section 5.3).
The mean of all IPDV(i) for a stream is usually zero. However, a The mean of all IPDV(i) for a stream is usually zero. However, a
slow delay change over the life of the stream, or a frequency error slow delay change over the life of the stream, or a frequency error
between the measurement system clocks, can result in a non-zero mean. between the measurement system clocks, can result in a non-zero mean.
4.2. PDV: Packet Delay Variation 4.2. PDV: Packet Delay Variation
The name Packet Delay Variation is used in [Y.1540] and its The name Packet Delay Variation is used in [Y.1540] and its
predecessors, and refers to a performance parameter equivalent to the predecessors, and refers to a performance parameter equivalent to the
metric described below. metric described below.
The Selection Function for PDV requires two specific roles for the The Selection Function for PDV requires two specific roles for the
packets in the pair. The first packet is any Type-P packet within packets in the pair. The first packet is any Type-P packet within
the specified interval. The second, or reference packet is the the specified interval. The second, or reference packet is the
Type-P packet within the specified interval with the minimum one-way- Type-P packet within the specified interval with the minimum one-way
delay. delay.
Therefore, PDV(i) = D(i)-D(min) (using the nomenclature introduced in Therefore, PDV(i) = D(i)-D(min) (using the nomenclature introduced in
the IPDV section). D(min) is the delay of the packet with the lowest the IPDV section). D(min) is the delay of the packet with the lowest
value for delay (minimum) over the current test interval. Values of value for delay (minimum) over the current test interval. Values of
PDV may be zero or positive, and quantiles of the PDV distribution PDV may be zero or positive, and quantiles of the PDV distribution
are direct indications of delay variation. are direct indications of delay variation.
PDV is a version of the one-way delay distribution, shifted to the PDV is a version of the one-way-delay distribution, shifted to the
origin by normalizing to the minimum delay. origin by normalizing to the minimum delay.
4.3. A "Point" about Measurement Points 4.3. A "Point" about Measurement Points
Both IPDV and PDV are derived from the one-way delay metric. One way Both IPDV and PDV are derived from the one-way-delay metric. One-way
delay requires knowledge of time at two points, e.g., the source and delay requires knowledge of time at two points, e.g., the source and
destination of an IP network path in end-to-end measurement. destination of an IP network path in end-to-end measurement.
Therefore, both IPDV and PDV can be categorized as 2-point metrics Therefore, both IPDV and PDV can be categorized as 2-point metrics
because they are derived from one-way delay. Specific methods of because they are derived from one-way delay. Specific methods of
measurement may make assumptions or have a priori knowledge about one measurement may make assumptions or have a priori knowledge about one
of the measurement points, but the metric definitions themselves are of the measurement points, but the metric definitions themselves are
based on information collected at two measurement points. based on information collected at two measurement points.
4.4. Examples and Initial Comparisons 4.4. Examples and Initial Comparisons
Note: This material originally presented in slides 2 and 3 of Note: This material originally presented in Slides 2 and 3 of
[Morton06]. [Morton06].
The Figure below gives a sample of packet delays and calculates IPDV The Figure below gives a sample of packet delays, calculates IPDV and
and PDV values and depicts a histogram for each one. PDV values, and depicts a histogram for each one.
Packet # 1 2 3 4 5 Packet # 1 2 3 4 5
------------------------------- -------------------------------
Delay, ms 20 10 20 25 20 Delay, ms 20 10 20 25 20
IPDV U -10 10 5 -5 IPDV U -10 10 5 -5
PDV 10 0 10 15 10 PDV 10 0 10 15 10
| | | |
skipping to change at page 13, line 39 skipping to change at page 13, line 30
H H 1| H H 1|H H H H H 1| H H 1|H H H
H H | H H |H H H H H | H H |H H H
---------+-------- +--------------- ---------+-------- +---------------
-10 -5 0 5 10 0 5 10 15 -10 -5 0 5 10 0 5 10 15
IPDV Histogram PDV Histogram IPDV Histogram PDV Histogram
Figure 1: IPDV and PDV Comparison Figure 1: IPDV and PDV Comparison
The sample of packets contains three packets with "typical" delays of The sample of packets contains three packets with "typical" delays of
20ms, one packet with a low delay of 10ms (the minimum of the sample) 20 ms, one packet with a low delay of 10 ms (the minimum of the
and one packet with 25ms delay. sample) and one packet with 25 ms delay.
As noted above, this example illustrates that IPDV may take on As noted above, this example illustrates that IPDV may take on
positive and negative values, while the PDV values are greater than positive and negative values, while the PDV values are greater than
or equal to zero. The Histograms of IPDV and PDV are quite different or equal to zero. The histograms of IPDV and PDV are quite different
in general shape, and the ranges are different, too (IPDV range = in general shape, and the ranges are different, too (IPDV range =
20ms, PDV range = 15 ms). Note that the IPDV histogram will change 20ms, PDV range = 15 ms). Note that the IPDV histogram will change
if the sequence of delays is modified, but the PDV histogram will if the sequence of delays is modified, but the PDV histogram will
stay the same. PDV normalizes the one-way delay distribution to the stay the same. PDV normalizes the one-way-delay distribution to the
minimum delay and emphasizes the variation independent from the minimum delay and emphasizes the variation independent from the
sequence of delays. sequence of delays.
5. Survey of Earlier Comparisons 5. Survey of Earlier Comparisons
This section summarizes previous work to compare these two forms of This section summarizes previous work to compare these two forms of
delay variation. delay variation.
5.1. Demichelis' Comparison 5.1. Demichelis' Comparison
In [Demichelis], Demichelis compared the early draft versions of two In [Demichelis], Demichelis compared the early versions of two forms
forms of delay variation. Although the IPDV form would eventually of delay variation. Although the IPDV form would eventually see
see widespread use, the ITU-T work-in-progress he cited did not widespread use, the ITU-T work-in-progress he cited did not utilize
utilize the same reference packets as PDV. Demichelis compared IPDV the same reference packets as PDV. Demichelis compared IPDV with the
with the alternatives of using the delay of the first packet in the alternatives of using the delay of the first packet in the stream and
stream and the mean delay of the stream as the PDV reference packet. the mean delay of the stream as the PDV reference packet. Neither of
Neither of these alternative references were used in practice, and these alternative references were used in practice, and they are now
they are now deprecated in favor of the minimum delay of the stream deprecated in favor of the minimum delay of the stream [Y.1540].
[Y.1540].
Active measurements of a transcontinental path (Torino to Tokyo) Active measurements of a transcontinental path (Torino to Tokyo)
provided the data for the comparison. The Poisson test stream had provided the data for the comparison. The Poisson test stream had
0.764 second average inter-packet interval, with more than 58 0.764 second average inter-packet interval, with more than 58
thousand packets over 13.5 hours. Among Demichelis' observations thousand packets over 13.5 hours. Among Demichelis' observations
about IPDV are the following: about IPDV are the following:
1. IPDV is a measure of the network's ability to preserve the 1. IPDV is a measure of the network's ability to preserve the
spacing between packets. spacing between packets.
2. The distribution of IPDV is usually symmetrical about the origin, 2. The distribution of IPDV is usually symmetrical about the origin,
having a balance of negative and positive values (for the most having a balance of negative and positive values (for the most
part). The mean is usually zero, unless some long-term delay part). The mean is usually zero, unless some long-term delay
trend is present. trend is present.
3. IPDV singletons distinguish quick delay variations (short-term, 3. IPDV singletons distinguish quick-delay variations (short-term,
on the order of the interval between packets) from longer term on the order of the interval between packets) from longer-term
variations. variations.
4. IPDV places reduced demands on the stability and skew of 4. IPDV places reduced demands on the stability and skew of
measurement clocks. measurement clocks.
He also notes these features of PDV: He also notes these features of PDV:
1. The PDV distribution does not distinguish short-term variation 1. The PDV distribution does not distinguish short-term variation
from variation over the complete test interval. (Comment: PDV from variation over the complete test interval. (Comment: PDV
can be determined over any sub-intervals when the singletons are can be determined over any sub-intervals when the singletons are
skipping to change at page 15, line 32 skipping to change at page 15, line 24
exceeding a +/-50ms range around the mean, the Inverse Percentiles, exceeding a +/-50ms range around the mean, the Inverse Percentiles,
and the Inter-Quartile Range. and the Inter-Quartile Range.
5.2. Ciavattone et al. 5.2. Ciavattone et al.
In [Cia03], the authors compared IPDV and PDV (referred to as delta) In [Cia03], the authors compared IPDV and PDV (referred to as delta)
using a periodic packet stream conforming to [RFC3432] with inter- using a periodic packet stream conforming to [RFC3432] with inter-
packet interval of 20 ms. packet interval of 20 ms.
One of the comparisons between IPDV and PDV involves a laboratory One of the comparisons between IPDV and PDV involves a laboratory
set-up where a queue was temporarily congested by a competing packet setup where a queue was temporarily congested by a competing packet
burst. The additional queuing delay was 85ms to 95ms, much larger burst. The additional queuing delay was 85ms to 95ms, much larger
than the inter-packet interval. The first packet in the stream that than the inter-packet interval. The first packet in the stream that
follows the competing burst spends the longest time queued, and follows the competing burst spends the longest time queued, and
others experience less and less queuing time until the queue is others experience less and less queuing time until the queue is
drained. drained.
The authors observed that PDV reflects the additional queuing time of The authors observed that PDV reflects the additional queuing time of
the packets affected by the burst, with values of 85, 65, 45, 25, and the packets affected by the burst, with values of 85, 65, 45, 25, and
5ms. Also, it is easy to determine (by looking at the PDV range) 5ms. Also, it is easy to determine (by looking at the PDV range)
that a de-jitter buffer of >85 ms would have been sufficient to that a de-jitter buffer of >85 ms would have been sufficient to
skipping to change at page 16, line 10 skipping to change at page 16, line 6
85, -20, -20, -20, -20, -5ms. Only the positive excursion of IPDV 85, -20, -20, -20, -20, -5ms. Only the positive excursion of IPDV
gives an indication of the de-jitter buffer size needed. Although gives an indication of the de-jitter buffer size needed. Although
the variation exceeds the inter-packet interval, the extent of the variation exceeds the inter-packet interval, the extent of
negative IPDV values is limited by that sending interval. This negative IPDV values is limited by that sending interval. This
preference for information from the positive IPDV values has prompted preference for information from the positive IPDV values has prompted
some to ignore the negative values, or to take the absolute value of some to ignore the negative values, or to take the absolute value of
each IPDV measurement (sacrificing key properties of IPDV in the each IPDV measurement (sacrificing key properties of IPDV in the
process, such as its ability to distinguish delay trends). process, such as its ability to distinguish delay trends).
Note that this example illustrates a case where the IPDV distribution Note that this example illustrates a case where the IPDV distribution
is asymmetrical, because the delay variation range (85ms) exceeds the is asymmetrical, because the delay variation range (85 ms) exceeds
inter-packet spacing (20ms). We see that the IPDV values 85, -20, the inter-packet spacing (20 ms). We see that the IPDV values 85,
-20, -20, -20, -5ms have zero mean, but the left side of the -20, -20, -20, -20, -5 ms have zero mean, but the left side of the
distribution is truncated at -20ms. distribution is truncated at -20ms.
Elsewhere, the authors considered the range as a summary statistic Elsewhere in the article, the authors considered the range as a
for IPDV, and the 99.9%-ile minus the minimum delay as a summary summary statistic for IPDV, and the 99.9th percentile minus the
statistic for delay variation, or PDV. minimum delay as a summary statistic for delay variation, or PDV.
5.3. IPPM List Discussion from 2000 5.3. IPPM List Discussion from 2000
Mike Pierce made many comments in the context of the 05 version of Mike Pierce made many comments in the context of a working version of
draft-ietf-ippm-ipdv. One of his main points was that a delay [RFC3393]. One of his main points was that a delay histogram is a
histogram is a useful approach to quantifying variation. Another useful approach to quantifying variation. Another point was that the
point was that the time duration of evaluation is a critical aspect. time duration of evaluation is a critical aspect.
Carlo Demichelis then mailed his comparison paper to the IPPM list, Carlo Demichelis then mailed his comparison paper [Demichelis] to the
[Demichelis] as discussed in more detail above. IPPM list, as discussed in more detail above.
Ruediger Geib observed that both IPDV and the delay histogram (PDV) Ruediger Geib observed that both IPDV and the delay histogram (PDV)
are useful, and suggested that they might be applied to different are useful, and suggested that they might be applied to different
variation time scales. He pointed out that loss has a significant variation time scales. He pointed out that loss has a significant
effect on IPDV, and encouraged that the loss information be retained effect on IPDV, and encouraged that the loss information be retained
in the arrival sequence. in the arrival sequence.
Several example delay variation scenarios were discussed, including: Several example delay variation scenarios were discussed, including:
Packet # 1 2 3 4 5 6 7 8 9 10 11 Packet # 1 2 3 4 5 6 7 8 9 10 11
skipping to change at page 17, line 33 skipping to change at page 17, line 33
PDV 0 10 50 U 20 0 10 50 30 20 0 PDV 0 10 50 U 20 0 10 50 30 20 0
Figure 2: Delay Examples Figure 2: Delay Examples
Clearly, the range of PDV values is 50 ms in both cases above, and Clearly, the range of PDV values is 50 ms in both cases above, and
this is the statistic that determines the size of a de-jitter buffer. this is the statistic that determines the size of a de-jitter buffer.
The IPDV range is minimal in response to the smooth variation in The IPDV range is minimal in response to the smooth variation in
Example A (20 ms). However, IPDV responds to the faster variations Example A (20 ms). However, IPDV responds to the faster variations
in Example B (60 ms range from 40 to -20). Here the IPDV range is in Example B (60 ms range from 40 to -20). Here the IPDV range is
larger than the PDV range, and over-estimates the buffer size larger than the PDV range, and overestimates the buffer size
requirements. requirements.
A heuristic method to estimate buffer size using IPDV is to sum the A heuristic method to estimate buffer size using IPDV is to sum the
consecutive positive or zero values as an estimate of PDV range. consecutive positive or zero values as an estimate of PDV range.
However, this is more complicated to assess than the PDV range, and However, this is more complicated to assess than the PDV range, and
has strong dependence on the actual sequence of IPDV values (any has strong dependence on the actual sequence of IPDV values (any
negative IPDV value stops the summation, and again causes an negative IPDV value stops the summation, and again causes an
underestimate). underestimate).
IPDV values can be viewed as the adjustments that an adaptive de- IPDV values can be viewed as the adjustments that an adaptive de-
jitter buffer would make, IF it could make adjustments on a packet- jitter buffer would make, if it could make adjustments on a packet-
by-packet basis. However, adaptive de-jitter buffers don't make by-packet basis. However, adaptive de-jitter buffers don't make
adjustments this frequently, so the value of this information is adjustments this frequently, so the value of this information is
unknown. The short-term variations may be useful to know in some unknown. The short-term variations may be useful to know in some
other cases. other cases.
5.4. Y.1540 Appendix II 5.4. Y.1540 Appendix II
Appendix II of [Y.1540] describes a secondary terminology for delay Appendix II of [Y.1540] describes a secondary terminology for delay
variation. It compares IPDV, PDV (referred to as 2-point PDV), and variation. It compares IPDV, PDV (referred to as 2-point PDV), and
1-point packet delay variation (which assumes a periodic stream and 1-point packet delay variation (which assumes a periodic stream and
assesses variation against an ideal arrival schedule constructed at a assesses variation against an ideal arrival schedule constructed at a
single measurement point). This early comparison discusses some of single measurement point). This early comparison discusses some of
the same considerations raised in section 6 below. the same considerations raised in Section 6 below.
5.5. Clark's ITU-T SG 12 Contribution 5.5. Clark's ITU-T SG 12 Contribution
Alan Clark's contribution to ITU-T Study Group 12 in January 2003, Alan Clark's contribution to ITU-T Study Group 12 in January 2003
provided an analysis of the root causes of delay variation and provided an analysis of the root causes of delay variation and
investigated different techniques for measurement and modeling of investigated different techniques for measurement and modeling of
"jitter" [COM12.D98]. Clark compared a metric closely related to "jitter" [COM12.D98]. Clark compared a metric closely related to
IPDV, Mean Packet-to-Packet Delay Variation, MPPDV = mean(abs(D(i)- IPDV, Mean Packet-to-Packet Delay Variation, MPPDV = mean(abs(D(i)-
D(i-1))) to the newly proposed Mean Absolute Packet Delay Variation D(i-1))) to the newly proposed Mean Absolute Packet Delay Variation
(MAPDV2, see [G.1020]). One of the tasks for this study was to (MAPDV2, see [G.1020]). One of the tasks for this study was to
estimate the number of packet discards in a de-jitter buffer. Clark estimate the number of packet discards in a de-jitter buffer. Clark
concluded that MPPDV did not track the ramp delay variation he concluded that MPPDV did not track the ramp delay variation he
associated access link congestion (similar to Figure 2, Example A associated access link congestion (similar to Figure 2, Example A
above), but MAPDV2 did. above), but MAPDV2 did.
Clark also briefly looked at PDV (as described in the 2002 version Clark also briefly looked at PDV (as described in the 2002 version of
of[Y.1541]). He concluded that if PDV was applied to a series of [Y.1541]). He concluded that if PDV was applied to a series of very
very short measurement intervals (e.g., 200ms), it could be used to short measurement intervals (e.g., 200 ms), it could be used to
determine the fraction of intervals with high packet discard rates. determine the fraction of intervals with high packet discard rates.
6. Additional Properties and Comparisons 6. Additional Properties and Comparisons
This section treats some of the earlier comparison areas in more This section treats some of the earlier comparison areas in more
detail, and introduces new areas for comparison. detail and introduces new areas for comparison.
6.1. Packet Loss 6.1. Packet Loss
The measurement packet loss is of great influence for the delay The measurement of packet loss is of great influence for the delay
variation results, as displayed in the figures 3 and 4 (L means Lost variation results, as displayed in the Figures 3 and 4 (L means Lost
and U means undefined). Figure 3 shows that in the extreme case of and U means Undefined). Figure 3 shows that in the extreme case of
every other packet loss, the IPDV doesn't produce any results, while every other packet loss, the IPDV metric doesn't produce any results,
the PDV produces results for all arriving packets. while the PDV produces results for all arriving packets.
Packet # 1 2 3 4 5 6 7 8 9 10 Packet # 1 2 3 4 5 6 7 8 9 10
Lost L L L L L Lost L L L L L
--------------------------------------- ---------------------------------------
Delay, ms 3 U 5 U 4 U 3 U 4 U Delay, ms 3 U 5 U 4 U 3 U 4 U
IPDV U U U U U U U U U U IPDV U U U U U U U U U U
PDV 0 U 2 U 1 U 0 U 1 U PDV 0 U 2 U 1 U 0 U 1 U
Figure 3: Path Loss Every Other Packet Figure 3: Path Loss Every Other Packet
In case of a burst of packet loss, as displayed in figure 3, both the In case of a burst of packet loss, as displayed in Figure 4, both the
IPDV and PDV produces some results. Note that PDV still produces IPDV and PDV metrics produce some results. Note that PDV still
more values than IPDV. produces more values than IPDV.
Packet # 1 2 3 4 5 6 7 8 9 10 Packet # 1 2 3 4 5 6 7 8 9 10
Lost L L L L L Lost L L L L L
--------------------------------------- ---------------------------------------
Delay, ms 3 4 U U U U U 5 4 3 Delay, ms 3 4 U U U U U 5 4 3
IPDV U 1 U U U U U U -1 -1 IPDV U 1 U U U U U U -1 -1
PDV 0 1 U U U U U 2 1 0 PDV 0 1 U U U U U 2 1 0
Figure 4: Burst of Packet Loss Figure 4: Burst of Packet Loss
In conclusion, the PDV results are affected by the packet loss ratio. In conclusion, the PDV results are affected by the packet-loss ratio.
The IPDV results are affected by both the packet loss ratio and the The IPDV results are affected by both the packet-loss ratio and the
packet loss distribution. In the extreme case of loss of every other packet-loss distribution. In the extreme case of loss of every other
packet, IPDV doesn't provide any results. packet, IPDV doesn't provide any results.
6.2. Path Changes 6.2. Path Changes
When there is little or no stability in the network under test, then When there is little or no stability in the network under test, then
the devices that attempt to characterize the network are equally the devices that attempt to characterize the network are equally
stressed, especially if the results displayed are used to make stressed, especially if the results displayed are used to make
inferences which may not be valid. inferences that may not be valid.
Sometimes the path characteristics change during a measurement Sometimes the path characteristics change during a measurement
interval. The change may be due to link or router failure, interval. The change may be due to link or router failure,
administrative changes prior to maintenance (e.g., link cost change), administrative changes prior to maintenance (e.g., link-cost change),
or re-optimization of routing using new information. All these or re-optimization of routing using new information. All these
causes are usually infrequent, and network providers take appropriate causes are usually infrequent, and network providers take appropriate
measures to ensure this. Automatic restoration to a back-up path is measures to ensure this. Automatic restoration to a back-up path is
seen as a desirable feature of IP networks. seen as a desirable feature of IP networks.
Frequent path changes and prolonged congestion with substantial Frequent path changes and prolonged congestion with substantial
packet loss clearly make delay variation measurements challenging. packet loss clearly make delay variation measurements challenging.
Path changes are usually accompanied by a sudden, persistent increase Path changes are usually accompanied by a sudden, persistent increase
or decrease in one-way-delay. [Cia03] gives one such example. We or decrease in one-way delay. [Cia03] gives one such example. We
assume that a restoration path either accepts a stream of packets, or assume that a restoration path either accepts a stream of packets or
is not used for that particular stream (e.g., no multi-path for is not used for that particular stream (e.g., no multi-path for
flows). flows).
In any case, a change in the TTL (or Hop Limit) of the received In any case, a change in the Time to Live (TTL) (or Hop Limit) of the
packets indicates that the path is no longer the same. Transient received packets indicates that the path is no longer the same.
packet reordering may also be observed with path changes, due to use Transient packet reordering may also be observed with path changes,
of non-optimal routing while updates propagate through the network due to use of non-optimal routing while updates propagate through the
(see [Casner] and [Cia03] ) network (see [Casner] and [Cia03] )
Many, if not all, packet streams experience packet loss in Many, if not all, packet streams experience packet loss in
conjunction with a path change. However, it is certainly possible conjunction with a path change. However, it is certainly possible
that the active measurement stream does not experience loss. This that the active measurement stream does not experience loss. This
may be due to use of a long inter-packet sending interval with may be due to use of a long inter-packet sending interval with
respect to the restoration time, and it becomes more likely as "fast respect to the restoration time, and it becomes more likely as "fast
restoration" techniques see wider deployment (e.g., [RFC4090]. restoration" techniques see wider deployment (e.g., [RFC4090]).
Thus, there are two main cases to consider, path changes accompanied Thus, there are two main cases to consider, path changes accompanied
by loss, and those that are lossless from the point of view of the by loss, and those that are lossless from the point of view of the
active measurement stream. The subsections below examine each of active measurement stream. The subsections below examine each of
these cases. these cases.
6.2.1. Lossless Path Change 6.2.1. Lossless Path Change
In the lossless case, a path change will typically affect only one In the lossless case, a path change will typically affect only one
IPDV singleton. For example, the delay sequence in the Figure below IPDV singleton. For example, the delay sequence in the Figure below
skipping to change at page 21, line 11 skipping to change at page 21, line 14
The use of the new path and its delay variation can be quantified by The use of the new path and its delay variation can be quantified by
treating the PDV distribution as bi-modal, and characterizing each treating the PDV distribution as bi-modal, and characterizing each
mode separately. This would involve declaring a new path within the mode separately. This would involve declaring a new path within the
sample, and using a new local minimum delay as the PDV reference sample, and using a new local minimum delay as the PDV reference
delay for the sub-sample (or time interval) where the new path is delay for the sub-sample (or time interval) where the new path is
present. present.
The process of detecting a bi-modal delay distribution is made The process of detecting a bi-modal delay distribution is made
difficult if the typical delay variation is larger than the delay difficult if the typical delay variation is larger than the delay
change associated with the new path. However, information on TTL (or change associated with the new path. However, information on a TTL
Hop Limit) change or the presence of transient reordering can assist (or Hop Limit) change or the presence of transient reordering can
in an automated decision. assist in an automated decision.
The effect of path changes may also be reduced by making PDV The effect of path changes may also be reduced by making PDV
measurements over short intervals (minutes, as opposed to hours). measurements over short intervals (minutes, as opposed to hours).
This way, a path change will affect one sample and its PDV values. This way, a path change will affect one sample and its PDV values.
Assuming that the mean or median one-way-delay changes appreciably on Assuming that the mean or median one-way delay changes appreciably on
the new path, then subsequent measurements can confirm a path change the new path, then subsequent measurements can confirm a path change
and trigger special processing on the interval to revise the PDV and trigger special processing on the interval to revise the PDV
result. result.
Alternatively, if the path change is detected, by monitoring the test Alternatively, if the path change is detected, by monitoring the test
packets TTL or Hop Limit, or monitoring the change in the IGP link- packets TTL or Hop Limit, or monitoring the change in the IGP link-
state database, the results of measurement before and after the path state database, the results of measurement before and after the path
change could be kept separated, presenting two different change could be kept separated, presenting two different
distributions. This avoids the difficult task of determining the distributions. This avoids the difficult task of determining the
different modes of a multi-modal distribution. different modes of a multi-modal distribution.
6.2.2. Path Change with Loss 6.2.2. Path Change with Loss
If the path change is accompanied by loss, such that there are no If the path change is accompanied by loss, such that there are no
consecutive packet pairs that span the change, then no IPDV consecutive packet pairs that span the change, then no IPDV
singletons will reflect the change. This may or may not be singletons will reflect the change. This may or may not be
desirable, depending on the ultimate use of the delay variation desirable, depending on the ultimate use of the delay variation
measurement. Figure 6, in which L means Lost and U means undefined, measurement. Figure 6, in which L means Lost and U means Undefined,
illustrates this case. illustrates this case.
Packet # 1 2 3 4 5 6 7 8 9 Packet # 1 2 3 4 5 6 7 8 9
Lost L L Lost L L
------------------------------------ ------------------------------------
Delay, ms 3 4 3 3 U U 8 9 8 Delay, ms 3 4 3 3 U U 8 9 8
IPDV U 1 -1 0 U U U 1 -1 IPDV U 1 -1 0 U U U 1 -1
PDV 0 1 0 0 U U 5 6 5 PDV 0 1 0 0 U U 5 6 5
Figure 6: Path Change with Loss Figure 6: Path Change with Loss
PDV will again produce a bi-modal distribution. But here, the
PDV will again produce a bimodal distribution. But here, the
decision process to define sub-intervals associated with each path is decision process to define sub-intervals associated with each path is
further assisted by the presence of loss, in addition to TTL, further assisted by the presence of loss, in addition to TTL,
reordering information, and use of short measurement intervals reordering information, and use of short measurement intervals
consistent with the duration of user sessions. It is reasonable to consistent with the duration of user sessions. It is reasonable to
assume that at least loss and delay will be measured simultaneously assume that at least loss and delay will be measured simultaneously
with PDV and/or IPDV. with PDV and/or IPDV.
IPDV does not help to detect path changes when accompanied by loss, IPDV does not help to detect path changes when accompanied by loss,
and this is a disadvantage for those who rely solely on IPDV and this is a disadvantage for those who rely solely on IPDV
measurements. measurements.
skipping to change at page 22, line 48 skipping to change at page 22, line 36
Maintaining time accuracy <<1ms has typically required access to Maintaining time accuracy <<1ms has typically required access to
dedicated time receivers at all measurement points. Global dedicated time receivers at all measurement points. Global
positioning system (GPS) receivers have often been installed to positioning system (GPS) receivers have often been installed to
support measurements. The GPS installation conditions are fairly support measurements. The GPS installation conditions are fairly
restrictive, and many prospective measurement efforts have found the restrictive, and many prospective measurement efforts have found the
deployment complexity and system maintenance too difficult. deployment complexity and system maintenance too difficult.
As mentioned above, [Demichelis] observed that PDV places greater As mentioned above, [Demichelis] observed that PDV places greater
demands on clock synchronization than for IPDV. This observation demands on clock synchronization than for IPDV. This observation
deserves more discussion. Synchronization errors have two deserves more discussion. Synchronization errors have two
components: time of day errors and clock frequency errors (resulting components: time-of-day errors and clock-frequency errors (resulting
in skew). in skew).
Both IPDV and PDV are sensitive to time-of-day errors when attempting Both IPDV and PDV are sensitive to time-of-day errors when attempting
to align measurement intervals at the source and destination. Gross to align measurement intervals at the source and destination. Gross
mis-alignment of the measurement intervals can lead to lost packets, misalignment of the measurement intervals can lead to lost packets,
for example if the receiver is not ready when the first test packet for example, if the receiver is not ready when the first test packet
arrives. However, both IPDV and PDV assess delay differences, so the arrives. However, both IPDV and PDV assess delay differences, so the
error present in any two one-way-delay singletons will cancel as long error present in any two one-way-delay singletons will cancel as long
as the error is constant. So, the demand for NTP or GPS as the error is constant. So, the demand for NTP or GPS
synchronization comes primarily from one-way delay measurement time- synchronization comes primarily from one-way-delay measurement time-
of-day accuracy requirements. Delay variation and measurement of-day accuracy requirements. Delay variation and measurement
interval alignment are relatively less demanding. interval alignment are relatively less demanding.
Skew is a measure of the change in clock time over an interval w.r.t. Skew is a measure of the change in clock time over an interval with
a reference clock. Both IPDV and PDV are affected by skew, but the respect to a reference clock. Both IPDV and PDV are affected by
error sensitivity in IPDV singletons is less because the intervals skew, but the error sensitivity in IPDV singletons is less because
between consecutive packets are rather small, especially when the intervals between consecutive packets are rather small,
compared to the overall measurement interval. Since PDV computes the especially when compared to the overall measurement interval. Since
difference between a single reference delay (the sample minimum) and PDV computes the difference between a single reference delay (the
all other delays in the measurement interval, the constraint on skew sample minimum) and all other delays in the measurement interval, the
error is greater to attain the same accuracy as IPDV. Again, use of constraint on skew error is greater to attain the same accuracy as
short PDV measurement intervals (on the order of minutes, not hours) IPDV. Again, use of short PDV measurement intervals (on the order of
provides some relief from the effects of skew error. Thus, the minutes, not hours) provides some relief from the effects of skew
additional accuracy demand of PDV can be expressed as a ratio of the error. Thus, the additional accuracy demand of PDV can be expressed
measurement interval to the inter-packet spacing. as a ratio of the measurement interval to the inter-packet spacing.
A practical example is a measurement between two hosts, one with a A practical example is a measurement between two hosts, one with a
synchronized clock and the other with a free-running clock having 50 synchronized clock and the other with a free-running clock having 50
part per million (ppm) long term accuracy. parts per million (ppm) long term accuracy.
o If IPDV measurements are made on packets with a 1 second spacing, o If IPDV measurements are made on packets with a 1 second spacing,
the maximum singleton error will be 1 x 5 x 10^-5 seconds, or the maximum singleton error will be 1 x 5 x 10^-5 seconds, or 0.05
0.05ms. ms.
o If PDV measurements are made on the same packets over a 60 second o If PDV measurements are made on the same packets over a 60 second
measurement interval, then the delay variation due to the max measurement interval, then the delay variation due to the max
free-running clock error will be 60 x 5 x 10-5 seconds, or 3ms free-running clock error will be 60 x 5 x 10-5 seconds, or 3ms
delay variation error from the first packet to the last. delay variation error from the first packet to the last.
Therefore, the additional accuracy required for equivalent PDV error Therefore, the additional accuracy required for equivalent PDV error
under these conditions is a factor of 60 more than for IPDV. This is under these conditions is a factor of 60 more than for IPDV. This is
a rather extreme scenario, because time-of-day error of 1 second a rather extreme scenario, because time-of-day error of 1 second
would accumulate in ~5.5 hours, potentially causing the measurement would accumulate in ~5.5 hours, potentially causing the measurement
interval alignment issue described above. interval alignment issue described above.
If skew is present in a sample of one-way-delays, its symptom is If skew is present in a sample of one-way delays, its symptom is
typically a nearly linear growth or decline over all the one-way- typically a nearly linear growth or decline over all the one-way-
delay values. As a practical matter, if the same slope appears delay values. As a practical matter, if the same slope appears
consistently in the measurements, then it may be possible to fit the consistently in the measurements, then it may be possible to fit the
slope and compensate for the skew in the one-way-delay measurements, slope and compensate for the skew in the one-way-delay measurements,
thereby avoiding the issue in the PDV calculations that follow. See thereby avoiding the issue in the PDV calculations that follow. See
[RFC3393] for additional information on compensating for skew. [RFC3393] for additional information on compensating for skew.
Values for IPDV may have non-zero mean over a sample when clock skew Values for IPDV may have non-zero mean over a sample when clock skew
is present. This tends to complicate IPDV analysis when using the is present. This tends to complicate IPDV analysis when using the
assumptions of a zero mean and a symmetric distribution. assumptions of a zero mean and a symmetric distribution.
There is a third factor related to clock error and stability: this is There is a third factor related to clock error and stability: this is
the presence of a clock synchronization protocol (e.g., NTP) and the the presence of a clock-synchronization protocol (e.g., NTP) and the
time adjustment operations that result. When a time error is time-adjustment operations that result. When a time error is
detected (typically on the order of a few milliseconds), the host detected (typically on the order of a few milliseconds), the host
clock frequency is continuously adjusted to reduce the time error. clock frequency is continuously adjusted to reduce the time error.
If these adjustments take place during a measurement interval, they If these adjustments take place during a measurement interval, they
may appear as delay variation when none was present, and therefore may appear as delay variation when none was present, and therefore
are a source of error (regardless of the DV form considered). are a source of error (regardless of the form of delay variation
considered).
6.4. Spatial Composition 6.4. Spatial Composition
ITU-T Recommendation [Y.1541] gives a provisional method to compose a ITU-T Recommendation [Y.1541] gives a provisional method to compose a
PDV metric using PDV measurement results from two or more sub-paths. PDV metric using PDV measurement results from two or more sub-paths.
Additional methods are considered in Additional methods are considered in [IPPM-Spatial].
[I-D.ietf-ippm-spatial-composition].
PDV has a clear advantage at this time, since there is no validated PDV has a clear advantage at this time, since there is no validated
method to compose an IPDV metric. In addition, IPDV results depend method to compose an IPDV metric. In addition, IPDV results depend
greatly on the exact sequence of packets and may not lend themselves greatly on the exact sequence of packets and may not lend themselves
easily to the composition problem, where segments must be assumed to easily to the composition problem, where segments must be assumed to
have independent delay distributions. have independent delay distributions.
6.5. Reporting a Single Number (SLA) 6.5. Reporting a Single Number (SLA)
Despite the risk of over-summarization, measurements must often be Despite the risk of over-summarization, measurements must often be
skipping to change at page 24, line 51 skipping to change at page 24, line 42
each measurement interval. The single number summary statistic is a each measurement interval. The single number summary statistic is a
key component of SLAs, where a threshold on that number must be met key component of SLAs, where a threshold on that number must be met
x% of the time. x% of the time.
The simplicity of the PDV distribution lends itself to this The simplicity of the PDV distribution lends itself to this
summarization process (including use of the percentiles, median or summarization process (including use of the percentiles, median or
mean). An SLA of the form "no more than x% of packets in a mean). An SLA of the form "no more than x% of packets in a
measurement interval shall have PDV >= y ms, for no less than z% of measurement interval shall have PDV >= y ms, for no less than z% of
time" is relatively straightforward to specify and implement. time" is relatively straightforward to specify and implement.
[Y.1541] introduced the notion of a pseudo-range when setting an [Y.1541] introduced the notion of a pseudo-range when setting an
objective for the 99.9%-ile of PDV. The conventional range (max-min) objective for the 99.9th percentile of PDV. The conventional range
was avoided for several reasons, including stability of the maximum (max-min) was avoided for several reasons, including stability of the
delay. The 99.9%-ile of PDV is helpful to performance planners maximum delay. The 99.9th percentile of PDV is helpful to
(seeking to meet some user-to-user objective for delay) and in design performance planners (seeking to meet some user-to-user objective for
of de-jitter buffer sizes, even those with adaptive capabilities. delay) and in design of de-jitter buffer sizes, even those with
adaptive capabilities.
IPDV does not lend itself to summarization so easily. The mean IPDV IPDV does not lend itself to summarization so easily. The mean IPDV
is typically zero. As the IPDV distribution will have two tails is typically zero. As the IPDV distribution will have two tails
(positive and negative) the range or pseudo-range would not match the (positive and negative), the range or pseudo-range would not match
needed de-jitter buffer size. Additional complexity may be the needed de-jitter buffer size. Additional complexity may be
introduced when the variation exceeds the inter-packet sending introduced when the variation exceeds the inter-packet sending
interval, as discussed above (in sections 5.2 and 6.2.1). Should the interval, as discussed above (in Sections 5.2 and 6.2.1). Should the
Inter-Quartile Range be used? Should the singletons beyond some Inter-Quartile Range be used? Should the singletons beyond some
threshold be counted (e.g., mean +/- 50ms)? A strong rationale for threshold be counted (e.g., mean +/- 50ms)? A strong rationale for
one of these summary statistics has yet to emerge. one of these summary statistics has yet to emerge.
When summarizing IPDV, some prefer the simplicity of the single-sided When summarizing IPDV, some prefer the simplicity of the single-sided
distribution created by taking the absolute value of each singleton distribution created by taking the absolute value of each singleton
result, abs(D(i)-D(i-1)). This approach sacrifices the two-sided result, abs(D(i)-D(i-1)). This approach sacrifices the two-sided
inter-arrival spread information in the distribution. It also makes inter-arrival spread information in the distribution. It also makes
the evaluation using percentiles more confusing, because a single the evaluation using percentiles more confusing, because a single
late packet that exceeds the variation threshold will cause two pairs late packet that exceeds the variation threshold will cause two pairs
of singletons to fail the criteria (one positive, the other negative of singletons to fail the criteria (one positive, the other negative
converted to positive). The single-sided PDV distribution is an converted to positive). The single-sided PDV distribution is an
advantage in this category. advantage in this category.
6.6. Jitter in RTCP Reports 6.6. Jitter in RTCP Reports
[RFC3550] gives the calculation of the inter-arrival Jitter field for Section 6.4.1 of [RFC3550] gives the calculation of the "inter-
the RTCP report, with a sample implementation in an Appendix. arrival jitter" field for the RTP Control Protocol (RTCP) report,
with a sample implementation in an Appendix.
The RTCP Jitter value can be calculated using IPDV singletons. If The RTCP "interarrival jitter" value can be calculated using IPDV
there is packet reordering, as defined in [RFC4737], then estimates singletons. If there is packet reordering, as defined in [RFC4737],
of Jitter based on IPDV may vary slightly, because [RFC3550] then estimates of Jitter based on IPDV may vary slightly, because
specifies the use of receive packet order. [RFC3550] specifies the use of receive-packet order.
Just as there is no simple way to convert PDV singletons to IPDV Just as there is no simple way to convert PDV singletons to IPDV
singletons without returning to the original sample of delay singletons without returning to the original sample of delay
singletons, there is no clear relationship between PDV and [RFC3550] singletons, there is no clear relationship between PDV and [RFC3550]
Jitter. "interarrival jitter".
6.7. MAPDV2 6.7. MAPDV2
MAPDV2 stands for Mean Absolute Packet Delay Variation (version) 2, MAPDV2 stands for Mean Absolute Packet Delay Variation (version) 2,
and is specified in [G.1020]. The MAPDV2 algorithm computes a and is specified in [G.1020]. The MAPDV2 algorithm computes a
smoothed running estimate of the mean delay using the one-way delays smoothed running estimate of the mean delay using the one-way delays
of 16 previous packets. It compares the current one-way-delay to the of 16 previous packets. It compares the current one-way delay to the
estimated mean, separately computes the means of positive and estimated mean, separately computes the means of positive and
negative deviations, and sums these deviation means to produce negative deviations, and sums these deviation means to produce
MAPVDV2. In effect, there is a MAPDV2 singleton for every arriving MAPVDV2. In effect, there is a MAPDV2 singleton for every arriving
packet, so further summarization is usually warranted. packet, so further summarization is usually warranted.
Neither IPDV or PDV forms assist in the computation of MAPDV2. Neither IPDV or PDV forms assist in the computation of MAPDV2.
6.8. Load Balancing 6.8. Load Balancing
Network traffic load balancing is a process to divide packet traffic Network traffic load balancing is a process to divide packet traffic
in order to provide a more even distribution over two or more equally in order to provide a more even distribution over two or more equally
viable paths. The paths chosen are based on the IGP cost metrics, viable paths. The paths chosen are based on the IGP cost metrics,
while the delay depends on the path's physical layout. Usually, the while the delay depends on the path's physical layout. Usually, the
balancing process is performed on a per-flow basis to avoid delay balancing process is performed on a per-flow basis to avoid delay
variation experienced when packets traverse different physical paths. variation experienced when packets traverse different physical paths.
If the sample includes test packets with different characteristics If the sample includes test packets with different characteristics
such as IP addresses/ports, there could be multi-modal delay such as IP addresses/ports, there could be multi-modal delay
distributions present. The PDV form makes the identification of distributions present. The PDV form makes the identification of
multiple modes possible. IPDV may also reveal that multiple paths multiple modes possible. IPDV may also reveal that multiple paths
are in use with a mixed flow sample, but the different delay modes are in use with a mixed-flow sample, but the different delay modes
are not easily divided and analyzed separately. are not easily divided and analyzed separately.
Should the delay singletons using multiple addresses/ports be Should the delay singletons using multiple addresses/ports be
combined in the same sample? Should we characterize each mode combined in the same sample? Should we characterize each mode
separately? (This question also applies to the Path Change case.) separately? (This question also applies to the Path Change case.)
It depends on the task to be addressed by the measurement. It depends on the task to be addressed by the measurement.
For the task of de-jitter buffer sizing or assessing queue For the task of de-jitter buffer sizing or assessing queue
occupation, the modes should be characterized separately because occupation, the modes should be characterized separately because
flows will experience only one mode on a stable path. Use of a flows will experience only one mode on a stable path. Use of a
single flow description (address/port combination) in each sample single flow description (address/port combination) in each sample
simplifies this analysis. Multiple modes may be identified by simplifies this analysis. Multiple modes may be identified by
collecting samples with different flow attributes, and collecting samples with different flow attributes, and
characterization of multiple paths can proceed with comparison of the characterization of multiple paths can proceed with comparison of the
delay distributions from each sample. delay distributions from each sample.
For the task of capacity planning and routing optimization, For the task of capacity planning and routing optimization,
characterizing the modes separately could offer an advantage. characterizing the modes separately could offer an advantage.
Network wide capacity planning (as opposed to link capacity planning) Network-wide capacity planning (as opposed to link capacity planning)
takes as input the core traffic matrix, which corresponds to a matrix takes as input the core traffic matrix, which corresponds to a matrix
of traffic transferred from every source to every destination in the of traffic transferred from every source to every destination in the
network. Applying the core traffic matrix along with the routing network. Applying the core traffic matrix along with the routing
information (typically the link state database of a routing protocol) information (typically the link state database of a routing protocol)
in a capacity planning tool offers the possibility to visualize the in a capacity planning tool offers the possibility to visualize the
paths where the traffic flows and to optimize the routing based on paths where the traffic flows and to optimize the routing based on
the link utilization. In the case where equal cost multiple paths the link utilization. In the case where equal cost multiple paths
(ECMP) are used, the traffic will be load balanced onto multiple (ECMPs) are used, the traffic will be load balanced onto multiple
paths. If each mode of the IP delay multi-modal distribution can be paths. If each mode of the IP delay multi-modal distribution can be
associated with a specific path, the delay performance offers an associated with a specific path, the delay performance offers an
extra optimization parameter, i.e. the routing optimization based on extra optimization parameter, i.e., the routing optimization based on
the IP delay variation metric. As an example, the load balancing the IP delay variation metric. As an example, the load balancing
across ECMPs could be suppressed so that the VoIP calls would only be across ECMPs could be suppressed so that the Voice over IP (VoIP)
routed via the path with the lower IP delay variation. Clearly, any calls would only be routed via the path with the lower IP delay
modifications can result in new delay performance measurements, so variation. Clearly, any modifications can result in new delay
there must be a verification step to ensure the desired outcome. performance measurements, so there must be a verification step to
ensure the desired outcome.
7. Applicability of the Delay Variation Forms and Recommendations 7. Applicability of the Delay Variation Forms and Recommendations
Based on the comparisons of IPDV and PDV presented above, this Based on the comparisons of IPDV and PDV presented above, this
section matches the attributes of each form with the tasks described section matches the attributes of each form with the tasks described
earlier. We discuss the more general circumstances first. earlier. We discuss the more general circumstances first.
7.1. Uses 7.1. Uses
7.1.1. Inferring Queue Occupancy 7.1.1. Inferring Queue Occupancy
skipping to change at page 27, line 34 skipping to change at page 27, line 32
stream. If the minimum delay is not the true minimum, then the PDV stream. If the minimum delay is not the true minimum, then the PDV
distribution captures the variation in queuing time and some distribution captures the variation in queuing time and some
additional amount of queuing time is experienced, but unknown. One additional amount of queuing time is experienced, but unknown. One
can summarize the PDV distribution with the mean, median, and other can summarize the PDV distribution with the mean, median, and other
statistics. statistics.
IPDV can capture the difference in queuing time from one packet to IPDV can capture the difference in queuing time from one packet to
the next, but this is a different distribution from the queue the next, but this is a different distribution from the queue
occupancy revealed by PDV. occupancy revealed by PDV.
7.1.2. Determining De-jitter Buffer Size (and FEC Design) 7.1.2. Determining De-Jitter Buffer Size (and FEC Design)
This task is complimentary to the problem of inferring queue This task is complimentary to the problem of inferring queue
occupancy through measurement. Again, use of the sample minimum as occupancy through measurement. Again, use of the sample minimum as
the reference delay for PDV yields a distribution that is very the reference delay for PDV yields a distribution that is very
relevant to de-jitter buffer size. This is because the minimum delay relevant to de-jitter buffer size. This is because the minimum delay
is an alignment point for the smoothing operation of de-jitter is an alignment point for the smoothing operation of de-jitter
buffers. A de-jitter buffer that is ideally aligned with the delay buffers. A de-jitter buffer that is ideally aligned with the delay
variation adds zero buffer time to packets with the longest variation adds zero buffer time to packets with the longest
accommodated network delay (any packets with longer delays are accommodated network delay (any packets with longer delays are
discarded). Thus, a packet experiencing minimum network delay should discarded). Thus, a packet experiencing minimum network delay should
be aligned to wait the maximum length of the de-jitter buffer. With be aligned to wait the maximum length of the de-jitter buffer. With
this alignment, the stream is smoothed with no unnecessary delay this alignment, the stream is smoothed with no unnecessary delay
added. [G.1020] illustrates the ideal relationship between network added. Figure 5 of [G.1020] illustrates the ideal relationship
delay variation and buffer time. between network delay variation and buffer time.
The PDV distribution is also useful for this task, but different The PDV distribution is also useful for this task, but different
statistics are preferred. The range (max-min) or the 99.9%-ile of statistics are preferred. The range (max-min) or the 99.9th
PDV (pseudo-range) are closely related to the buffer size needed to percentile of PDV (pseudo-range) are closely related to the buffer
accommodate the observed network delay variation. size needed to accommodate the observed network delay variation.
The PDV distribution directly addresses the FEC waiting time The PDV distribution directly addresses the FEC waiting time
question. When the PDV distribution has a 99th percentile of 10ms, question. When the PDV distribution has a 99th percentile of 10ms,
then waiting 10ms longer than the FEC protection interval will allow then waiting 10ms longer than the FEC protection interval will allow
99% of late packets to arrive and be used in the FEC block. 99% of late packets to arrive and be used in the FEC block.
In some cases, the positive excursions (or series of positive In some cases, the positive excursions (or series of positive
excursions) of IPDV may help to approximate the de-jitter buffer excursions) of IPDV may help to approximate the de-jitter buffer
size, but there is no guarantee that a good buffer estimate will size, but there is no guarantee that a good buffer estimate will
emerge, especially when the delay varies as a positive trend over emerge, especially when the delay varies as a positive trend over
several test packets. several test packets.
7.1.3. Spatial Composition 7.1.3. Spatial Composition
PDV has a clear advantage at this time, since there is no validated PDV has a clear advantage at this time, since there is no validated
method to compose an IPDV metric. method to compose an IPDV metric.
7.1.4. Service Level Specification: Reporting a Single Number 7.1.4. Service-Level Specification: Reporting a Single Number
The one-sided PDV distribution can be constrained with a single The one-sided PDV distribution can be constrained with a single
statistic, such as an upper percentile, so it is preferred. The IPDV statistic, such as an upper percentile, so it is preferred. The IPDV
distribution is two-sided, usually has zero mean, and no universal distribution is two-sided, usually has zero mean, and no universal
summary statistic that relates to a physical quantity has emerged in summary statistic that relates to a physical quantity has emerged in
years of experience. years of experience.
7.2. Challenging Circumstances 7.2. Challenging Circumstances
Note that measurement of delay variation may not be the primary Note that measurement of delay variation may not be the primary
concern under unstable and unreliable circumstances. concern under unstable and unreliable circumstances.
7.2.1. Clock and Storage Issues 7.2.1. Clock and Storage Issues
When appreciable skew is present between measurement system clocks, When appreciable skew is present between measurement system clocks,
then IPDV has an advantage because PDV would require processing over IPDV has an advantage because PDV would require processing over the
the entire sample to remove the skew error. However, significant entire sample to remove the skew error. However, significant skew
skew can invalidate IPDV analysis assumptions, such as the zero mean can invalidate IPDV analysis assumptions, such as the zero-mean and
and symmetric distribution characteristics. Small skew may well be symmetric-distribution characteristics. Small skew may well be
within the error tolerance, and both PDV and IPDV results will be within the error tolerance, and both PDV and IPDV results will be
usable. There may be a portion of the skew, measurement interval, usable. There may be a portion of the skew, measurement interval,
and required accuracy 3-D space where IPDV has an advantage, and required accuracy 3-D space where IPDV has an advantage,
depending on the specific measurement specifications. depending on the specific measurement specifications.
Neither form of delay variation is more suited than the other to on- Neither form of delay variation is more suited than the other to
the-fly summarization without memory, and this may be one of the on-the-fly summarization without memory, and this may be one of the
reasons that [RFC3550] RTCP Jitter and MAPDV2 in [G.1020] have reasons that [RFC3550] RTCP Jitter and MAPDV2 in [G.1020] have
attained deployment in low-cost systems. attained deployment in low-cost systems.
7.2.2. Frequent Path Changes 7.2.2. Frequent Path Changes
If the network under test exhibits frequent path changes, on the If the network under test exhibits frequent path changes, on the
order of several new routes per minute, then IPDV appears to isolate order of several new routes per minute, then IPDV appears to isolate
the delay variation on each path from the transient effect of path the delay variation on each path from the transient effect of path
change (especially if there is packet loss at the time of path change (especially if there is packet loss at the time of path
change). However, if one intends to use IPDV to indicate path change). However, if one intends to use IPDV to indicate path
skipping to change at page 29, line 29 skipping to change at page 29, line 29
When path changes are frequent and cause packet loss, delay variation When path changes are frequent and cause packet loss, delay variation
is probably less important than the loss episodes and attention is probably less important than the loss episodes and attention
should be turned to the loss metric instead. should be turned to the loss metric instead.
7.2.3. Frequent Loss 7.2.3. Frequent Loss
If the network under test exhibits frequent loss, then PDV may If the network under test exhibits frequent loss, then PDV may
produce a larger set of singletons for the sample than IPDV. This is produce a larger set of singletons for the sample than IPDV. This is
due to IPDV requiring consecutive packet arrivals to assess delay due to IPDV requiring consecutive packet arrivals to assess delay
variation, compared to PDV where any packet arrival is useful. The variation, compared to PDV where any packet arrival is useful. The
worst case is when no consecutive packets arrive, and the entire IPDV worst case is when no consecutive packets arrive and the entire IPDV
sample would be undefined. PDV would successfully produce a sample sample would be undefined, yet PDV would successfully produce a
based on the arriving packets. sample based on the arriving packets.
7.2.4. Load Balancing 7.2.4. Load Balancing
PDV distributions offer the most straightforward way to identify that PDV distributions offer the most straightforward way to identify that
a sample of packets have traversed multiple paths. The tasks of de- a sample of packets have traversed multiple paths. The tasks of
jitter buffer sizing or assessing queue occupation with PDV should be de-jitter buffer sizing or assessing queue occupation with PDV should
use a sample with a single flow because flows will experience only be use a sample with a single flow because flows will experience only
one mode on a stable path, and it simplifies the analysis. one mode on a stable path, and it simplifies the analysis.
7.3. Summary 7.3. Summary
+---------------+----------------------+----------------------------+ +---------------+----------------------+----------------------------+
| Comparison | PDV | IPDV | | Comparison | PDV = D(i)-D(min) | IPDV = D(i)-D(i-1) |
| Area | | | | Area | | |
+---------------+----------------------+----------------------------+ +---------------+----------------------+----------------------------+
| Challenging | Less sensitive to | Preferred when path | | Challenging | Less sensitive to | Preferred when path |
| Circumstances | packet loss, and | changes are frequent or | | Circumstances | packet loss, and | changes are frequent or |
| | simplifies analysis | when measurement clocks | | | simplifies analysis | when measurement clocks |
| | when load balancing | exhibit some skew | | | when load balancing | exhibit some skew |
| | or multiple paths | | | | or multiple paths | |
| | are present | | | | are present | |
|---------------|----------------------|----------------------------| |---------------|----------------------|----------------------------|
| Spatial | All validated | Has sensitivity to | | Spatial | All validated | Has sensitivity to |
skipping to change at page 31, line 7 skipping to change at page 31, line 13
Summary of Comparisons Summary of Comparisons
8. Measurement Considerations 8. Measurement Considerations
This section discusses the practical aspects of delay variation This section discusses the practical aspects of delay variation
measurement, with special attention to the two formulations compared measurement, with special attention to the two formulations compared
in this memo. in this memo.
8.1. Measurement Stream Characteristics 8.1. Measurement Stream Characteristics
As stated in the background section, there is a strong dependency As stated in Section 1.2, there is a strong dependency between the
between the active measurement stream characteristics and the active measurement stream characteristics and the results. The IPPM
results. The IPPM literature includes two primary methods for literature includes two primary methods for collecting samples:
collecting samples: Poisson sampling described in [RFC2330], and Poisson sampling described in [RFC2330], and Periodic sampling in
Periodic sampling in[RFC3432]. The Poisson method was intended to [RFC3432]. The Poisson method was intended to collect an unbiased
collect an unbiased sample of performance, while the Periodic method sample of performance, while the Periodic method addresses a "known
addresses a "known bias of interest". Periodic streams are required bias of interest". Periodic streams are required to have random
to have random start times and limited stream duration, in order to start times and limited stream duration, in order to avoid unwanted
avoid unwanted synchronization with some other periodic process, or synchronization with some other periodic process, or cause
cause congestion-aware senders to synchronize with the stream and congestion-aware senders to synchronize with the stream and produce
produce atypical results. The random start time should be different atypical results. The random start time should be different for each
for each new stream. new stream.
It is worth noting that [RFC3393] was developed in parallel with It is worth noting that [RFC3393] was developed in parallel with
[RFC3432]. As a result, all the stream metrics defined in [RFC3393] [RFC3432]. As a result, all the stream metrics defined in [RFC3393]
specify the Poisson sampling method. specify the Poisson sampling method.
Periodic sampling is frequently used in measurements of delay Periodic sampling is frequently used in measurements of delay
variation. Several factors foster this choice: variation. Several factors foster this choice:
1. Many application streams that are sensitive to delay variation 1. Many application streams that are sensitive to delay variation
also exhibit periodicity, and so exemplify the bias of interest. also exhibit periodicity, and so exemplify the bias of interest.
skipping to change at page 32, line 4 skipping to change at page 32, line 11
when the send spacing and send rate are constant. For example, when the send spacing and send rate are constant. For example,
the inter-arrival times at the destination can be compared with the inter-arrival times at the destination can be compared with
an ideal sending schedule, and allowing a one-point measurement an ideal sending schedule, and allowing a one-point measurement
of delay variation (described in [Y.1540]) that approximates the of delay variation (described in [Y.1540]) that approximates the
IPDV form. Simplified methods that approximate PDV are possible IPDV form. Simplified methods that approximate PDV are possible
as well (some are discussed in Appendix II of [Y.1541]). as well (some are discussed in Appendix II of [Y.1541]).
4. Analysis of truncated, or non-symmetrical IPDV distributions is 4. Analysis of truncated, or non-symmetrical IPDV distributions is
simplified. Delay variations in excess of the periodic sending simplified. Delay variations in excess of the periodic sending
interval can cause multiple singleton values at the negative interval can cause multiple singleton values at the negative
limit of the packet spacing (see section 5.2 and [Cia03]). Only limit of the packet spacing (see Section 5.2 and [Cia03]). Only
packet reordering can cause the negative spacing limit to be packet reordering can cause the negative spacing limit to be
exceeded. exceeded.
Despite the emphasis on inter-packet delay differences with IPDV, Despite the emphasis on inter-packet delay differences with IPDV,
both Poisson [Demichelis] and Periodic [Li.Mills] streams have been both Poisson [Demichelis] and Periodic [Li.Mills] streams have been
used, and these references illustrate the different analyses that are used, and these references illustrate the different analyses that are
possible. possible.
The advantages of using a Poisson distribution are discussed in The advantages of using a Poisson distribution are discussed in
[RFC2330]. The main properties are to avoid predicting the sample [RFC2330]. The main properties are to avoid predicting the sample
skipping to change at page 33, line 16 skipping to change at page 33, line 29
At several points in this memo, we have recommended use of test At several points in this memo, we have recommended use of test
intervals on the order of minutes. In their paper examining the intervals on the order of minutes. In their paper examining the
stability of Internet path properties[Zhang.Duff], Zhang et al. stability of Internet path properties[Zhang.Duff], Zhang et al.
concluded that consistency was present on the order of minutes for concluded that consistency was present on the order of minutes for
the performance metrics considered (loss, delay, and throughput) for the performance metrics considered (loss, delay, and throughput) for
the paths they measured. the paths they measured.
The topic of temporal aggregation of performance measured in small The topic of temporal aggregation of performance measured in small
intervals to estimate some larger interval is described in the Metric intervals to estimate some larger interval is described in the Metric
Composition Framework [I-D.ietf-ippm-framework-compagg]. Composition Framework [IPPM-Framework].
The primary recommendation here is to test using durations that are The primary recommendation here is to test using durations that are
similar in length to the session time of interest. This applies to similar in length to the session time of interest. This applies to
both IPDV and PDV, but is possibly more relevant for PDV since the both IPDV and PDV, but is possibly more relevant for PDV since the
duration determines how often the D_min will be determined, and the duration determines how often the D_min will be determined, and the
size of the associated sample. size of the associated sample.
8.5. Clock Sync Options 8.5. Clock Sync Options
As with one-way delay measurements, local clock synchronization is an As with one-way-delay measurements, local clock synchronization is an
important matter for delay variation measurements. important matter for delay variation measurements.
There are several options available: There are several options available:
1. Global Positioning System receivers 1. Global Positioning System receivers
2. In some parts of the world, Cellular Code Division Multiple 2. In some parts of the world, Cellular Code Division Multiple
Access (CDMA) systems distribute timing signals that are derived Access (CDMA) systems distribute timing signals that are derived
from GPS and traceable to UTC. from GPS and traceable to UTC.
3. Network Time Protocol [RFC1305] is a convenient choice in many 3. Network Time Protocol [RFC1305] is a convenient choice in many
cases, but usually offers lower accuracy than the options above. cases, but usually offers lower accuracy than the options above.
When clock synchronization is inconvenient or subject to appreciable When clock synchronization is inconvenient or subject to appreciable
errors, then round-trip measurements may give a cumulative indication errors, then round-trip measurements may give a cumulative indication
of the delay variation present on both directions of the path. of the delay variation present on both directions of the path.
However, delay distributions are rarely symmetrical, so it is However, delay distributions are rarely symmetrical, so it is
difficult to infer much about the one-way delay variation from round- difficult to infer much about the one-way-delay variation from round-
trip measurements. Also, measurements on asymmetrical paths add trip measurements. Also, measurements on asymmetrical paths add
complications for the one-way delay metric. complications for the one-way-delay metric.
8.6. Distinguishing Long Delay from Loss 8.6. Distinguishing Long Delay from Loss
Lost and delayed packets are separated by a waiting time threshold. Lost and delayed packets are separated by a waiting time threshold.
Packets that arrive at the measurement destination within their Packets that arrive at the measurement destination within their
waiting time have finite delay and are not lost. Otherwise, packets waiting time have finite delay and are not lost. Otherwise, packets
are designated lost and their delay is undefined. Guidance on are designated lost and their delay is undefined. Guidance on
setting the waiting time threshold may be found in [RFC2680] and setting the waiting time threshold may be found in [RFC2680] and
[I-D.morton-ippm-reporting-metrics]. [IPPM-Reporting].
In essence, [I-D.morton-ippm-reporting-metrics] suggests to use a In essence, [IPPM-Reporting] suggests to use a long waiting time to
long waiting time to serve network characterization and revise serve network characterization and revise results for specific
results for specific application delay thresholds as needed. application delay thresholds as needed.
8.7. Accounting for Packet Reordering 8.7. Accounting for Packet Reordering
Packet reordering, defined in [RFC4737], is essentially an extreme Packet reordering, defined in [RFC4737], is essentially an extreme
form of delay variation where the packet stream arrival order differs form of delay variation where the packet stream arrival order differs
from the sending order. from the sending order.
PDV results are not sensitive to packet arrival order, and are not PDV results are not sensitive to packet arrival order, and are not
affected by reordering other than to reflect the more extreme affected by reordering other than to reflect the more extreme
variation. variation.
skipping to change at page 34, line 45 skipping to change at page 35, line 11
when sequence has changed. If IPDV measurements are made without when sequence has changed. If IPDV measurements are made without
regard to packet arrival order, the IPDV will be under-reported when regard to packet arrival order, the IPDV will be under-reported when
reordering occurs. reordering occurs.
8.8. Results Representation and Reporting 8.8. Results Representation and Reporting
All of the references that discuss or define delay variation suggest All of the references that discuss or define delay variation suggest
ways to represent or report the results, and interested readers ways to represent or report the results, and interested readers
should review the various possibilities. should review the various possibilities.
For example, [I-D.morton-ippm-reporting-metrics] suggests to report a For example, [IPPM-Reporting] suggests reporting a pseudo-range of
pseudo range of delay variation based on calculating the difference delay variation based on calculating the difference between a high
between a high percentile of delay and the minimum delay. The 99.9%- percentile of delay and the minimum delay. The 99.9th percentile
ile minus the minimum will give a value that can be compared with minus the minimum will give a value that can be compared with
objectives in [Y.1541]. objectives in [Y.1541].
9. IANA Considerations 9. Security Considerations
This document makes no request of IANA.
Note to RFC Editor: this section may be removed on publication as an
RFC.
10. Security Considerations
The security considerations that apply to any active measurement of The security considerations that apply to any active measurement of
live networks are relevant here as well. See the security live networks are relevant here as well. See the "Security
considerations sections in [RFC2330], [RFC2679], [RFC3393], Considerations" sections in [RFC2330], [RFC2679], [RFC3393],
[RFC3432], and[RFC4656]. [RFC3432], and[RFC4656].
Security considerations do not contribute to the selection of PDV or Security considerations do not contribute to the selection of PDV or
IPDV forms of delay variation. IPDV forms of delay variation, because measurements using these
metrics involve exactly the same security issues.
11. Acknowledgements 10. Acknowledgments
The authors would like to thank Phil Chimento for his suggestion to The authors would like to thank Phil Chimento for his suggestion to
employ the convention of conditional distributions of Delay to deal employ the convention of conditional distributions of delay to deal
with packet loss, and his encouragement to "write the memo" after with packet loss, and his encouragement to "write the memo" after
hearing "the talk" on this topic at IETF-65. We also acknowledge hearing "the talk" on this topic at IETF 65. We also acknowledge
constructive comments from Alan Clark, Loki Jorgenson, Carsten constructive comments from Alan Clark, Loki Jorgenson, Carsten
Schmoll, and Robert Holley. Schmoll, and Robert Holley.
12. Appendix on Calculating the D(min) in PDV 11. Appendix on Calculating the D(min) in PDV
Practitioners have raised questions several questions that this Practitioners have raised several questions that this section intends
section intends to answer: to answer:
- how is this D_min calculated? Is it DV(99%) as mentioned in - How is this D_min calculated? Is it DV(99%) as mentioned in
[Krzanowski]? [Krzanowski]?
- do we need to keep all the values from the interval, then take the - Do we need to keep all the values from the interval, then take the
minimum? Or do we keep the minimum from previous intervals? minimum? Or do we keep the minimum from previous intervals?
The value of D_min used as the reference delay for PDV calculations The value of D_min used as the reference delay for PDV calculations
is simply the minimum delay of all packets in the current sample. is simply the minimum delay of all packets in the current sample.
The usual single value summary of the PDV distribution is D_99.9%-ile The usual single value summary of the PDV distribution is D_(99.9th
minus D_min. percentile) minus D_min.
It may be appropriate to segregate sub-sets and revise the minimum It may be appropriate to segregate sub-sets and revise the minimum
value during a sample. For example, if it can be determined with value during a sample. For example, if it can be determined with
certainty that the path has changed by monitoring the Time to Live or certainty that the path has changed by monitoring the Time to Live or
Hop Count of arriving packets, this may be sufficient justification Hop Count of arriving packets, this may be sufficient justification
to reset the minimum for packets on the new path. There is also a to reset the minimum for packets on the new path. There is also a
simpler approach to solving this problem: use samples collected over simpler approach to solving this problem: use samples collected over
short evaluation intervals (on the order of minutes). Intervals with short evaluation intervals (on the order of minutes). Intervals with
path changes may be more interesting from the loss or one-way delay path changes may be more interesting from the loss or one-way-delay
perspective (possibly failing to meet one or more SLAs), and it may perspective (possibly failing to meet one or more SLAs), and it may
not be necessary to conduct delay variation analysis. Short not be necessary to conduct delay variation analysis. Short
evaluation intervals are preferred for measurements that serve as a evaluation intervals are preferred for measurements that serve as a
basis for troubleshooting, since the results are available to report basis for troubleshooting, since the results are available to report
soon after collection. soon after collection.
It is not necessary to store all delay values in a sample when It is not necessary to store all delay values in a sample when
storage is a major concern. D_min can be found by comparing each new storage is a major concern. D_min can be found by comparing each new
singleton value with the current value and replacing it when singleton value with the current value and replacing it when
required. In a sample with 5000 packets, evaluation of the 99.9%-ile required. In a sample with 5000 packets, evaluation of the 99.9th
can also be achieved with limited storage. One method calls for percentile can also be achieved with limited storage. One method
storing the top 50 delay singletons and revising the top value list calls for storing the top 50 delay singletons and revising the top
each time 50 more packets arrive. value list each time 50 more packets arrive.
13. References 12. References
13.1. Normative References 12.1. 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, March 1997. Requirement Levels", BCP 14, RFC 2119, March 1997.
[RFC2330] Paxson, V., Almes, G., Mahdavi, J., and M. Mathis, [RFC2330] Paxson, V., Almes, G., Mahdavi, J., and M. Mathis,
"Framework for IP Performance Metrics", RFC 2330, "Framework for IP Performance Metrics", RFC 2330,
May 1998. May 1998.
[RFC2679] Almes, G., Kalidindi, S., and M. Zekauskas, "A One-way [RFC2679] Almes, G., Kalidindi, S., and M. Zekauskas, "A One-
Delay Metric for IPPM", RFC 2679, September 1999. way Delay Metric for IPPM", RFC 2679,
September 1999.
[RFC2680] Almes, G., Kalidindi, S., and M. Zekauskas, "A One-way [RFC2680] Almes, G., Kalidindi, S., and M. Zekauskas, "A One-
Packet Loss Metric for IPPM", RFC 2680, September 1999. way Packet Loss Metric for IPPM", RFC 2680,
September 1999.
[RFC3393] Demichelis, C. and P. Chimento, "IP Packet Delay Variation [RFC3393] Demichelis, C. and P. Chimento, "IP Packet Delay
Metric for IP Performance Metrics (IPPM)", RFC 3393, Variation Metric for IP Performance Metrics
November 2002. (IPPM)", RFC 3393, November 2002.
[RFC3432] Raisanen, V., Grotefeld, G., and A. Morton, "Network [RFC3432] Raisanen, V., Grotefeld, G., and A. Morton,
performance measurement with periodic streams", RFC 3432, "Network performance measurement with periodic
November 2002. streams", RFC 3432, November 2002.
[RFC4090] Pan, P., Swallow, G., and A. Atlas, "Fast Reroute [RFC4090] Pan, P., Swallow, G., and A. Atlas, "Fast Reroute
Extensions to RSVP-TE for LSP Tunnels", RFC 4090, Extensions to RSVP-TE for LSP Tunnels", RFC 4090,
May 2005. May 2005.
[RFC4656] Shalunov, S., Teitelbaum, B., Karp, A., Boote, J., and M. [RFC4656] Shalunov, S., Teitelbaum, B., Karp, A., Boote, J.,
Zekauskas, "A One-way Active Measurement Protocol and M. Zekauskas, "A One-way Active Measurement
(OWAMP)", RFC 4656, September 2006. Protocol (OWAMP)", RFC 4656, September 2006.
[RFC4737] Morton, A., Ciavattone, L., Ramachandran, G., Shalunov, [RFC4737] Morton, A., Ciavattone, L., Ramachandran, G.,
S., and J. Perser, "Packet Reordering Metrics", RFC 4737, Shalunov, S., and J. Perser, "Packet Reordering
November 2006. Metrics", RFC 4737, November 2006.
13.2. Informative References 12.2. Informative References
[COM12.D98] [COM12.D98] Clark, A., "Analysis, measurement and modelling of
Clark, Alan., "ITU-T Delayed Contribution COM 12 - D98, Jitter", ITU-T Delayed Contribution COM 12 - D98,
"Analysis, measurement and modelling of Jitter"",
January 2003. January 2003.
[Casner] "A Fine-Grained View of High Performance Networking, NANOG [Casner] Casner, S., Alaettinoglu, C., and C. Kuan, "A Fine-
22 Conf.; http://www.nanog.org/mtg-0105/agenda.html", May Grained View of High Performance Networking",
20-22 2001. NANOG 22, May 20-22, 2001,
<http://www.nanog.org/mtg-0105/agenda.html>.
[Cia03] "Standardized Active Measurements on a Tier 1 IP Backbone, [Cia03] Ciavattone, L., Morton, A., and G. Ramachandran,
IEEE Communications Mag., pp 90-97.", June 2003. "Standardized Active Measurements on a Tier 1 IP
Backbone", IEEE Communications Magazine, p. 90-97,
June 2003.
[Demichelis] [Demichelis] Demichelis, C., "Packet Delay Variation Comparison
http://www.advanced.org/ippm/archive.3/att-0075/ between ITU-T and IETF Draft Definitions",
01-pap02.doc, "Packet Delay Variation Comparison between November 2000, <http://www.advanced.org/ippm/
ITU-T and IETF Draft Definitions", November 2000. archive.3/att-0075/01-pap02.doc>.
[G.1020] ITU-T Recommendation G.1020, ""Performance parameter [G.1020] ITU-T, "Performance parameter definitions for the
definitions for the quality of speech and other voiceband quality of speech and other voiceband applications
applications utilizing IP networks"", 2006. utilizing IP networks", ITU-T
Recommendation G.1020, 2006.
[G.1050] ITU-T Recommendation G.1050, ""Network model for [G.1050] ITU-T, "Network model for evaluating multimedia
evaluating multimedia transmission performance over transmission performance over Internet Protocol",
Internet Protocol"", November 2005. ITU-T Recommendation G.1050, November 2005.
[I-D.ietf-ippm-framework-compagg] [I.356] ITU-T, "B-ISDN ATM Layer Cell Transfer
Morton, A., "Framework for Metric Composition", Performance", ITU-T Recommendation I.356,
draft-ietf-ippm-framework-compagg-07 (work in progress), March 2000.
October 2008.
[I-D.ietf-ippm-spatial-composition] [IPPM-Framework] Morton, A., "Framework for Metric Composition",
Morton, A. and E. Stephan, "Spatial Composition of Work in Progress, October 2008.
Metrics", draft-ietf-ippm-spatial-composition-07 (work in
progress), July 2008.
[I-D.morton-ippm-reporting-metrics] [IPPM-Reporting] Morton, A., Ramachandran, G., and G. Maguluri,
Morton, A., Ramachandran, G., and G. Maguluri, "Reporting "Reporting Metrics: Different Points of View", Work
Metrics: Different Points of View", in Progress, January 2009.
draft-morton-ippm-reporting-metrics-06 (work in progress),
January 2009.
[I.356] ITU-T Recommendation Y.1540, "B-ISDN ATM layer cell [IPPM-Spatial] Morton, A. and E. Stephan, "Spatial Composition of
transfer performance", March 2000. Metrics", Work in Progress, July 2008.
[Krzanowski] [Krzanowski] Presentation at IPPM, IETF-64, "Jitter Definitions:
Presentation at IPPM, IETF-64, "Jitter Definitions: What What is What?", November 2005.
is What?", November 2005.
[Li.Mills] [Li.Mills] Li, Q. and D. Mills, "The Implications of Short-
Li, Quong. and David. Mills, ""The Implications of Short- Range Dependency on Delay Variation Measurement",
Range Dependency on Delay Variation Measurement", Second Second IEEE Symposium on Network Computing
IEEE Symposium on Network Computing and Applications", and Applications, 2003.
2003.
[Morton06] [Morton06] Morton, A., "A Brief Jitter Metrics Comparison, and
Morton, A., ""A Brief Jitter Metrics Comparison, and not not the last word, by any means...", slide
the last word, by any means...", Slide Presentation at presentation at IETF 65, IPPM Session, March 2006.
IETF-65, IPPM Session,", March 2006.
[RFC1305] Mills, D., "Network Time Protocol (Version 3) [RFC1305] Mills, D., "Network Time Protocol (Version 3)
Specification, Implementation", RFC 1305, March 1992. Specification, Implementation", RFC 1305,
March 1992.
[RFC3357] Koodli, R. and R. Ravikanth, "One-way Loss Pattern Sample [RFC3357] Koodli, R. and R. Ravikanth, "One-way Loss Pattern
Metrics", RFC 3357, August 2002. Sample Metrics", RFC 3357, August 2002.
[RFC3550] Schulzrinne, H., Casner, S., Frederick, R., and V. [RFC3550] Schulzrinne, H., Casner, S., Frederick, R., and V.
Jacobson, "RTP: A Transport Protocol for Real-Time Jacobson, "RTP: A Transport Protocol for Real-Time
Applications", STD 64, RFC 3550, July 2003. Applications", STD 64, RFC 3550, July 2003.
[Y.1540] ITU-T Recommendation Y.1540, "Internet protocol data [Y.1540] ITU-T, "Internet protocol data communication
communication service - IP packet transfer and service - IP packet transfer and availability
availability performance parameters", November 2007. performance parameters", ITU-T Recommendation
Y.1540, November 2007.
[Y.1541] ITU-T Recommendation Y.1541, "Network Performance [Y.1541] ITU-T, "Network Performance Objectives for IP-Based
Objectives for IP-Based Services", February 2006. Services", ITU-T Recommendation Y.1541,
February 2006.
[Zhang.Duff] [Zhang.Duff] Zhang, Y., Duffield, N., Paxson, V., and S.
Zhang, Yin., Duffield, Nick., Paxson, Vern., and Scott. Shenker, "On the Constancy of Internet Path
Shenker, ""On the Constancy of Internet Path Properties", Properties", Proceedings of ACM SIGCOMM Internet
Proceedings of ACM SIGCOMM Internet Measurement Measurement Workshop, November 2001.
Workshop,", November 2001.
Authors' Addresses Authors' Addresses
Al Morton Al Morton
AT&T Labs AT&T Labs
200 Laurel Avenue South 200 Laurel Avenue South
Middletown,, NJ 07748 Middletown, NJ 07748
USA USA
Phone: +1 732 420 1571 Phone: +1 732 420 1571
Fax: +1 732 368 1192 Fax: +1 732 368 1192
Email: acmorton@att.com EMail: acmorton@att.com
URI: http://home.comcast.net/~acmacm/ URI: http://home.comcast.net/~acmacm/
Benoit Claise Benoit Claise
Cisco Systems, Inc. Cisco Systems, Inc.
De Kleetlaan 6a b1 De Kleetlaan 6a b1
Diegem, 1831 Diegem, 1831
Belgium Belgium
Phone: +32 2 704 5622 Phone: +32 2 704 5622
Fax: EMail: bclaise@cisco.com
Email: bclaise@cisco.com
URI:
 End of changes. 157 change blocks. 
418 lines changed or deleted 406 lines changed or added

This html diff was produced by rfcdiff 1.35. The latest version is available from http://tools.ietf.org/tools/rfcdiff/