wdiff draft-ietf-ippm-model-based-metrics-00.txt draft-ietf-ippm-model-based-metrics-01.txt

IP Performance Working Group M. Mathis
Internet-Draft Google, Inc
Intended status: Experimental A. Morton
Expires: December 23, 2013 April 24, 2014 AT&T Labs
June
October 21, 2013

Model Based Bulk Performance Metrics
draft-ietf-ippm-model-based-metrics-00.txt
draft-ietf-ippm-model-based-metrics-01.txt

Abstract

We introduce a new class of model based metrics designed to determine
if a long network path can meet predefined end-to-end application
performance
targets. This is done by subpath at a time testing -- targets by applying a suite of single property IP diagnostic tests to
successive subpaths of a long path.
In many cases these single property tests are based on existing IPPM
metrics, with the addition of success and validity criteria. subpaths. The subpath at a time tests are designed to facilitate IP providers
eliminating
exclude all known conditions that which might prevent the full end-to-
end end-to-end
path from meeting the users user's target application performance.

This approach makes it possible to to determine the IP performance
requirements needed to support the desired end-to-end TCP
performance. The IP metrics are based on traffic patterns that mimic
TCP or other transport protocol but are precomputed independently of
the actual behavior of TCP the transport protocol over the subpath under
test. This makes the measurements open loop, eliminating nearly all
of the difficulties encountered by traditional bulk transport
metrics, which rely fundamentally depend on congestion control equilibrium
behavior.

A natural consequence of this methodology is verifiable network
measurement: measurements from any given vantage point are repeatable can be
verified by repeating them from other vantage points.

Formatted: Fri Jun Mon Oct 21 18:23:29 15:42:35 PDT 2013

Status of this Memo

This Internet-Draft is submitted in full conformance with the
provisions of BCP 78 and BCP 79.

Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF). Note that other groups may also distribute
working documents as Internet-Drafts. The list of current Internet-
Drafts is at http://datatracker.ietf.org/drafts/current/.

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."

This Internet-Draft will expire on December 23, 2013. April 24, 2014.

This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents
(http://trustee.ietf.org/license-info) in effect on the date of
publication of this document. Please review these documents
carefully, as they describe your rights and restrictions with respect
to this document. Code Components extracted from this document must
include Simplified BSD License text as described in Section 4.e of
the Trust Legal Provisions and are provided without warranty as
described in the Simplified BSD License.

Table of Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.1. TODO . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 6
3. New requirements relative to RFC 2330 . . . . . . . . . . . . 8 9
4. Background . . . . . . . . . . . . . . . . . . . . . . . . . . 9 10
4.1. TCP properties . . . . . . . . . . . . . . . . . . . . . . 11 12
5. Common Models and Parameters . . . . . . . . . . . . . . . . . 12 14
5.1. Target End-to-end parameters . . . . . . . . . . . . . . . 13 14
5.2. Common Model Calculations . . . . . . . . . . . . . . . . 13 15
5.3. Parameter Derating . . . . . . . . . . . . . . . . . . . . 14 16
6. Common testing procedures . . . . . . . . . . . . . . . . . . 15 16
6.1. Traffic generating techniques . . . . . . . . . . . . . . 15 16
6.1.1. Paced transmission . . . . . . . . . . . . . . . . . . 15 16
6.1.2. Constant window pseudo CBR . . . . . . . . . . . . . . 16
6.1.2.1. 17
6.1.3. Scanned window pseudo CBR . . . . . . . . . . . . 16
6.1.3. . . 18
6.1.4. Concurrent or channelized testing . . . . . . . . . . 18
6.1.5. Intermittent Testing . . . . . . . . . . . . . . . . . 16
6.1.4. 19
6.1.6. Intermittent Scatter Testing . . . . . . . . . . . . . 17 20
6.2. Interpreting the Results . . . . . . . . . . . . . . . . . 17 20
6.2.1. Test outcomes . . . . . . . . . . . . . . . . . . . . 17 20
6.2.2. Statistical criteria for measuring run_length . . . . 17 21
6.2.3. Classifications of tests . . . . . . . . . . . . . . . 19
6.2.4. Reordering Tolerance . . . . . . . . . . . . . . . . . 20 23
6.3. Test Qualifications . . . . . . . . . . . . . . . . . . . 20 23
6.3.1. Verify the Traffic Generation Accuracy . . . . . . . . 20 23
6.3.2. Verify the absence of cross traffic . . . . . . . . . 21 24
6.3.3. Additional test preconditions . . . . . . . . . . . . 22 25
7. Single Property Diagnostic Tests . . . . . . . . . . . . . . . . . . . . 22 . . . 25
7.1. Basic Data and Loss Rate and Run Length Tests . . . . . . . . . . . . . . 22 25
7.1.1. Loss Rate Run Length at Paced Full Data Rate . . . . . . . . . . 22 26
7.1.2. Loss Rate run length at Full Data Windowed Rate . . . . . . . . . 23 26
7.1.3. Background Loss Rate Run Length Tests . . . . . . . . . . . . . . 23 26
7.2. Standing Queue tests . . . . . . . . . . . . . . . . . . . 24 26
7.2.1. Congestion Avoidance . . . . . . . . . . . . . . . . . 24 28
7.2.2. Buffer Bloat Bufferbloat . . . . . . . . . . . . . . . . . . . . . 25 28
7.2.3. Non excessive loss . . . . . . . . . . . . . . . . . . 28
7.2.4. Duplex Self Interference . . . . . . . . . . . . . . . 25 28
7.3. Slowstart tests . . . . . . . . . . . . . . . . . . . . . 25 29
7.3.1. Full Window slowstart test . . . . . . . . . . . . . . 25 29
7.3.2. Slowstart AQM test . . . . . . . . . . . . . . . . . . 26 29
7.4. Sender Rate Burst tests . . . . . . . . . . . . . . . . . 26
7.4.1. Sender TCP Send Offload (TSO) tests . . 29
7.5. Combined Tests . . . . . . . 26
7.4.2. Sender Full Window burst test . . . . . . . . . . . . 26
8. Combined Tests . . . 30
7.5.1. Sustained burst test . . . . . . . . . . . . . . . . . 30
7.5.2. Live Streaming Media . . . . 27
8.1. Sustained burst test . . . . . . . . . . . . . 31
8. Examples . . . . . . 27
9. Calibration . . . . . . . . . . . . . . . . . . . . . 32
8.1. Near serving HD streaming video . . . . 28
10. Acknowledgements . . . . . . . . . 32
8.2. Far serving SD streaming video . . . . . . . . . . . . . . 28
11. Informative References 32
8.3. Bulk delivery of remote scientific data . . . . . . . . . 33
9. Validation . . . . . . . . . . . 28
Appendix A. Model Derivations . . . . . . . . . . . . . . . 33
10. Acknowledgements . . . 29
Appendix B. old text . . . . . . . . . . . . . . . . . . . . 34
11. Informative References . . 29
B.1. An earlier document . . . . . . . . . . . . . . . . . . 35
Appendix A. Model Derivations . 30
B.2. End-to-end parameters from subpaths . . . . . . . . . . . 31
B.3. Per subpath parameters . . . . . . 36
A.1. Aggregate Reno . . . . . . . . . . . . 32
B.4. Version Control . . . . . . . . . . 37
A.2. CUBIC . . . . . . . . . . . 32 . . . . . . . . . . . . . . . 37
Appendix B. Version Control . . . . . . . . . . . . . . . . . . . 38
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 32 38

1. Introduction

Model based bulk performance metrics evaluate an Internet paths path's
ability to carry bulk data. TCP models are used to design a targeted
diagnostic suite (TDS) of IP performance tests which can be applied
independently to each subpath of the full end-to-end path. The A
targeted diagnostic suites are suite is constructed such that independent tests
of the subpaths will accurately predict if the full end-to-end path
can deliver bulk data at the specified performance target,
independent of the measurement vantage points or other details of the
test procedures used to measure each subpath.

Each test in the targeted diagnostic suite TDS consists of a precomputed traffic pattern and
statistical criteria for evaluating packet delivery.

TCP models are used to design traffic patterns that mimic TCP or
other bulk transport protocol operating at the target performance and
RTT over a full range of conditions, including flows that are bursty
at multiple time scales. The traffic patterns are computed in
advance based on the properties of the full end-to-end path and
independent of the properties of individual subpaths. As much as
possible the traffic is generated deterministically in ways that
minimizes the extent to which test methodology, measurement points,
measurement vantage or path partitioning effect the details of the
traffic.

Models are also used to compute the statistical criteria bounds on the packet delivery
statistics for
evaluating acceptable the IP diagnostics tests. performance. The criteria for
passing each test must be are determined from the end-to-end target
performance and are independent of the RTT or other properties of the subpath under test. In
addition to passing or failing, a test can be inconclusive if the
precomputed traffic pattern was not authentically generated, test
preconditions were not met or the measurement results were not
statistically significantly. significant.

TCP's ability to compensate for less than ideal network conditions is
fundamentally affected by the RTT and MTU of the end-to-end Internet
path that it traverses which are both traverses. The end-to-end path determines fixed properties of the end-to-
end path. bounds
on these parameters. The target values for these three parameters,
Data Rate, RTT and MTU, are determined by the application, its
intended use and the physical infrastructure over which it traverses. They is
intended to traverse. These parameters are used to inform the models
used to design the targeted TDS.

This document describes a framework for deriving the traffic and
delivery statistics for model based metrics. It does not fully
specify any measurement techniques. Important details such as packet
type-p selection, sampling techniques, vantage selection, etc are out
of scope for this document. We imagine Fully Specified Targeted
Diagnostic Suites (FSTDS), that fully defines all of these details.
We use TDS to refer to the subset of such a specification that is in
scope for this document. A TDS includes specification for the
traffic and delivery statistics for the diagnostic suite. tests themselves,
documentation of the models and any assumptions or derating used to
derive the test parameters and a description of the test setup used
to calibrate the models, as described in later sections.

Section 2 defines terminology used throughout this document.

It has been difficult to develop BTC metrics due to some overlooked
requirements described in Section 3 and some intrinsic problems with
using protocols for measurement, described in Section 4.

In Section 5 we describe the models and common parameters used to
derive the targeted diagnostic suite. In Section 6 we describe
common testing procedures used by all of the tests. procedures. Each subpath is evaluated using suite of
far simpler and more predictable single
property diagnostic tests described in
Section 7. In Section 8 describes some
combined tests that are more efficient we present three example TDS, one that might
be representative of HD video, when served fairly close to implement and deploy.
However, if they fail they may not clearly indicate the nature user,
a second that might be representative of the
problem. standard video, served from
a greater distance, and a third that might be representative of an
network designed to support high performance bulk download.

There exists a small risk that model based metric itself might yield
a false pass result, in the sense that every subpath of an end-to-end
path passes every IP diagnostic test and yet a real application falls
to attain the performance target over the end-to-end path. If this
happens, then the calibration validation procedure described in Section 9 needs
to be used to validate prove and potentially revise the models.

Future document will define model based metrics for other traffic
classes and application types, such as real time. time streaming media.

1.1. TODO

Please send comments on this draft to ippm@ietf.org. See
http://goo.gl/02tkD for more information including: interim drafts,
an up to date todo list and information on contributing.

Formatted: Fri Jun Mon Oct 21 18:23:29 15:42:35 PDT 2013

2. Terminology

Terminology about paths, etc. See [RFC2330] and
[I-D.morton-ippm-lmap-path].

[data] sender Host sending data and receiving ACKs, typically via
TCP.
[data] receiver Host receiving data and sending ACKs, typically via
TCP.
subpath A portion of the full path. Note that there is no
requirement that subpaths be non-overlapping.
Measurement Point Measurement points as described in
[I-D.morton-ippm-lmap-path].
test path A path between two measurement points that includes a
subpath of the end-to-end path under test, plus possibly
additional infrastructure between the measurement points and the
subpath.
[Dominant] Bottleneck The Bottleneck that determines a flow's self
clock. It generally determines the traffic statistics for the
entire path. See Section 4.1.
front path The subpath from the data sender to the dominant
bottleneck.
back path The subpath from the dominant bottleneck to the receiver.
return path The path taken by the ACKs from the data receiver to the
data sender.
cross traffic Other, potentially interfering, traffic competing for
resources (network and/or queue capacity).

Properties determined by the end-to-end path and application. They
are described in more detail in Section 5.1.

end-to-end target parameters:

Application or transport performance
goals Data Rate General term for the end-to-end path. They include data rate as seen by the target
application above the transport layer. This is the payload data
rate,
RTT and MTU described below.
Target excludes TCP/IP (or other protocol) headers and
retransmits.
Link Data Rate: The application or ultimate user's performance
goal. Rate General term for the data rate as seen by the link or
lower layers. It includes transport and IP headers, retransmits
and other transport layer overhead. This document is agnostic as
to whether the link data rate includes or excludes framing, MAC or
other lower layer overheads, except that they must be treated
uniformly.
end-to-end target parameters: Application or transport performance
goals for the end-to-end path. They include the target data rate,
RTT and MTU described below.
Target Data Rate: The application or ultimate user's performance
goal. When converted to link data rate, it must be slightly
smaller than the actual link data rate, otherwise there is no
margin for compensating for RTT or other path protperties. properties. These
test will be excessively brittle if the target data rate does not
include any built in headroom.

Target RTT (Round Trip Time): The baseline (minimum) RTT of the
longest end-to-end path the over which the application
must expects to
meet the target performance. This must be specified considering
authentic packets sizes: MTU sized packets on the forward path,
header_overhead sized packets on the return (ACK) path.
Target MTU (Maximum Transmission Unit): The maximum MTU supported by
the end-to-end path the over which the application expects to meet
the target performance. Assume 1500 Bytes per packet unless
otherwise specified. If some subpath forces a smaller MTU, then
it becomes the target MTU, and all subpaths model calculations and subpath
tests must be tested
with use the same smaller MTU.
Effective Bottleneck Data Rate: This is the bottleneck data rate
that might be inferred from the ACK stream, by looking at how much
data the ACK stream reports was delivered per unit time. See
Section 4.1 for more details.
Permitted Number of Connections: The target rate can be more easily
obtained by dividing the traffic across more than one connection.
In general the number of concurrent connections is determined by
the application, however see the comments below on multiple
connections.
[sender] [interface] rate: The burst data rate, constrained by the
data sender's interfaces. Today 1 or 10 Gb/s are typical.
Header overhead: The IP and TCP header sizes, which are the portion
of each MTU not available for carrying application payload.
Without loss of generality this is assumed to be the size for
returning acknowledgements (ACKs). For TCP, the Maximum Segment
Size (MSS) is the Target MTU minus the header overhead.

Terminology about paths, etc. See [RFC2330] and
[I-D.morton-ippm-lmap-path].

[data] sender Host sending data and receiving ACKs, typically via
TCP.
[data] receiver Host receiving data

Basic parameters common to models and sending ACKs, typically via
TCP. subpath Subpath as defined in [RFC2330].
Measurement Point Measurement points as tests. They are
described in
[I-D.morton-ippm-lmap-path].
test path more detail in Section 5.2.

pipe size A path between two measurement points that includes a
subpath of the end-to-end path under test, plus possibly
additional infrastructure between the measurement points and the
subpath.
[Dominant] Bottleneck The Bottleneck that determines a flow's self
clock. It generally determines the traffic statistics for the
entire path. See Section 4.1.
front path The subpath from the data sender to the dominant
bottleneck.
back path The subpath from the dominant bottleneck to the receiver.
return path The path taken by the ACKs from the data receiver to the
data sender.
cross traffic Other, potentially interfering, traffic competing for
resources (network and/or queue capacity).

Basic parameters common to all models and subpath tests. They are
described in more detail in Section 5.2.

@ @@@
pipe size The number general term for number of packets needed in flight (the
window size) to exactly fill some network path or sub path. The subpath. This
is the window size which in normally the onset of queueing.
target_pipe_size: The number of packets in flight (the window size)
needed to exactly meet the target rate, with a single stream and
no cross traffic for the specified target data rate, RTT and MTU.
subpath pipe size
run length Observed, A general term for the observed, measured or specified
number of packets that are (to be) delivered between losses or ECN
marks. Nominally one over the loss or ECN marking probability.
target_run_length Required run length computed from the target data
rate, RTT and MTU.
reference_target_run_length: One specific conservative estimate of
the number of packets that must be delivered between loss episodes
in most diagnostic tests.

Ancillary parameters used for some tests

derating: Under some conditions the standard models are too
conservative. The modeling framework permits some latitude in
relaxing or derating some specific test parameters as described in
Section 5.3.

Test types [These need work] 5.3 in exchange for a more stringent TDS validation
procedures, described in Section 9.

subpath_data_rate The maximum IP data rate supported by a subpath.
This typically includes TCP/IP overhead, including headers,
retransmits, etc.
test_path_RTT The RTT (using appropriate packet sizes) between two
measurement points.
test_path_pipe The amount of data necessary to fill a test path.
Nominally the test path RTT times the subpath_data_rate (which
should be part of the end-to-end subpath).
test_window The window necessary to meet the target_rate over a
subpath. Typically test_window=target_data_rate*test_RTT/
target_MTU.

Tests can be classified into groups according to their applicability

Capacity tests determine if a network subpath has sufficient
capacity tests: For "capacity tests" is required that as to deliver the target performance. As long as the test
traffic is within the proper envelope for the target end-to-
end end-to-end
performance, the average packet losses or ECN must be below the
threshold computed by the model.
Engineering tests: Engineering tests verify As such, they reflect parameters
that can transition from passing to failing as a consequence of
additional presented load or the actions of other network users.
By definition, capacity tests also consume significant network
resources (data capacity and/or buffer space), and the subpath under test interacts well
schedules must be balanced by their cost.
Monitoring tests are design to capture the most important aspects of
a capacity test, but without causing unreasonable ongoing load
themselves. As such they may miss some details of the network
performance, but can serve as a useful reduced cost proxy for a
capacity test.
Engineering tests evaluate how network algorithms (such as AQM and
channel allocation) interact with TCP style self clocked protocols using
and adaptive congestion control based on packet loss and ECN
marks. These tests are likely to have complicated interactions
with other traffic and under some conditions can be inversely
sensitive to load. For example "AQM Tests" a test to verify that when the presented load
exceeds the capacity of the subpath, the subpath signals for the
transport protocol to slow down, by appropriately an AQM
algorithm causes ECN marking marks or
dropping some of packet drops early enough to limit
queue occupancy may experience a false pass results in the packets. Note while that
presence of bursty cross traffic traffic. It is
can cause capacity tests to fail, it has the potential to cause
AQM important that
engineering tests to false pass, which be performed under a wide range of conditions,
including both in situ and bench testing, and over a wide variety
of load conditions. Ongoing monitoring is why AQM tests require separate
test procedures. less likely to be
useful for engineering tests, although sparse in situ testing
might be appropriate.

3. New requirements relative to RFC 2330

Model Based

[Move this entire section to a future paper]
Model Based Metrics are designed to fulfil fulfill some additional
requirement that were not recognized at the time RFC 2330 [RFC2330]
was written. These missing requirements may have significantly
contributed to policy difficulties in the IP measurement space. Some
additional requirements are:
o Metrics must be actionable by the ISP - they have to be
interpreted in terms of behaviors or properties at the IP or lower
layers, that an ISP can test, repair and verify.
o Metrics must be vantage point invariant over a significant range
of measurement point choices (e.g., measurement points as
described in [I-D.morton-ippm-lmap-path]), including off path
measurement points. The only requirements on MP selection should
be that the portion of the path that is not under test is
effectively ideal (or is non ideal in calibratable ways) and the
end-to-end
RTT between MPs is below some reasonable bound.
o Metrics must be repeatable by multiple parties. It must be
possible for different parties to make the same measurement and
observe the same results. In particular it is specifically
important that both a consumer (or their delegate) and ISP be able
to perform the same measurement and get the same result.

NB: All of the metric requirements in RFC 2330 should be reviewed and
potentially revised. If such a document is opened soon enough, this
entire section should be dropped.

4. Background

[Move to a future paper, abridge here, ]

At the time the IPPM WG was chartered, sound Bulk Transport Capacity
measurement was known to be beyond our capabilities. By hindsight it
is now clear why it is such a hard problem:
o TCP is a control system with circular dependencies - everything
affects performance, including components that are explicitly not
part of the test.
o Congestion control is an equilibrium process, transport protocols
change the network (raise loss probability and/or RTT) to conform
to their behavior.
o TCP's ability to compensate for network flaws is directly
proportional to the number of roundtrips per second (i.e.
inversely proportional to the RTT). As a consequence a flawed
link may pass a short RTT local test even though it fails when the
path is extended by a perfect network to some larger RTT.
o TCP has a meta Heisenberg problem - Measurement and cross traffic
interact in unknown and ill defined ways. The situation is
actually worse than the traditional physics problem where you can
at least estimate the relative momentum of the measurement and
measured particles. For network measurement you can not in
general determine the relative "elasticity" of the measurement
traffic and cross traffic, so you can not even gage the relative
magnitude of their effects on each other.

The MBM approach is to "open loop" TCP by precomputing traffic
patterns that are typically generated by TCP operating at the given
target parameters, and evaluating delivery statistics (losses (losses, ECN
marks and delay). In this approach the measurement software
explicitly controls the data rate, transmission pattern or cwnd
(TCP's primary congestion control state variables) to create
repeatable traffic patterns that mimic TCP behavior but are
independent of the actual network behavior of the subpath under test.
These patterns are manipulated to probe the network to verify that it
can deliver all of the traffic patterns that a transport protocol is
likely to generate under normal operation at the target rate and RTT.

Models are used to determine the actual test parameters (burst size,
loss rate, etc) from the target parameters. The basic method is to
use models to estimate specific network properties required to
sustain a given transport flow (or set of flows), and using a suite
of metrics to confirm that the network meets the required properties.

A network is expected to be able to sustain a Bulk TCP flow of a
given data rate, MTU and RTT when the following conditions are met:
o The raw link rate is higher than the target data rate.
o The raw packet loss rate run length is lower larger than required by a suitable
TCP performance model
o There is sufficient buffering at the dominant bottleneck to absorb
a slowstart rate burst large enough to get the flow out of
slowstart at a suitable window size.
o There is sufficient buffering in the front path to absorb and
smooth sender interface rate bursts at all scales that are likely
to be generated by the application, any channel arbitration in the
ACK path or other mechanisms.
o When there is a standing queue at a bottleneck for a shared media
subpath, there are suitable bounds on how the data and ACKs
interact, for example due to the channel arbitration mechanism.
o When there is a slowly rising standing queue at the bottleneck the
onset of packet loss has to be at an appropriate point (time or
queue depth) and progressive.

The tests to verify these condition are described in Section 7.

Note that this procedure is not invertible: a

A singleton [RFC2330] measurement is a pass/fail evaluation of a
given path or subpath at a given performance. Measurements Note that measurements
to confirm that a link passes at one particular performance may might not
be generally be useful to predict if the link will pass at a different
performance.

Although they are not invertible, they do

A TDS does have several other valuable properties, such as natural ways to
define several different composition metrics [RFC5835].

[Add text on algebra on metrics (A-Frame from [RFC2330]) and
tomography.] The Spatial Composition of fundamental IPPM metrics has
been studied and standardized. For example, the algebra to combine
empirical assessments of loss ratio to estimate complete path
performance is described in section 5.1.5. of [RFC6049]. We intend
to use this and other composition metrics as necessary.

We are developing a tool that can perform many of the tests described
here[MBMSource].

4.1. TCP properties

[Move this entire section to a future paper]

TCP and SCTP are self clocked protocols. The dominant steady state
behavior is to have an approximately fixed quantity of data and
acknowledgements (ACKs) circulating in the network. The receiver
reports arriving data by returning ACKs to the data sender, the data
sender most frequently responds by sending exactly the same quantity
of data back into the network. The quantity of data plus the data
represented by ACKs circulating in the network is referred to as the
window. The mandatory congestion control algorithms incrementally
adjust the widow by sending slightly more or less data in response to
each ACK. The fundamentally important property of this systems is
that it is entirely self clocked: The data transmissions are a
reflection of the ACKs that were delivered by the network, the ACKs
are a reflection of the data arriving from the network.

A number of phenomena can cause bursts of data, even in idealized
networks that are modeled as simple queueing systems.

During slowstart the data rate is doubled on each RTT by sending
twice as much data as was delivered to the receiver. receiver on the prior RTT.
For slowstart to be able to fill such a network the network must be
able to tolerate slowstart bursts up to the full pipe size inflated
by the anticipated window reduction on the first loss. loss or ECN mark.
For example, with classic Reno congestion control, an optimal
slowstart has to end with a burst that is twice the bottleneck rate
for exactly one RTT in duration. This burst causes a queue which is
exactly equal to the pipe size (the window is exactly twice the pipe
size) so when the window is halved, the new window will be exactly
the pipe size.

Another source of bursts are application pauses. If the application
pauses (stops reading or writing data) for some fraction of one RTT,
state-of-the-art TCP to "catches up" to the earlier window size by
sending a burst of data at the full sender interface rate. To fill
such a network with a realistic application, the network has to be
able to tolerate interface rate bursts from the data sender large
enough to cover the worst case application pause. pauses.

Note that if the bottleneck data rate is significantly slower than
the rest of the path, the slowstart bursts will not cause significant
queues anywhere else along the path; they primarily exercise the
queue at the dominant bottleneck. Furthermore Furthermore, although the
interface rate bursts caused by the application are likely to be
smaller than
burst at the last RTT burst of a slowstart, they are at a higher rate so
they can exercise queues at arbitrary points along the "front path"
from the data sender up to and including the queue at the bottleneck.

For many network technologies a simple queueing model does not apply:
the network schedules, thins or otherwise alters the timing of ACKs
and data
stream, data, generally to raise the efficiency of the channel allocation
process when confronted with relatively widely spaced small ACKs.
These efficiency strategies are ubiquitous for wireless and other half
duplex duplex, wireless
or broadcast media.

Altering the ACK stream generally has two consequences: raising the
effective bottleneck data rate making slowstart burst at higher rates
(possibly as high as the sender's interface rate) and effectively
raising the RTT by the time that the ACKs were postponed. The first
effect can be partially mitigated by reclocking ACKs once they are
through
beyond the bottleneck on the return path to the sender, however this
further raises the effective RTT. The most extreme example of this
class of behaviors is a half duplex channel that is never released
until the current sender end point has no pending traffic. Such
environments
intrinsically cause self clocked protocols revert to extremely
inefficient stop and wait behavior, where they send an entire window
of data as a single burst, followed by the entire window of ACKs on
the return path.

If a particular end-to-end path contains a link or device that alters
the ACK stream, then the entire path from the sender up to the
bottleneck must be tested at the burst parameters implied by the ACK
scheduling algorithms. algorithm. The most important parameter is the Effective
Bottleneck Data Rate, which is the average rate at which the ACKs
advance snd.una. Note that thinning the ACKs (relying on the
cumulative nature of seg.ack to permit discarding some ACKs) is
implies an effectively infinite bottleneck data rate.

To verify that a path can meet the performance target, Model Based
Metrics need it is
necessary to independently confirm that the entire path can tolerate
bursts of in the dimensions that are likely to be induced by the
application and any data or ACK scheduling. scheduling anywhere in the path. Two
common cases are the most important: slowstart bursts of with more than the
target_pipe_size data at twice the
effective bottleneck data rate; and somewhat smaller sender interface
rate bursts.

5. Common Models

The slowstart rate bursts must be at least as least as large
target_pipe_size packets and Parameters

Transport performance models are used to derive should be twice as large (so the test parameters peak
queue occupancy at the dominant bottleneck would be approximately
target_pipe_size).

There is no general model for test suites how well the network needs to tolerate
sender interface rate bursts. All existing TCP implementations send
full sized full rate bursts under some typically uncommon conditions,
such as application pauses that approximately match the RTT, or when
ACKs are lost or thinned. Strawman: partial window bursts (some
fraction of simple diagnostics from target_pipe_size) should be tolerated without
significantly raising the end-to-end target
parameters loss probability. Full target_pipe_size
bursts may slightly increase the loss probability. Interface rate
bursts as large as twice target_pipe_size should not cause
deterministic packet drops.

5. Common Models and additional ancillary parameters. Parameters

5.1. Target End-to-end parameters

The target end to end parameters are the target data rate, target RTT
and target MTU as defined in Section 2 These parameters are
determined by the needs of the application or the ultimate end user
and the end-to-end Internet path. They are in units that make path over which the application is
expected to operate. The target parameters are in units that make
sense to the upper layer: payload bytes delivered, excluding header delivered to the application,
above TCP. They exclude overheads for IP, associated with TCP and IP
headers, retransmitts and other protocol.

Ancillary protocols (e.g. DNS). In addition,
other end-to-end parameters include the effective bottleneck data
rate, the sender interface data rate and the
permitted number of connections (numb_cons).

The use of multiple connections has been very controversial since the
beginning of TCP/IP header sizes
(overhead).

Note that the World-Wide-Web[first complaint]. Modern browsers
open many connections [BScope]. Experts associated with IETF
transport area have frequently spoken against this practice [long
list]. It is not inappropriate target parameters can be specified for a hypothetical
path, for example to assume some small number construct TDS designed for bench testing in the
absence of
concurrent connections (e.g. 4 a real application, or 6), to compensate for limitation in
TCP. However, choosing too large a real physical test, for in
situ testing of production infrastructure.

The number is at risk of being
interpreted as concurrent connections is explicitly not a signal by the web browser community that parameter to
this
practice has been embraced by model [unlike earlier drafts]. If a subpath requires multiple
connections in order to meet the Internet service provider
community. It may not specified performance, that must be desirable to send such a signal.
stated explicitly and the procedure described in Section 6.1.4
applies.

5.2. Common Model Calculations

The most important derived parameter is target_pipe_size (in
packets), which is the window size --- the number of packets needed
exactly meet the target rate, with numb_cons connections and no cross traffic for the specified
target RTT and MTU. It is given by:

target_pipe_size = (target_rate / numb_cons) target_rate * target_RTT / ( target_MTU -
header_overhead )

If the transport protocol (e.g. TCP) average window size is smaller
than this, it will not meet the target rate.

The reference_target_run_length, which reference target_run_length, is the most a very conservative model for the
minimum required spacing between losses, losses or ECN marks. The reference
target_run_length can derived as follows: assume the link_data_rate
subpath_data_rate is infinitesimally larger than the
target_data_rate. target_data_rate
plus the required header overheads. Then target_pipe_size also
predicts the onset of queueing. If the transport protocol (e.g.
TCP) has an average a window size that is larger than the target_pipe_size, the
excess packets will form raise the RTT, typically by forming a standing
queue at the bottleneck.

Assume the transport protocol is using standard Reno style Additive
Increase, Multiplicative Decrease congestion control [RFC5681], then [RFC5681] and
the receiver is using standard delayed ACKs. With delayed ACKs there
must be target_pipe_size 2*target_pipe_size roundtrips between losses. Otherwise the
multiplicative window reduction triggered by a loss would cause the
network to be underfilled. Following [MSMO97], we We derive the number of packets between
losses from the area under the AIMD sawtooth following [MSMO97].
They must be no more frequent than every 1 in
(3/2)(target_pipe_size^2)
(3/2)*target_pipe_size*(2*target_pipe_size) packets. This provides the reference value
for simplifies
to:

target_run_length which is typically the number of packets that
must be delivered between loss episodes in the tests below:

reference_target_run_length = (3/2)(target_pipe_size^2) 3*(target_pipe_size^2)

Note that this calculation is very conservative and is based on a
number of assumptions that may not apply. Appendix A discusses these
assumptions and provides some alternative models. The If a less
conservative model is used, a fully specified TDS or FSTDS MUST
document the actual method for computing target_run_length MUST be documented along with
the rationale for the underlying assumptions and the ratio of chosen
target_run_length to
reference_target_run_length. @@@ MOVE

Although this document gives a lot of latitude for calculating the reference target_run_length calculated
above.

These two parameters, target_pipe_size and target_run_length, people designing suites
directly imply most of the individual parameters for the tests need to consider below.
Target_pipe_size is the effect window size, the amount of their choices on circulating data
required to meet the ongoing conversation target data rate, and tussle
about implies the relevance scale of "TCP friendliness" as an appropriate model for
capacity allocation. Choosing a target_run_length that is
substantially smaller than reference_target_run_length is equivalent
to saying the
bursts that it is appropriate for the transport research community
to abandon "TCP friendliness" as a fairness model and to develop more
aggressive Internet transport protocols, and for applications to
continue (or even increase) network might experience. Target_run_length is the number
amount of connections that they open
concurrently.

The calculations data required between losses or ECN marks standard for
standard congestion control.

The individual parameters are presented with the for each single property test. In general these calculations permit some
derating as described in Section 5.3. For diagnostic test parameters that can
be derated and is described
below. In a few case there are proportional to target_pipe_size, it not well established models for what
is
recommended that the derating be specified relative to
target_pipe_size calculations using numb_cons=1, although considered correct network operation. In many of these cases the
derating may additionally
problems might either be specified relative partially mitigated by future improvements
to the
target_pipe_size common to other tests. TCP implementations.

5.3. Parameter Derating

Since some aspects of the models are very conservative, the modeling this
framework permits some latitude in derating some specific test parameters. For example classical performance models suggest that in
order to be sure that a single TCP stream can fill a link, it needs
to have a full bandwidth-delay-product worth of buffering at the
bottleneck[QueueSize]. In real networks with real applications this
is often overly conservative. Rather
than trying to formalize more complicated models we permit some test
parameters to be relaxed as long as they meet some additional
procedural constraints:
o The TDS or FSTDS MUST document and justify the actual method used
compute and justify the derated metrics is
published in such a way that it becomes a matter of public record.
@@@ introduce earlier metric parameters.
o The calibration validation procedures described in Section 9 are must be used to
demonstrate the feasibility of meeting the performance targets
with infrastructure that infinitessimally passes the derated test parameters.
tests.
o The calibration validation process itself is must be documented is such a way
that other researchers can duplicate the experiments and validate the
results.

In the test specifications in Section 7 validation experiments.

Except as noted, all tests below assume 0 < derate <= 1, no derating. Tests where
there is not currently a
derating parameter. These will be individually named in the final
document. In all cases making derate smaller makes well established model for the test more
tolerant. Derate = 1 is "full strenght".

Note that some test required
parameters are not permitted include derating as a way to be derated. indicate flexibility in the
parameters.

6. Common testing procedures

6.1. Traffic generating techniques

6.1.1. Paced transmission

Paced (burst) transmissions: send bursts of data on a timer to meet a
particular target rate and pattern.
Single: In all cases the specified data
rate can either be the application or link rates. Header overheads
must be included in the calculations as appropriate.

Paced single packets: Send individual packets at the specified rate
or headway.
Burst: Send sender interface rate bursts on a timer. Specify any 3
of
of: average rate, packet size, burst size (number of packets) and
burst headway (burst start to start). These bursts are typically
sent as back-to-back packets at the testers interface rate.
Slowstart:
Slowstart bursts: Send 4 packet sender interface rate bursts at an
average data rate equal to the minimum of twice effective bottleneck link rate
or
(but not more than the sender interface rate. rate). This corresponds
to the average rate during a TCP slowstart when Appropriate Byte
Counting [ABC] is present or delayed ack is disabled.
Repeated Slowstart: Slowstart pacing itself is bursts: Slowstart bursts are typically part of
larger scale pattern of repeated bursts, such as sending
target_pipe_size packets as slowstart bursts on a target_RTT
headway (burst start to burst start). Such a stream has three
different average rates, depending on the averaging time scale.
At the finest time scale the average rate is the same as the
sender interface rate, at a medium scale the average rate is twice
the effective bottleneck link rate and at the longest time scales
the average rate is the target data rate, adjusted to include header
overhead. rate.

Note that if the effective bottleneck link rate is more than half of
the sender interface rate, slowstart bursts become sender interface
rate bursts.

6.1.2. Constant window pseudo CBR

Implement pseudo CBR constant bit rate by running a standard protocol
such as TCP with a fixed bound on the window size. This has the advantage that it can be
implemented as part of real content delivery. The rate is only
maintained in average over each RTT, and is subject to limitations of
the transport protocol.

For tests that have strongly prescribed data rates, if

The bound on the window size is computed from the target_data_rate
and the actual RTT of the test path.

If the transport protocol fails to maintain the test rate for any reason related to within
prescribed data rates, the test MUST NOT be considered passing. If
there is a signature of a network itself, such as problem (e.g. the run length is too
small) then the test can be considered to fail. Since packet losses or congestion, loss
and ECN marks are required to reduce the data rate for standard
transport protocols, the test
should specification must include suitable
allowances in the prescribed data rates. If there is not sufficient
signature of a network problem, then failing to make the prescribed
data rate must be considered inconclusive. Otherwise there are some
cases where tester failures might cause false negative link test results.

6.1.2.1.

6.1.3. Scanned window pseudo CBR

Same as the above, except the window is incremented once per
2*target_pipe_size, starting from below target_pipe[@@@ test pipe]
and sweeping up to first loss or some other event. This is analogous scanned across a range of
sizes designed to include two key events, the tests implemented in Windowed Ping [WPING] and pathdiag
[Pathdiag]

6.1.3. Intermittent Testing

Any test which does not depend on onset of queueing (e.g. and
the CBR tests) onset of packet loss or
experiences periodic zero outstanding data during normal operation
(e.g. between bursts for burst tests), can be formulated as an
intermittent test. ECN marks. The Intermittent testing can be used for ongoing monitoring window is scanned by
incrementing it by one packet for
changes in subpath quality with minimal disruption users. It should
be used in conjunction with every 2*target_pipe_size delivered
packets. This mimics the full rate test because additive increase phase of standard
congestion avoidance and normally separates the the window increases
by approximately twice the target_RTT.

There are two versions of this method
assesses an average_run_length over test: one built by applying a long time interval w.r.t. user
sessions. It may false fail due window
clamp to other legitimate standard congestion
causing traffic or may false pass changes in underlying link
properties (e.g. control and one one built by stiffening
a modem retraining to an out of contract lower
rate).

[Need text about bias (false pass) non-standard transport protocol. When standard congestion control
is in effect, any losses or ECN marks cause the shadow of loss caused by
excessive bursts]

6.1.4. Intermittent Scatter Testing

Intermittent scatter testing: when testing the network path transport to revert
to or
from an ISP subscriber aggregation point (CMTS, DSLAM, etc),
intermittent tests can be spread across a pool of users window smaller than the clamp such that no
one users experiences the full impact of the testing, even though the
traffic to or from the ISP subscriber aggregation point is sustained
at full rate.

6.2. Interpreting scanning clamp
looses control the Results

6.2.1. Test outcomes

A singleton window size. The NPAD pathdiag tool is an example
of this class of algorithms [Pathdiag].

Alternatively a pass fail measurement. If any subpath fails any
test it non-standard congestion control algorithm can be assumed respond
to losses by transmitting extra data, such that the end-to-end path will also fail it (attempts) to
attain
maintain the target performance specified window size independent of losses or ECN
marks. Such a stiffened transport explicitly violates mandatory
Internet congestion control and is not suitable for in situ testing.
It is only appropriate for engineering testing under some laboratory
conditions.

In addition we use "inconclusive" outcome to indicate that The Windowed Ping tools implemented such a test
failed to attain the required test conditions. [WPING].
This tool has been updated and is important under test.[mpingSource]

The test procedures in Section 7.2 describe how to the extent that the tests themselves use protocols that have built in
control systems which might interfere with some aspect of partition the test.
For example consider a test is implemented by adding rate controls
scans into regions and instrumentation to TCP: failing how to attain interpret the specified data rate
has results.

6.1.4. Concurrent or channelized testing

The procedures described in his document are only directly applicable
to be treated single stream performance measurement, e.g. one TCP connection.
In an inconclusive, unless the test clearly fails
(target_run_lenght is too small). This Ideal world, we would disallow all performance claims based
multiple concurrent stream but this is because failing not practical due to reach
the target at least
two different issues. First, many very high rate is an ambiguous signature for link technologies
are channelized, and pin individual flows to specific channels to
minimize reordering or solve other problems with either and second TCP itself has
scaling limits. Although the test procedure (a former problem with might be overcome
through different design decisions, the TCP implementation or later problem is more deeply
rooted.

All standard [RFC 5681] and de facto standard [CUBIC] congestion
control algorithms have scaling limits, in the test
path sense that as a
network over a fixed RTT is too long) or and MTU gets faster all congestion control
algorithms get less accurate. In general their noise immunity drops
(a single packet drop should have less effect as individual packets
become smaller relative to the subpath itself.

The vantage independence properties of Model Based Metrics depends on window size) and the accuracy control frequency
of the distinction between failing and inconclusive
tests. One of AIMD sawtooth also drops, meaning that as TCP is using more
total capacity it gets less information about the goals state of evolving test designs will be to keep
sharpening the distinction between failing
network and inconclusive tests.

One of the goals other traffic. These properties are a direct consequence
of evolving the testing process, procedures original Reno design and
measurement point selection should are implicitly required by the
requirement that all transport protocols be "TCP friendly"
[Guidelines] There are a number of reason to minimize the want to specify
performance in term of multiple concurrent flows. Although there are
a number of
inconclusive tests.

6.2.2. Statistical criteria for measuring run_length

When evaluating the observed run_length, we need to determine
appropriate packet stream sizes and acceptable error levels downsides to test
efficiently. In practice, can we compare @@@@

The use of multiple connections in the empirically estimated
loss probabilities with Internet has been very
controversial since the targets as beginning of the sample size grows? How
large a sample World-Wide-Web[first
complaint]. Modern browsers open many connections [BScope]. Experts
associated with IETF transport area have frequently spoken against
this practice [long list]. It is needed not inappropriate to say that the measurements assume some
small number of packet
transfer indicate concurrent connections (e.g. 4 or 6), to compensate
for limitation in TCP. However, choosing too large a particular run-length number is present?

The generalized measurement can be described at
risk of being interpreted as recursive testing:

send a flight of packets and observe signal by the packet transfer performance
(loss ratio or other metric, any defect we define).

As each flight is sent and measured, we have an ongoing estimate of web browser community
that this practice has been embraced by the performance in terms of defect to total packet ratio (or an
empirical probability). Continue Internet service provider
community. It may not be desirable to send until conditions support such a
conclusion or signal.

Note that the current proposal for httpbis [SPDY] is specifically
designed to work best with a maximum single TCP connection per client server
pair, because it uses adaptive compression which requires sending limit has been reached.

We have a target_defect_probability, 1 defect
separate compression dictionaries per target_run_length,
where a "defect" is defined connection. As long as TCP can
use IW10 and some of the transport parameter can be cached, multiple
connections provide a lost packet, a packet with ECN mark,
or other impairment. This constitutes the null Hypothesis:

H0: no more than one defects in target_run_length = (3/2)*(flight)^2
packets

and we can stop sending flights of packets if measurements support
accepting H0 with negative gain, due to the specified Type I error = alpha (= 0.05 for
example).

We also have an alternative Hypothesis replicated
compression overhead.

The specification to evaluate: if performance use multiple connections is
significantly lower than the target_defect_probability, say half the
target:

H1: one or more defects in target_run_length/2 packets

and we not recommended for
data rates below several Mb/s, which can stop sending flights of packets if measurements support
rejecting H0 be attained with run lengths
under 10000. Since run length goes as the specified Type II error = beta, thus preferring
the alternate H1.

H0 and H1 constitute square of the Success and Failure outcomes described
elsewhere in data rates,
at higher rates (see Section 8.3) the memo, run lengths can be unfeasibly
large, and while multiple connection might be the ongoing measurements do only feasible approach.

6.1.5. Intermittent Testing

Any test which does not
support either hypothesis depend on queueing (e.g. the current status of measurements is
inconclusive.

The problem above is formulated to match CBR tests) or
experiences periodic zero outstanding data during normal operation
(e.g. between bursts for the Sequential Probability
Ratio Test (SPRT) [StatQC] [temp ref:
http://en.wikipedia.org/wiki/Sequential_probability_ratio_test ],
which also starts with a pair of hypothesis specified various burst tests), can be formulated
as above:

H0: p = p0 = one defect an intermittent test.

The Intermittent testing can be used for ongoing monitoring for
changes in target_run_length
H1: p = p1 = one defect subpath quality with minimal disruption users. It should
be used in target_run_length/2
As flights are sent and measurements collected, the tester evaluates conjunction with the cumulative log-likelihood ratio:

S_i full rate test because this method
assesses an average_run_length over a long time interval w.r.t. user
sessions. It may false fail due to other legitimate congestion
causing traffic or may false pass changes in underlying link
properties (e.g. a modem retraining to an out of contract lower
rate).

[Need text about bias (false pass) in the shadow of loss caused by
excessive bursts]

6.1.6. Intermittent Scatter Testing

Intermittent scatter testing: when testing the network path to or
from an ISP subscriber aggregation point (CMTS, DSLAM, etc),
intermittent tests can be spread across a pool of users such that no
one users experiences the full impact of the testing, even though the
traffic to or from the ISP subscriber aggregation point is sustained
at full rate.

6.2. Interpreting the Results

6.2.1. Test outcomes

A singleton is a pass/fail measurement of a subpath. If any subpath
fails any test then the end-to-end path is also expected to fail to
attain the target performance under some conditions.

In addition we use "inconclusive outcome" to indicate that a test
failed to attain the required test conditions. A test is
inconclusive if the precomputed traffic pattern was not authentically
generated, test preconditions were not met or the measurement results
were not statistically significantly.

This is important to the extent that the diagnostic tests use
protocols which themselves include built in control systems which
might interfere with some aspect of the test. For example consider a
test that is implemented by adding rate controls and loss
instrumentation to TCP: meeting the run length specification while
failing to attain the specified data rate must be treated as an
inconclusive result, because we can not a priori determine if the
reduced data rate was caused by a TCP problem or a network problem,
or if the reduced data rate had a material effect on the run length
measurement. (Note that if the measured run length was too small,
the test can be considered to have failed because it doesn't really
matter that the test didn't attain the required data rate).

The vantage independence properties of Model Based Metrics depends on
the accuracy of the distinction between conclusive (pass or fail) and
inconclusive tests. One way to view inconclusive tests is that they
reflect situations where the signature is ambiguous between problems
with the the subpath and problems with the diagnostic test itself.
One of the goals for evolving diagnostic test designs will be to keep
sharpening this distinction.

One of the goals of evolving the testing process, procedures and
measurement point selection should be to minimize the number of
inconclusive tests.

Note that procedures that attempt to sweep the target parameter space
to find the bounds on some parameter (for example to find the highest
data rate for a subpath) are likely to break the location independent
properties of Model Based Metrics, because the boundary between
passing and inconclusive is extremely likely to be RTT sensitive,
because TCP's ability to compensate for problems scales with the
number of round trips per second.

6.2.2. Statistical criteria for measuring run_length

When evaluating the observed run_length, we need to determine
appropriate packet stream sizes and acceptable error levels for
efficient methods of measurement. In practice, can we compare the
empirically estimated loss probabilities with the targets as the
sample size grows? How large a sample is needed to say that the
measurements of packet transfer indicate a particular run-length is
present?

The generalized measurement can be described as recursive testing:
send packets (individually or in patterns) and observe the packet
transfer performance (loss ratio or other metric, any defect we
define).

As each packet is sent and measured, we have an ongoing estimate of
the performance in terms of defect to total packet ratio (or an
empirical probability). We continue to send until conditions support
a conclusion or a maximum sending limit has been reached.

We have a target_defect_probability, 1 defect per target_run_length,
where a "defect" is defined as a lost packet, a packet with ECN mark,
or other impairment. This constitutes the null Hypothesis:

H0: no more than one defect in target_run_length =
3*(target_pipe_size)^2 packets

and we can stop sending packets if on-going measurements support
accepting H0 with the specified Type I error = alpha (= 0.05 for
example).

We also have an alternative Hypothesis to evaluate: if performance is
significantly lower than the target_defect_probability. Based on
analysis of typical values and practical limits on measurement
duration, we choose four times the H0 probability:

H1: one or more defects in (target_run_length/4) packets

and we can stop sending packets if measurements support rejecting H0
with the specified Type II error = beta (= 0.05 for example), thus
preferring the alternate hypothesis H1.

H0 and H1 constitute the Success and Failure outcomes described
elsewhere in the memo, and while the ongoing measurements do not
support either hypothesis the current status of measurements is
inconclusive.

The problem above is formulated to match the Sequential Probability
Ratio Test (SPRT) [StatQC], which also starts with a pair of
hypothesis specified as above:

H0: p0 = one defect in target_run_length
H1: p1 = one defect in target_run_length/4
As packets are sent and measurements collected, the tester evaluates
the cumulative defect count against two boundaries representing H0
Acceptance or Rejection (and acceptance of H1):

Acceptance line: Xa = -h1 + sn
Rejection line: Xr = S_i-1 h2 + log(Lambda_i) sn
where Lambda_i n increases linearly for each packet sent and

h1 = { log((1-alpha)/beta) }/k
h2 = { log((1-beta)/alpha) }/k
k = log{ (p1(1-p0)) / (p0(1-p1)) }
s = [ log{ (1-p0)/(1-p1) } ]/k
for p0 and p1 as defined in the null and alternative Hypotheses
statements above, and alpha and beta as the Type I and Type II error.

The SPRT specifies simple stopping rules:

o Xa < defect_count(n) < Xb: continue testing
o defect_count(n) <= Xa: Accept H0
o defect_count(n) >= Xb: Accept H1

The calculations above are implemented in the R-tool for Statistical
Analysis, in the add-on package for Cross-Validation via Sequential
Testing (CVST) [http://www.r-project.org/] [Rtool] [CVST] .

Using the equations above, we can calculate the minimum number of
packets (n) needed to accept H0 when x defects are observed. For
example, when x = 0:

Xa = 0 = -h1 + sn
and n = h1 / s

6.2.3. Reordering Tolerance

All tests must be instrumented for reordering [RFC4737].

NB: there is no global consensus for how much reordering tolerance is
appropriate or reasonable. ("None" is absolutely unreasonable.)

Section 5 of [RFC4737] proposed a metric that may be sufficient to
designate isolated reordered packets as effectively lost, because
TCP's retransmission response would be the same.

[As a strawman, we propose the following:] TCP should be able to
adapt to reordering as long as the reordering extent is no more than
the maximum of one half window or 1 mS, whichever is larger. Note
that there is a fundamental tradeoff between tolerance to reordering
and how quickly algorithms such as fast retransmit can repair losses.
Within this limit on reorder extent, there should be no bound on
reordering density.

NB: Traditional TCP implementations were not compatible with this
metric, however newer implementations still need to be evaluated

Parameters:
Reordering displacement: the ratio maximum of one half of target_pipe_size
or 1 mS.

6.3. Test Qualifications

This entire section might be summarized as "needs to be specified in
a FSTDS"

Things to monitor before, during and after a test.

6.3.1. Verify the two likelihood functions
(calculated Traffic Generation Accuracy

[Excess detail for this doc. To be summarized]

for most tests, failing to accurately generate the test traffic
indicates an inconclusive tests, since it has to be presumed that the
error in traffic generation might have affected the test outcome. To
the extent that the network itself had an effect on the measurement at packet i, and index i increases
linearly over all flights the traffic
generation (e.g. in the standing queue tests) the possibility exists
that allowing too large of packets ) for p0 and p1 [temp ref:
http://en.wikipedia.org/wiki/Likelihood_function ]. error margin in the traffic generation
might introduce feedback loops that comprise the vantage independents
properties of these tests.

Parameters:
Maximum Data Rate Error The SPRT specifies simple stopping rules:

o a < S_i < b: continue testing
o S_i <= a: Accept H0
o S_i >= b: Accept H1
where a and b are based on permitted amount that the Type I and II errors, alpha and beta: test traffic
can be different than specified for the current test. This is a ~= Log((beta/(1-alpha)) and b ~= Log((1-beta)/alpha)
symmetrical bound.
Maximum Data Rate Overage The permitted amount that the test traffic
can be above than specified for the current test.
Maximum Data Rate Underage The permitted amount that the test
traffic can be less than specified for the current test.

6.3.2. Verify the absence of cross traffic

[Excess detail for this doc. To be summarized]

The proper treatment of cross traffic is different for different
subpaths. In general when testing infrastructure which is associated
with only one subscriber, the error probabilities decided beforehand, test should be treated as above.

The calculations above are implemented in inconclusive
it that subscriber is active on the R-tool network. However, for Statistical
Analysis, in shared
infrastructure, the add-on package for Cross-Validation via Sequential
Testing (CVST) [http://www.r-project.org/] [Rtool] [CVST] .

6.2.3. Classifications of tests

Tests are annotated with "(capacity)", "(engineering)" or
"(monitoring)". @@@@MOVE question at hand is likely to definitions?

Capacity tests determine be testing if a network subpath
provider has sufficient capacity
to deliver total capacity. In such cases the target performance. As such, they reflect parameters
that can transition from passing presence
of cross traffic due to failing as a consequence other subscribers is explicitly part of
additional presented load or the actions of other network users. By
definition, capacity tests also consume
network resources (capacity
and/or buffer space), conditions and their test schedules must be balanced by
their cost.

Monitoring tests its effects are design to capture the most important aspects of
a capacity test, but without causing unreasonable ongoing load
themselves. As such they may miss some details explicitly part of the network
performance, but can serve as a useful reduced cost proxy for a
capacity test.

Engineering tests evaluate how network algorithms (such as AQM

@@@@ Need to distinguish between ISP managed sharing and
channel allocation) interact with transport protocols. These unmanaged
sharing. e.g. WiFi

Note that canceling tests
are likely due to have complicated interactions with load on subscriber lines may
introduce sampling errors for testing other network
traffic and can be inversely sensitive parts of the
infrastructure. For this reason tests that are scheduled but not run
due to load. For example load should be treated as a test special case of "inconclusive".

Use a passive packet or SNMP monitoring to verify that an AQM algorithm causes ECN marks or packet drops
early enough to limit queue occupancy may experience a false pass
results in the presence of bursty cross traffic. It is important
that engineering tests traffic
volume on the subpath agrees with the traffic generated by a test.
Ideally this should be performed under a wide range of conditions,
including both in situ before, during and bench testing, after each test.

The goal is provide quality assurance on the overall measurement
process, and under specifically to detect the following measurement
failure: a variety of load
conditions. Ongoing monitoring user observes unexpectedly poor application performance,
the ISP observes that the access link is less likely running at the rated
capacity. Both fail to observe that the user's computer has been
infected by a virus which is spewing traffic as fast as it can.

Parameters:
Maximum Cross Traffic Data Rate The amount of excess traffic
permitted. Note that this will be useful for these
tests, although sparse in situ testing might be appropriate.

@@@ Add single property vs combined tests here?

6.2.4. Reordering Tolerance

All tests must be instrumented different for reordering [RFC4737].

NB: there different tests.

One possible method is no global consensus an adaptation of: www-didc.lbl.gov/papers/
SCNM-PAM03.pdf D Agarwal etal. "An Infrastructure for how much reordering tolerance is
appropriate or reasonable. ("None" is absolutely unreasonable.)

Section 5 Passive
Network Monitoring of [RFC4737] proposed a metric that may be sufficient to
designate isolated reordered packets Application Data Streams". Use the same
technique as effectively lost, because
TCP's retransmission response would be that paper to trigger the same.

[As a strawman, we propose capture of SNMP statistics for
the following:] TCP should link.

6.3.3. Additional test preconditions

[Excess detail for this doc. To be able summarized]

Send pre-load traffic as needed to
adapt activate radios with a sleep mode,
or other "reactive network" elements (term defined in
[draft-morton-ippm-2330-update-01]).

Use the procedure above to reordering as long as confirm that the reordering extent pre-test background
traffic is no more than low enough.

7. Diagnostic Tests

The diagnostic tests are organized by which properties are being
tested: run length, standing queues; slowstart bursts; sender rate
bursts; and combined tests. The combined tests reduce overhead at
the maximum expense of one half window conflating the signatures of multiple failures.

7.1. Basic Data Rate and Run Length Tests

We propose several versions of the basic data rate and run length
test. All measure the number of packets delivered between losses or 1 mS, whichever is larger. Note
ECN marks, using a data stream that there is a fundamental tradeoff between tolerance to reordering
and rate controlled at or below
the target_data_rate.

The tests below differ in how quickly algorithms such as fast retransmit the data rate is controlled. The data
can repair losses.
Within this limit on reorder extent, there should be no bound paced on
reordering frequency.

NB: Current TCP implementations a timer, or window controlled at full target data
rate. The first two tests implicitly confirm that sub_path has
sufficient raw capacity to carry the target_data_rate. They are not compatible with this metric.
We view this
recommend for relatively infrequent testing, such as bugs in current TCP implementations.

Parameters:
Reordering displacement: the maximum of one half of target_pipe_size an installation
or 1 mS.

6.3. Test Qualifications

Things to monitor before, during and after auditing process. The third, background run length, is a test.

6.3.1. Verify the Traffic Generation Accuracy low rate
test designed for most tests, failing ongoing monitoring for changes in subpath quality.

All rely on the receiver accumulating packet delivery statistics as
described in Section 6.2.2 to accurately generate score the outcome:

Pass: it is statistically significant that the observed run length is
larger than the target_run_length.

Fail: it is statistically significant that the observed run length is
smaller than the target_run_length.

A test traffic
indicates an is considered to be inconclusive tests, since if it has failed to be presumed that meet the
error data
rate as specified below, meet the qualifications defined in
Section 6.3 or neither run length statistical hypothesis was
confirmed in traffic generation might have affected the allotted test outcome. To
the extent duration.

7.1.1. Run Length at Paced Full Data Rate

Confirm that the network itself had an effect observed run length is at least the
target_run_length while relying on timer to send data at the
target_rate using the traffic
generation (e.g. procedure described in the standing queue tests) the possibility exists
that allowing too large of error margin in Section 6.1.1 with a
burst size of 1 (single packets).

The test is considered to be inconclusive if the traffic generation
might introduce feedback loops packet transmission
can not be accurately controlled for any reason.

7.1.2. run length at Full Data Windowed Rate

Confirm that comprise the vantage independents observed run length is at least the
target_run_length while sending at an average rate equal to the
target_data_rate, by controlling (or clamping) the window size of a
conventional transport protocol to a fixed value computed from the
properties of these tests.

Parameters:

Maximum Data Rate Error The permitted amount that the test traffic
can path, typically
test_window=target_data_rate*test_RTT/target_MTU.

Since losses and ECN marks generally cause transport protocols to at
least temporarily reduce their data rates, this test is expected to
be different than specified for less precise about controlling its data rate. It should not be
considered inconclusive as long as at least some of the current test. This is a
symmetrical bound.
Maximum Data Rate Overage The permitted amount that round trips
reached the full target_data_rate, without incurring losses. To pass
this test traffic
can be above than specified for the current network MUST deliver target_pipe_size packets in
target_RTT time without any losses or ECN marks at least once per two
target_pipe_size round trips, in addition to meeting the run length
statistical test.
Maximum Data Rate Underage

7.1.3. Background Run Length Tests

The permitted amount that background run length is a low rate version of the target target
rate test
traffic can be less than specified above, designed for the current test.

6.3.2. Verify the absence of cross traffic

The proper treatment of cross traffic is different ongoing lightweight monitoring for different
subpaths. In general when testing infrastructure which is associated
with only one subscriber,
changes in the test observed subpath run length without disrupting users.
It should be treated as inconclusive used in conjunction with one of the above full rate
tests because it does not confirm that subscriber is active on the network. However, for shared
infrastructure, the question at hand is likely to be testing if
provider has sufficient total capacity. In subpath can support raw
data rate.

Existing loss metrics such cases as [RFC 6673] might be appropriate for
measuring background run length.

7.2. Standing Queue tests

These test confirm that the presence
of cross traffic due to other subscribers bottleneck is explicitly part well behaved across the
onset of packet loss, which typically follows after the
network conditions and its effects are explicitly part onset of
queueing. Well behaved generally means lossless for transient
queues, but once the test.

Note that canceling tests due to load on subscriber lines may
introduce sampling errors queue has been sustained for testing other parts a sufficient period
of the
infrastructure. For this reason tests that are scheduled but not run
due to load time (or a sufficient queue depth) there should be treated as a special case small number
of "inconclusive".

Use a passive packet or SNMP monitoring losses to verify that the traffic
volume on the subpath agrees with the traffic generated by a test.
Ideally this should be performed before during and after each test.

The goal is provide quality assurance on the overall measurement
process, and specifically signal to detect the following measurement
failure: a user observes unexpectedly poor application performance, the ISP observes transport protocol that it should reduce
its window. Losses that are too early can prevent the access link is running transport from
averaging at the rated
capacity. Both fail to observe target_data_rate. Losses that the user's computer has been
infected by a virus which is spewing traffic as fast as it can.

Parameters:
Maximum Cross Traffic Data Rate The amount of excess traffic
permitted. Note are too late indicate
that this will the queue might be different for different tests.

One possible method is an adaptation of: www-didc.lbl.gov/papers/
SCNM-PAM03.pdf D Agarwal etal. "An Infrastructure subject to bufferbloat [Bufferbloat] and
inflict excess queuing delays on all flows sharing the bottleneck.
Excess losses make loss recovery problematic for Passive
Network Monitoring the transport
protocol. Non-linear or erratic RTT fluctuations suggest poor
interactions between the channel acquisition systems and the
transport self clock. All of Application Data Streams". Use the tests in this section use the same
technique as that paper to trigger
basic scanning algorithm but score the capture link on the basis of SNMP statistics for how well
it avoids each of these problems.

For some technologies the link.

6.3.3. Additional test preconditions

Send pre-load traffic as needed data might not be subject to activate radios with a sleep mode,
or other "reactive network" elements (term defined increasing
delays, in
[draft-morton-ippm-2330-update-01]).

Use which case the procedure above data rate will vary with the window size
all the way up to confirm that the pre-test background
traffic is low enough.

7. Single Property Tests

7.1. Basic Data and Loss Rate Tests

We propose several versions onset of the loss rate test. All are rate
controlled at losses or below ECN marks. For theses
technologies, the target_data_rate. The first, performed at
constant full data rate, is intrusive and recommend for infrequent
testing, such as when a service discussion of queueing does not apply, but it is first turned up or as part
still required that the onset of losses (or ECN marks) be at an
auditing process.
appropriate point and progressive.

Use the procedure in Section 6.1.3 to sweep the window across the
onset of queueing and the onset of loss. The second, background loss rate, is designed for
ongoing monitoring for change is subpath quality.

7.1.1. Loss Rate at Paced Full Data Rate

Confirm tests below all assume
that the observed run length is at least scan emulates standard additive increase and delayed ACK by
incrementing the
target_run_lenght while sending window by one packet for every 2*target_pipe_size
packets delivered. A scan can be divided into three regions: below
the onset of queueing, a standing queue, and at or beyond the target_rate. This test
implicitly confirms that sub_path has sufficient raw capacity to
carry onset
of loss.

Below the target_data_rate. This version onset of queueing the loss RTT is typically fairly constant, and
the data rate test
relies on timers varies in proportion to schedule the window size. Once the data transmission at a true constant bit
rate (CBR).

Test Parameters:
Run Length Same as target_run_lenght
Data Rate Same as target_data_rate
Maximum Cross Traffic A specified small fraction of
target_data_rate.

Note that target_run_lenght reaches the link rate, the data rate becomes fairly constant,
and target_data_rate parameters MUST NOT
be derated. If the default parameters are too stringent an alternate
model as described RTT increases in Appendix A proportion to the the window size. The
precise transition from one region to the other can be used identified by
the maximum network power, defined to compute
target_run_lenght.

The test traffic is sent using be the procedures in Section 6.1.1 ratio data rate over the
RTT[POWER].

For technologies that do not have conventional queues, start the scan
at
target_data_rate with a burst size of 1, subject to the
qualifications in Section 6.3. The receiver accumulates packet
delivery statistics as described in Section 6.2 window equal to score the outcome:

Pass: it is statistically significantly that test_window, i.e. starting at the observed run length
is larger than target
rate, instead of the target_run_length.

Fail: it power point.

If there is statistically significantly that random background loss (e.g. bit errors, etc), precise
determination of the observed run length
is smaller than onset of packet loss may require multiple scans.
Above the target_run_length.

Inconclusive: The test failed onset of loss, all transport protocols are expected to meet
experience periodic losses. For the stiffened transport case they
will be determined by the qualifications defined AQM algorithm in
Section 6.3 the network or neither test was statistically significant.

7.1.2. Loss Rate at Full Data Windowed Rate

Confirm that the observed run length is at least details
of how the
target_run_lenght while sending at the target_rate. This test
implicitly confirms that sub_path has sufficient raw capacity window increase function responds to
carry loss. For the target_data_rate. This version
standard transport case the details of periodic losses are typically
dominated by the loss rate test
relies on a fixed window to self clock data transmission into behavior of the
network. This is more authentic.

Test Parameters:
Run Length Same as target_run_lenght
Data Rate Same as target_data_rate
Maximum Cross Traffic transport protocol itself.

7.2.1. Congestion Avoidance

A specified small fraction of
target_data_rate.

Note that target_run_lenght and target_data_rate parameters MUST NOT
be derated. If link passes the default parameters congestion avoidance standing queue test if more
than target_run_length packets are too stringent an alternate
model as described in Appendix A can be used to compute
target_run_lenght.

The delivered between the power point
(or test_window) and the first loss or ECN mark. If this test traffic is sent
implemented using the procedures in Section 6.1.1 at
target_data_rate a standards congestion control algorithm with a burst size
clamp, it can be used in situ in the production internet as a
capacity test. For an example of 1, subject such a test see [NPAD].

7.2.2. Bufferbloat

This test confirms that there is some mechanism to limit buffer
occupancy (e.g. prevents bufferbloat). Note that this is not
strictly a requirement for single stream bulk performance, however if
there is no mechanism to limit buffer occupancy then a single stream
with sufficient data to deliver is likely to cause the
qualifications in Section 6.3. The receiver accumulates packet
delivery statistics as problems
described in Section 6.2 [RFC 2309] and [Bufferbloat]. This may cause only minor
symptoms for the dominant flow, but has the potential to score make the outcome:

Pass: it is statistically significantly that
link unusable for all other flows and applications.

Pass if the observed run length onset of loss is larger before a standing queue has introduced
more delay than the target_run_length.

Fail: it than twice target_RTT, or other well defined limit.
Note that there is statistically significantly not yet a model for how much standing queue is
acceptable. The factor of two chosen here reflects a rule of thumb.
Note that in conjunction with the observed run length previous test, this test implies
that the first loss should occur at a queueing delay which is smaller than between
one and two times the target_run_length.

Inconclusive: The target_RTT.

7.2.3. Non excessive loss

This test failed to meet confirm that the qualifications defined in
Section 6.3 or neither test was statistically significant.

7.1.3. Background Loss Rate Tests

The background onset of loss rate is a low rate version of not excessive. Pass if
losses are bound by the target rate
test above, designed for ongoing monitoring for changes in subpath
quality without disrupting users. It should be used the fluctuations in conjunction
with the above full rate test because it may be subject to false
results under some conditions, cross traffic, such
that transient load (bursts) do not cause dips in particular it may false aggregate raw
throughput. e.g. pass
changes in underlying link properties (e.g. a modem retraining to an
out of contract lower rate).

Parameters:
Run Length Same as target_run_lenght
Data Rate Some small fraction of target_data_rate, such long as 1%.

Once the preconditions described in Section 6.3 losses are met, the no more bursty than
are expected from a simple drop tail queue. Although this test could
be made more precise it is really included here for pedantic
completeness.

7.2.4. Duplex Self Interference

This engineering test
data is sent at the prescribed rate with confirms a burst size of 1. The
receiver accumulates packet delivery statistics and bound on the procedures
described in Section 6.2.1 interactions between
the forward data path and Section 6.3 are used to score the
outcome:

Pass: it is statistically significantly that ACK return path. Fail if the observed run length
is larger RTT rises
by more than some fixed bound above the target_run_length.

Fail: it is statistically significantly that the observed run length
is smaller than expected queueing time
computed from trom the target_run_length.

Inconclusive: Neither test was statistically significant or there was excess cross traffic during window divided by the test.

7.2. Standing Queue link data rate.
@@@@ This needs further testing.

7.3. Slowstart tests

These test confirm that the bottleneck tests mimic slowstart: data is well behaved across the
onset of queueing. For conventional bottlenecks this will be from
the onset of queuing to sent at twice the point where there is a full target_pipe
of standing data. Well behaved generally means lossless for
target_run_length, followed by a small number of losses to signal effective
bottleneck rate to exercise the transport protocol that it should slow down. Losses that are too
early can prevent the transport from averaging above queue at the target_rate.
Losses that dominant bottleneck.

They are too late indicate that deemed inconclusive if the queue might be subject to
bufferbloat and subject other flows elapsed time to excess queuing delay. Excess
losses (more send the data
burst is not less than half of of target_pipe) make loss recovery
problematic for the transport protcol.

These tests can also observe some problems with channel acquisition
systems, especially at time to receive the ACKs. (i.e.
sending data too fast is ok, but sending it slower than twice the onset of persistent queueing. Details
TBD.

7.2.1. Congestion Avoidance

Use
actual bottleneck rate as indicated by the procedure in Section 6.1.2.1 ACKs is deemed
inconclusive). Space the bursts such that the average data rate is
equal to sweep the window (rate) from
below link_pipe up target_data_rate.

7.3.1. Full Window slowstart test

This is a capacity test to beyond target_pipe+link_pipe. Depending on
events confirm that happen during the scan, slowstart is not likely to
exit prematurely. Send slowstart bursts that are target_pipe_size
total packets. Accumulate packet delivery statistics as described in
Section 6.2.2 to score the link. Identify the
power_point=MAX(rate/RTT) as outcome. Pass if it is statistically
significant that the start of observed run length is larger than the test.
target_run_length. Fail if first loss it is too early (loss rate too high) on repeated
tests or if statistically significant that the losses are more
observed run length is smaller than half of the outstanding data. (a
capacity test)

7.2.2. Buffer Bloat

Use the procedure in Section 6.1.2.1 to sweep the window (rate) from
below link_pipe up to beyond target_pipe+link_pipe. Depending on
events target_run_length.

Note that happen during the scan, score the link. Identify these are the
"power point:MAX(rate/RTT) same parameters as the start of Sender Full Window
burst test, except the test (should be
window=target_pipe)

Fail if first loss burst rate is too late (insufficient at slowestart rate, rather than
sender interface rate.

7.3.2. Slowstart AQM and subject to
bufferbloat - an engineering test). NO THEORY

7.2.3. Duplex Self Interference

Use test

Do a continuous slowstart (send data continuously at slowstart_rate),
until the procedure in Section 6.1.2.1 to sweep first loss, stop, allow the window (rate) from
below link_pipe up network to beyond target_pipe+required_queue. Depending drain and repeat,
gathering statistics on events that happen during the scan, score last packet delivered before the link. Identify loss,
the
"power point:MAX(rate/RTT) as loss pattern, maximum RTT and window size. Justify the start results.
There is not currently sufficient theory justifying requiring any
particular result, however design decisions that affect the outcome
of this tests also affect how the test (should be
window=target_pipe) @@@ add required_queue network balances between long and power_point

Fail if RTT
short flows (the "mice and elephants" problem)

This is non-monotonic by more than an engineering test: It would be best performed on a small number of packet
times (channel allocation self interference - engineering) IS THIS
SUFFICIENT?

7.3. Slowstart
quiescent network or testbed, since cross traffic has the potential
to change the results.

7.4. Sender Rate Burst tests

These tests mimic slowstart: data is determine how well the network can deliver bursts sent at slowstart_rate (twice
subpath_rate). They are deemed inconclusive if
sender's interface rate. Note that this test most heavily exercises
the elapsed time front path, and is likely to
send the data burst include infrastructure nominally out
of scope.

Also, there are a several details that are not precisely defined.
For starters there is not less than half of the (extrapolated) time a standard server interface rate. 1 Gb/s is
very common today, but higher rates (e.g. 10 Gb/s) are becoming cost
effective and can be expected to receive be dominant some time in the ACKs. (i.e. sending data too fast future.

Current standards permit TCP to send a full window bursts following
an application pause. Congestion Window Validation [RFC 2861], is ok,
not required, but sending even if was it slower does not take effect until an
application pause is longer than twice an RTO. Since this is standard
behavior, it is desirable that the actual bottleneck rate network be able to deliver it,
otherwise application pauses will cause unwarranted losses.

It is deemed
inconclusive). Space also understood in the application and serving community that
interface rate bursts such have a cost to the network that has to be
balanced against other costs in the average ACK rate servers themselves. For example
TCP Segmentation Offload [TSO] reduces server CPU in exchange for
larger network bursts, which increase the stress on network buffer
memory.

There is
equal not yet theory to unify these costs or to provide a
framework for trying to or faster than the target_data_rate.

These tests are optimize global efficiency. We do not useful at burst sizes smaller than yet
have a model for how much the sender
interface network should tolerate server rate tests, since
bursts. Some bursts must be tolerated by the sender interface rate tests are more
strenuous. If network, but it is necessary
probably unreasonable to derate the sender interface rate
tests, then the full window slowstart test (un-derated) would be
important.

7.3.1. Full Window slowstart test

Send (target_pipe_size+required_queue)*derate bursts must have fewer
than one loss per target_run_length*derate. Note that these are expect the
same parameters network to efficiently deliver
all data as the Sender Full Window burst test, except the
burst rate is at slowestart rate, rather than sender interface rate.
SHOULD derate=1.

Otherwise TCP will exit from slowstart prematurely, and only reach a
full target_pipe_size window by way series of congestion avoidance.

This bursts.

For this reason, this is a capacity test: cross traffic may cause premature losses.

7.3.2. Slowstart AQM the only test

Do for which we explicitly
encourage detrateing. A TDS should include a continuous slowstart (date rate = slowstart_rate), until first
loss, and repeat, gathering statistics on table of pairs of
derating parameters: what burst size to use as a fraction of the last delivered packet's
RTT
target_pipe_size, and window size. Fail if too large (NO THEORY for value).

This how much each burst size is an engineering test: It would be best performed on a
quiescent network or testbed, since cross traffic might cause a false
pass.

7.4. Sender Rate Burst tests permitted to reduce
the run length, relative to to the target_run_length. @@@@ Needs more
work and experimentation.

7.5. Combined Tests

These tests us "sender are more efficient from a deployment/operational
perspective, but may not be possible to diagnose if they fail.

7.5.1. Sustained burst test

Send target_pipe_size*derate sender interface rate" bursts. Although this rate bursts every
target_RTT*derate, for derate between 0 and 1. Verify that the
observed run length meets target_run_length. Key observations:
o This test is not
well defined it should be assumed subpath RTT invariant, as long as the tester can
generate the required pattern.
o The subpath under test is expected to be current state go idle for some fraction of
the art
server grade hardware (often 10Gb/s today). (load)

7.4.1. Sender TCP Send Offload (TSO) tests

If MIN(target_pipe_size, 42) packet bursts meet target_run_lenght
(Not derated!).

Otherwise time: (subpath_data_rate-target_rate)/subpath_data_rate.
Failing to do so suggests a problem with the procedure.

o This test is more strenuous than the slowstart tests: they are not
needed if the link will interact badly passes this test with modern server NIC
implementations, which as an optimization derate=1.
o A link that passes this test is likely to reduce host side
interactions (interrupts etc) accept up be able to 64kB super packets and
send them as 42 seperate packets on sustain
higher rates (close to subpath_data_rate) for paths with RTTs
smaller than the wire side.cc (load)

7.4.2. Sender Full Window burst target_RTT. Offsetting this performance
underestimation is part of the rationale behind permitting
derating in general.
o This test

target_pipe_size*derate bursts have fewer than one loss per
target_run_length*derate.

Otherwise can be implemented with standard instrumented TCP[RFC
4898], using a specialized measurement application pauses will cause unwarranted losses. Current
standards permit at one end and
a minimal service at the other end [RFC 863, RFC 864]. It may
require tweaks to the TCP implementation.
o This test is efficient to send a full cwnd burst following an
application pause. (Cwnd validation in not required, but even so implement, since it does not take effect until require
per-packet timers, and can make use of TSO in modern NIC hardware.
o This test is not totally sufficient: the pause is longer than RTO).

NB: there is no model here for what is good enough. derate=1 standing window
engineering tests are also needed to be sure that the link is
safest, but may well
behaved at and beyond the onset of congestion.
o This one test can be unnecessarily conservative for some applications.
Some application, proven to be the one capacity test to
supplant them all.

7.5.2. Live Streaming Media

Model Based Metrics can be implemented as a side effect of serving
any non-throughput maximizing traffic, such as streaming video need derate=1 media, by
applying some additional controls to be
efficient when the application pacing quanta traffic. The essential
requirement is larger than cwnd.
(load)

8. Combined Tests

These tests are more efficient from a deployment/operational
perspective, but may not that the traffic be possible to diagnose if they fail.

8.1. Sustained burst test

Send target_pipe_size sender interface rate constrained such that even with
arbitrary application pauses, bursts every and data rate fluctuations the
traffic stays within the envelope determined by all of the individual
tests described above, for a specific TDS.

If the serving RTT is less than the target_RTT,
verify that this constraint is
most easily implemented by clamping the observed run length meets target_run_length. Key
observations:
o transport window size to
test_window=target_data_rate*serving_RTT/target_MTU. This test is RTT invariant, as long as
test_window size will limit the tester can generate both the
required pattern.
o The subpath under test is expected serving data rate and burst
sizes to go idle for some fraction be no larger than the procedures in Section 7.1.2 and
Section 7.4, assuming burst size derating equal to the serving_RTT
divided by the target_RTT.

Note that if the application tolerates fluctuations in its actual
data rate (say by use of a playout buffer) it is important that the time: (link_rate-target_rate)/link_rate. Failing to do
target_data_rate be above the actual average rate needed by the
application so
suggests a problem with it can recover after transient pauses caused by
congestion or the procedure.
o This test application itself. Since the serving RTT is more strenuous
smaller than the slowstart tests: they are not
needed if target_RTT, the link passes underated sender interface rate burst
tests.
o This test could worst case bursts that might be derated
generated under these conditions are smaller than called for by reducing both the burst size and
headway (same average data rate).
o A link that passes
Section 7.4

8. Examples

In this test section we present TDS for a couple of performance
specifications.

Tentatively: 5 Mb/s*50 ms, 1 Mb/s*50ms, 250kbp*100mS

8.1. Near serving HD streaming video

Today the best quality HD video requires slightly less than 5 Mb/s
[HDvideo]. Since it is likely desirable to serve such content locally, we
assume that the content will be able within 50 mS, which is enough to sustain
higher rates (close
cover continental Europe or either US coast.

5 Mb/s over a 50 ms path

+----------------------+-------+---------+
| End to link_rate) for paths with RTTs smaller than
the target_RTT. Offsetting this performance underestimation is End Parameter | Value | units |
+----------------------+-------+---------+
| target_rate | 5 | Mb/s |
| target_RTT | 50 | ms |
| traget_MTU | 1500 | bytes |
| target_pipe_size | 22 | packets |
| target_run_length | 1452 | packets |
+----------------------+-------+---------+

Table 1

This example uses the rationale behind permitting derating most conservative TCP model and no derating.

8.2. Far serving SD streaming video

Standard Quality video typically fits in general.
o 1 Mb/s [SDvideo]. This test should can
be implementable reasonably delivered via longer paths with standard instrumented TCP,
[RFC 4898] using a specialized measurement application at one end
and larger. We assume
100mS.

5 Mb/s over a minimal service at the other end [RFC 863, RFC 864]. It may
require tweaks 50 ms path

+----------------------+-------+---------+
| End to End Parameter | Value | units |
+----------------------+-------+---------+
| target_rate | 1 | Mb/s |
| target_RTT | 100 | ms |
| traget_MTU | 1500 | bytes |
| target_pipe_size | 9 | packets |
| target_run_length | 243 | packets |
+----------------------+-------+---------+

Table 2

This example uses the most conservative TCP implementation.
o This test is efficient to implement, since it does not require
per-packet timers, model and can make maximal use no derating.

8.3. Bulk delivery of TSO in modern NIC
hardware.
o remote scientific data

This test is not totally sufficient: the standing window
engineering tests are also needed example corresponds to be sure 100 Mb/s bulk scientific data over a
moderately long RTT. Note that the link target_run_length is well
behaved at infeasible
for most networks.

100 Mb/s over a 200 ms path

+----------------------+---------+---------+
| End to End Parameter | Value | units |
+----------------------+---------+---------+
| target_rate | 100 | Mb/s |
| target_RTT | 200 | ms |
| traget_MTU | 1500 | bytes |
| target_pipe_size | 1741 | packets |
| target_run_length | 9093243 | packets |
+----------------------+---------+---------+

Table 3

9. Validation

This document permits alternate models and beyond parameter derating, as
described in Section 5.2 and Section 5.3. In exchange for this
latitude in the onset of congestion.
o I believe modelling process it requires the ability to
demonstrate authentic applications and protocol implementations
meeting the target end-to-end performance goals over infrastructure
that this infinitessimally passes the TDS.

The validation process relies on constructing a test network such
that all of the individual load tests pass only infinitessimally, and
proving that an authentic application running over a real TCP
implementation (or other protocol as appropriate) can be proven expected to be
meet the one capacity test
to supplant them all.

Example

To confirm that end-to-end target parameters on such a 100 Mb/s link can reliably deliver single 10
MByte/s stream at network.

For example using our example in our HD streaming video TDS described
in Section 8.1, the bottleneck data rate should be 5 Mb/s, the per
packet random background loss probability should be 1/1453, for a distance run
length of 50 mS, test 1452 packets, the bottleneck queue should be 22 packets and
the front path should have just enough buffering to withstand 22
packet line rate bursts. We want every one of the TDS tests to fail
if we slightly increase the link by relevant test parameter, so for example
sending 346 a 23 packet slowstart bursts every should cause excess (possibly
deterministic) packet drops at the dominant queue at the bottleneck.
On this infinitessimally passing network it should be possible for a
real ral application using a stock TCP implementation in the vendor's
default configuration to attain 5 Mb/s over an 50 mS (10 MByte/s payload rate, assuming a 1500
Byte IP MTU and 52 Byte TCP/IP headers). These bursts are 4196288
bits on path.

@@@@ Need to better specify the wire (assuming 16 bytes of link overhead workload: both short and framing) for
an aggregate test data rate long flows.

The difficult part of 8.4 Mb/s.

To this process is arranging for each subpath to
infinitesimally pass the test using the most conservative TCP model for individual tests. We suggest two
approaches: constraining resources in devices by configuring them not
to use all available buffer space or data rate; and preloading
subpaths with cross traffic. Note that is it important that a single
stream the observed run length must be larger than 179574 packets.

This
environment is the same as less than one loss per 519 bursts (1.5*346) or
every 26 seconds.

Note constructed that this test potentially cause transient 346 packet queues at
the bottleneck.

9. Calibration infinitessimally passes all tests,
otherwise there is a chance that TCP can exploit extra latitude in
some parameters (such as data rate) to partially compensate for
constraints in other parameters.

If using derated metrics, or when something goes wrong, the results
must be calibrated against a traditional BTC. The preferred
diagnostic follow-up TDS validated according to calibration issues these procedures is used to run open end-to-end
measurements on an inform
public dialog, the validation experiment itself should also be public
with sufficient precision for the experiment to be replicated by
other researchers. All components should either be open platform, such as Measurement Lab
[http://www.measurementlab.net/] source of
fully specified proprietary implementations that are available to the
research community.

TODO: paper proving the validation process.

10. Acknowledgements

Ganga Maguluri suggested the statistical test for measuring loss
probability in the target run length.

Meredith Whittaker for improving the clarity of the communications.

11. Informative References

[RFC2330] Paxson, V., Almes, G., Mahdavi, J., and M. Mathis,
"Framework for IP Performance Metrics", RFC 2330,
May 1998.

[RFC4737] Morton, A., Ciavattone, L., Ramachandran, G., Shalunov,
S., and J. Perser, "Packet Reordering Metrics", RFC 4737,
November 2006.

[RFC5681] Allman, M., Paxson, V., and E. Blanton, "TCP Congestion
Control", RFC 5681, September 2009.

[RFC5835] Morton, A. and S. Van den Berghe, "Framework for Metric
Composition", RFC 5835, April 2010.

[RFC6049] Morton, A. and E. Stephan, "Spatial Composition of
Metrics", RFC 6049, January 2011.

[I-D.morton-ippm-lmap-path]
Bagnulo, M., Burbridge, T., Crawford, S., Eardley, P., and
A. Morton, "A Reference Path and Measurement Points for
LMAP", draft-morton-ippm-lmap-path-00 (work in progress),
January 2013.

[MSMO97] Mathis, M., Semke, J., Mahdavi, J., and T. Ott, "The
Macroscopic Behavior of the TCP Congestion Avoidance
Algorithm", Computer Communications Review volume 27,
number3, July 1997.

[WPING] Mathis, M., "Windowed Ping: An IP Level Performance
Diagnostic", INET 94, June 1994.

[mpingSource]
Fan, X., Mathis, M., and D. Hamon, "Git Repository for
mping: An IP Level Performance Diagnostic", Sept 2013,
<https://github.com/m-lab/mping>.

[MBMSource]
Hamon, D., "Git Repository for Model Based Metrics",
Sept 2013, <https://github.com/m-lab/MBM>.

[Pathdiag]
Mathis, M., Heffner, J., O'Neil, P., and P. Siemsen,
"Pathdiag: Automated TCP Diagnosis", Passive and Active
Measurement , June 2008.

[BScope] Broswerscope, "Browserscope Network tests", Sept 2012,
<http://www.browserscope.org/?category=network>.

[Rtool] R Development Core Team, "R: A language and environment
for statistical computing. R Foundation for Statistical
Computing, Vienna, Austria. ISBN 3-900051-07-0, URL
http://www.R-project.org/", , 2011.

[StatQC] Montgomery, D., "Introduction to Statistical Quality
Control - 2nd ed.", ISBN 0-471-51988-X, 1990.

[CVST] Krueger, T. and M. Braun, "R package: Fast Cross-
Validation via Sequential Testing", version 0.1, 11 2012.

[LMCUBIC] Ledesma Goyzueta, R. and Y. Chen, "A Deterministic Loss
Model Based Analysis of CUBIC, IEEE International
Conference on Computing, Networking and Communications
(ICNC), E-ISBN : 978-1-4673-5286-4", January 2013.

Appendix A. Model Derivations

This appendix describes several different ways to calculate

The reference target_run_length and the implication of the chosen calculation.

Rederive MSMO97 under two different assumptions: target_rate =
link_rate and target_rate < 2 * link_rate.

Show equivalent derivation for CUBIC.

Commentary on the consequence of the choice.

Appendix B. old text

This entire section is contains scraps of text to be moved, removed
or absorbed elsewhere described in the document

B.1. An earlier document

Step 0: select target end-to-end parameters: a target rate and target
RTT. The primary test will be to confirm that the link quality Section 5.2 is
sufficient to meet the specified target rate for the link under test,
when extended to the target RTT by an ideal network. The target rate
must be below the actual link rate and nominally the target RTT would
be longer than the link RTT. There should probably be a convention
for the relationship between link and target rates (e.g. 85%).

For example based on a 10 Mb/s link, the target rate might be 1 MBytes/s,
at an RTT of 100 mS (a typical continental scale path).

Step 1: On the basis of the target rate and RTT and your favorite TCP
performance model, compute the "required run length", which is the
required number of consecutive non-losses between loss episodes. The
run length resembles one over the loss probability, if clustered
losses only count as a single event. Also select "test duration" and
"test rate". The latter would nominally the same as the target rate,
but might be different in some situations. There must be
documentation connecting the test rate, duration and required run
length,
very conservative assumptions: that all window above target_pipe_size
contributes to a standing queue that raises the target rate RTT, and RTT selected in step 0.

Continuing the above example: Assuming a 1500 Byte MTU. The
calculated model loss rate for a single TCP stream that classic
Reno congestion control is about 0.01% (1
loss in 1E4 packets).

Step 2, the actual measurement proceeds as follows: Start an
unconstrained bulk data flow effect. In this section we provide two
alternative calculations using any modern TCP (with large buffers
and/or autotuning). During the first interval (no rate limits)
observe the slowstart (e.g. tcpdump) and measure: Peak burst size;
link clock rate (delivery rate for each round); peak data rate for
the fastest single RTT interval; fraction of segments lost at the end different assumptions.

It may seem out of slowstart. After the flow has fully recovered from the slowstart
(details not important) throttle the flow down to the test rate (by
clamping cwnd or application pacing at the sender or receiver).
While clamped place to allow such latitude in a measurement
standard, but the test rate, observe the losses (run length) for
the chosen test duration. The link passes section provides offsetting requirements.

These models provide estimates that make the test most sense if network
performance is viewed logarithmically. In the slowstart
ends with less operational internet,
data rates span more than approximately 50% losses 8 orders of magnitude, RTT spans more than
3 orders of magnitude, and no timeouts, the
peak rate is loss probability spans at least the target rate, and the measured run length 8 orders
of magnitude. When viewed logarithmically (as in decibels), these
correspond to 80 dB of dynamic range. On an 80 db scale, a 3 dB
error is
better less than 4% of the required run length. There will also need to be some
ancillary metrics, for example to discard tests where the receiver
closes the window, invalidating the slowstart test. [This needs to
be separated into multiple subtests]

Optional step 3: In some cases scale, even though it might make sense to compute an
"extrapolated rate", which is the minimum represent a
factor of 2 in raw parameter.

Although this document gives a lot of latitude for calculating
target_run_length, people designing suites of tests need to consider
the observed peak rate,
and the rate computed from effect of their choices on the specified target RTT ongoing conversation and tussle
about the observed
run length by using relevance of "TCP friendliness" as an appropriate model for
capacity allocation. Choosing a suitable TCP performance model. The
extrapolated rate should be annotated to indicate if it was run
length or peak rate limited, since these have different predictive
values.

Other issues:

If the link RTT target_run_length that is not
substantially smaller than the target RTT and
the actual run length reference target_run_length specified
in Section 5.2 is close to the target rate, a standards
compliant TCP implementation might not be effective at accurately
controlling the data rate. To be independent of the details of the
TCP implementation, failing equivalent to control saying that it is appropriate for the rate has
transport research community to be treated abandon "TCP friendliness" as a spoiled measurement, not a infrastructure failure. This can be
overcome by "stiffening" TCP by using a non-standard congestion
control algorithm. For example if the rate controlling by clamping
cwnd then use "relentless TCP" style reductions on loss,
fairness model and lock
ssthresh to the cwnd clamp. Alternatively, implement an explicit
rate controller develop more aggressive Internet transport
protocols, and for TCP. In either case the test must be abandoned
(aborted) if applications to continue (or even increase) the measured run length
number of connections that they open concurrently.

A.1. Aggregate Reno

In Section 5.2 it is substantially below assumed that the target run length.

If the test rate is run "in situ" in the same as the
link rate, and any excess window causes a production environment, there also
needs to standing queue at the
bottleneck. This might be baseline tests using alternate paths representative of a non-shared access
link. An alternative situation would be a heavily aggregated subpath
where individual flows do not significantly contribute to confirm that
there are no bottlenecks or congested links between the test end
points
queueing delay, and losses are determined monitoring the link under test.

It might make sense to run multiple tests with different parameters, average data
rate, for example infrequent tests with test rate equal to the target rate,
and more frequent, less disruptive tests with the same target rate
but by the test rate equal to 1% use of a virtual queue as in [AFD]. In such
a scheme the target rate. To observe the
required run length, RTT is constant and TCP's AIMD congestion control causes
the low data rate test would take 100 times longer to
run.

Returning to fluctuate in a sawtooth. If the example: traffic is being
controlled in a full rate test would entail sending 690
pps (1 MByte/s) for several tens of seconds (e.g. 50k packets), and
observing manner that the total loss rate is below 1:1e4. A less disruptive
test might consistent with the metrics here, goal
would be to send at 6.9 pps for 100 times longer, and observing

B.2. End-to-end parameters from subpaths

[This entire section needs make the actual average rate equal to be overhauled and should be skipped on the
target_data_rate.

We can derive a first reading. The concepts defined here are not used elsewhere.]

The following optional parameters apply model for testing generalized end-
to-end paths that include subpaths with known specific types of
behaviors that are not well represented by simple queueing models:

Bottleneck link clock rate: This applies to links that are using
virtual queues or other techniques to police or shape users
traffic at lower rates full link rate. The bottleneck link clock
rate should be representative Reno TCP and delayed ACK under the above
set of queue drain times assumptions: for short
bursts some value of Wmin, the window will sweep
from Wmin to 2*Wmin in 2*Wmin RTT. Between losses each sawtooth
delivers (1/2)(Wmin+2*Wmin)(2Wmin) packets on an otherwise unloaded link.
Channel hold time: For channels that have relatively expensive
channel arbitration algorithms, this is in 2*Wmin round trip
times. However, unlike the typical (maximum?)
time that data queueing case where Wmin =
Target_pipe_size, we want the average of Wmin and or ACKs are held pending acquiring 2*Wmin to be the channel.
While under heavy load,
target_pipe_size, so the RTT may be inflated by this parameter,
unless it average rate is built into the target RTT
Preload traffic volume: If the user's traffic rate. Thus we
want Wmin = (2/3)*target_pipe_size.

(@@@@ something is shaped on the basis
of average traffic volume, wrong above) Substituting these together we get:

target_run_length = (8/3)(target_pipe_size^2)

Note that this is volume necessary to invoke
"heavy hitter" policies.
Unloaded traffic volume: If always 88% of the user's traffic is shaped on reference run length.

A.2. CUBIC

CUBIC has three operating regions. The model for the
basis expected value
of average traffic volume, this is window size derived in [LMCUBIC] assumes operation in the maximum traffic
volume that
"concave" region only, which is a test can use non-TCP friendly region for long-
lived flows. The authors make the following assumptions: packet loss
probability, p, is independent and stay within periodic, losses occur one at a "light user"
policies.

Note
time, and they are true losses due to tail drop or corruption. This
definition of p aligns very well with our definition of
target_run_length and the requirement for progressive loss (AQM).

Although CUBIC window increase depends on a ConEx enabled network [ConEx], continuous time, the word "traffic" in
authors transform the
last two items should be replaced by "congestion" i.e. "preload
congestion volume" and "unloaded congestion volume".

B.3. Per subpath parameters

[This entire section needs time to be overhauled reach the maximum Window size in terms
of RTT and should be skipped on a first reading. The concepts defined here are not used elsewhere.]

Some single parameter tests for the multiplicative rate decrease on
observing loss, beta (whose default value is 0.2 in CUBIC). The
expected value of Window size, E[W], is also need dependent on C, a
parameter of CUBIC that determines its window-growth aggressiveness
(values from 0.01 to 4).

E[W] = ( C*(RTT/p)^3 * ((4-beta)/beta) )^-4

and, further assuming Poisson arrival, the subpath.

subpath RTT: RTT of the subpath mean throughput, x, is

x = E[W]/RTT

We note that under test.
subpath link clock rate: If different these conditions (deterministic single losses),
the value of E[W] is always greater than 0.8 of the Bottleneck link clock
rate

B.4. maximum window
size ~= reference_run_length. (as far as I can tell)

Commentary on the consequence of the choice.

Appendix B. Version Control

Formatted: Fri Jun Mon Oct 21 18:23:29 15:42:35 PDT 2013

Authors' Addresses

Matt Mathis
Google, Inc
1600 Amphitheater Parkway
Mountain View, California 93117
USA

Email: mattmathis@google.com

Al Morton
AT&T Labs
200 Laurel Avenue South
Middletown, NJ 07748
USA

Phone: +1 732 420 1571
Email: acmorton@att.com
URI: http://home.comcast.net/~acmacm/