RFC 8337
Model-Based Metrics for Bulk Transport Capacity
Model-Based Metrics for Bulk Transport Capacity 🎀
Part 1 of 13 💖
- RFC 8337 🦋 (IETF, Request for Comments)
- Category: Experimental 🧪 (this is a trial run, not a finished standard)
- Date: March 2018 📅
- Authors: M. Mathis (Google, Inc) and A. Morton (AT&T Labs) 👑
- ISSN: 2070-1721
- Obsoletes / Updates: the header doesn't list anything it obsoletes or updates 🌸
Abstract ✨
okay so this doc introduces a brand new class of Model-Based Metrics 💅 The goal is to figure out whether a whole Internet path can be expected to hit a Target Transport Performance you've already picked out. It does this by running a bunch of IP diagnostic tests on the path one subpath at a time. The best part is that these subpath tests are sturdy enough to use on important infrastructure, like network interconnections or even single devices 🔥 So if one piece of the route is going to stop paths going through it from hitting the target, these tests can catch it.
Model-Based Metrics lean on mathematical models to build a Targeted IP Diagnostic Suite. That's a set of IP diagnostic tests made to check whether common transport protocols can be expected to hit a set Target Transport Performance over an Internet path 🦋
For Bulk Transport Capacity, the diagnostics are built from test streams and statistical rules for judging packet transfer. They copy how TCP behaves over the whole path 🧚 The timing of the test stream (bursts and so on) mimics TCP or other transport protocols moving bulk data over a long path. But here's the iconic part: the tests are built to be independent of the subpath being tested, the end systems, and the applications. The success rules check the subpath's packet transfer stats against limits worked out by protocol performance models, using the Target Transport Performance of the full path. Those rules don't depend on the subpath, end systems, or applications either 💜
Status of This Memo 🌈
ngl this part is a bit formal, but here's the tea. This document is not an Internet Standards Track specification. It's been published so people can examine it, try out experimental implementations, and evaluate it 🧪
It does define an Experimental Protocol for the Internet community. It's a product of the Internet Engineering Task Force (IETF) and represents the consensus of the IETF community. It's had public review and got approved for publication by the Internet Engineering Steering Group (IESG). Not every document the IESG approves can become any level of Internet Standard, so check out Section 2 of RFC 7841 if you want the details 🙈
For info on the current status of this document, any errata, and how to send feedback, head to https://www.rfc-editor.org/info/rfc8337 🫶
Copyright Notice 💌
Copyright (c) 2018 IETF Trust and the persons identified as the document authors. All rights reserved ✨
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents, found at https://trustee.ietf.org/license-info, as they stand on the date this document was published. Please read these carefully, bestie, because they spell out your rights and restrictions for this document 📜 Code Components pulled out of this document must include Simplified BSD License text, as described in Section 4.e of the Trust Legal Provisions. They come without warranty, just like the Simplified BSD License says 💀
Table of Contents 🗺️
1. Introduction ....................................................4
2. Overview ........................................................5
3. Terminology .....................................................8
3.1. General Terminology ........................................8
3.2. Terminology about Paths ...................................10
3.3. Properties ................................................11
3.4. Basic Parameters ..........................................12
3.5. Ancillary Parameters ......................................13
3.6. Temporal Patterns for Test Streams ........................14
3.7. Tests .....................................................15
4. Background .....................................................16
4.1. TCP Properties ............................................18
4.2. Diagnostic Approach .......................................20
4.3. New Requirements Relative to RFC 2330 .....................21
5. Common Models and Parameters ...................................22
5.1. Target End-to-End Parameters ..............................22
5.2. Common Model Calculations .................................22
5.3. Parameter Derating ........................................23
5.4. Test Preconditions ........................................24
6. Generating Test Streams ........................................24
6.1. Mimicking Slowstart .......................................25
6.2. Constant Window Pseudo CBR ................................27
6.3. Scanned Window Pseudo CBR .................................28
6.4. Concurrent or Channelized Testing .........................28
7. Interpreting the Results .......................................29
7.1. Test Outcomes .............................................29
7.2. Statistical Criteria for Estimating run_length ............31
7.3. Reordering Tolerance ......................................33
8. IP Diagnostic Tests ............................................34
8.1. Basic Data Rate and Packet Transfer Tests .................34
8.1.1. Delivery Statistics at Paced Full Data Rate ........35
8.1.2. Delivery Statistics at Full Data Windowed Rate .....35
8.1.3. Background Packet Transfer Statistics Tests ........35
8.2. Standing Queue Tests ......................................36
8.2.1. Congestion Avoidance ...............................37
8.2.2. Bufferbloat ........................................37
8.2.3. Non-excessive Loss .................................38
8.2.4. Duplex Self-Interference ...........................38
8.3. Slowstart Tests ...........................................39
8.3.1. Full Window Slowstart Test .........................39
8.3.2. Slowstart AQM Test .................................39
8.4. Sender Rate Burst Tests ...................................40
8.5. Combined and Implicit Tests ...............................41
8.5.1. Sustained Full-Rate Bursts Test ....................41
8.5.2. Passive Measurements ...............................42
9. Example ........................................................43
9.1. Observations about Applicability ..........................44
10. Validation ....................................................45
11. Security Considerations .......................................46
12. IANA Considerations ...........................................47
13. Informative References ........................................47
Appendix A. Model Derivations ....................................52
A.1. Queueless Reno ............................................52
Appendix B. The Effects of ACK Scheduling ........................53
Acknowledgments ...................................................55
Authors' Addresses ................................................551. Introduction ✨
okay so Model-Based Metrics (MBM) run on peer-reviewed mathematical models 📚 They use those models to build a Targeted IP Diagnostic Suite (TIDS), which is a set of IP diagnostic tests made to check whether common transport protocols can be expected to hit a predetermined Target Transport Performance over an Internet path. This doc explains the modeling framework used to work out the test parameters for checking whether an Internet path can support a predetermined Bulk Transport Capacity 🦋
Each test in TIDS measures some piece of IP packet transfer that's needed to hit the Target Transport Performance 🎯 For Bulk Transport Capacity, TIDS has IP diagnostic tests to confirm there's enough IP capacity (data rate) 🏎️ and enough queue space at bottlenecks to soak up and deliver the usual transport bursts 🫧 It also checks that background packet loss is low enough not to mess with congestion control, plus a few other properties we'll get to below. The big difference from typical IP Performance Metrics (IPPM), which give you measures of network properties, is that Model-Based Metrics basically give a pass/fail verdict on whether standard transport protocols can hit the specific performance goal over some network path 💅
Most of the time, the IP diagnostic tests can be built by combining existing IPPM metrics with extra controls 🧩 Those controls generate test streams with a set temporal structure (bursts or standing queues caused by constant bit rate streams, and so on) and add statistical criteria for judging packet transfer. The temporal structure copies how transport protocols behave over the whole path, and the statistical criteria model how the transport protocol responds to less-than-ideal packet transfer 🌸 In control theory terms, the tests are "open loop" 🔁 Heads up: running a test needs the sending and receiving measurement points to work together 🤝
This document is all about Bulk Transport Capacity 💖 It offers an alternative to the approach described in "A Framework for Defining Empirical Bulk Transfer Capacity Metrics" [RFC3148]. Other Model-Based Metrics might cover other apps and transports, like Voice over IP (VoIP) over UDP, RTP, and brand-new transport protocols 🧚
This document assumes a traditional Reno TCP-style, self-clocked, window-controlled transport protocol 🪟 It uses packet loss and Explicit Congestion Notification (ECN) Congestion Experienced (CE) marks as its congestion feedback. Some experimental protocols and congestion control algorithms are rate based, or otherwise don't fit these assumptions 🙈 Down the road, those newer protocols and algorithms might call for revised models 🔮
The MBM approach maps a Target Transport Performance onto a Targeted IP Diagnostic Suite (TIDS) of IP tests, and it fixes some intrinsic problems with using TCP or other throughput-maximizing protocols for measurement 💅 Here's the drama: all throughput-maximizing protocols (TCP congestion control especially) cause some congestion just to find out when they've hit the network's available capacity limit 😭 That self-inflicted congestion muddies the network properties you actually care about, and it creates non-linear dynamic equilibrium behaviors. So the measurements end up useless as metrics, because they can't predict anything for conditions or paths other than the measurement itself 💀 To prevent this, the measurement method has to avoid the effects of TCP congestion control. Section 4 covers these issues at length. If you're new to the basics of TCP and TCP-like congestion control, you might find it easier to start at Section 4 or Section 4.1 🎀
A Targeted IP Diagnostic Suite doesn't have those problems 👑 IP diagnostics can be built to make strong statistical statements about path properties that don't depend on measurement details, like the vantage point or which measurement points you pick 🫶
2. Overview ✨
okay so bestie, this section is the big-picture tour 🗺️🦋. It's a modeling framework for coming up with a Targeted IP Diagnostic Suite (TIDS) from a Target Transport Performance you've already picked. It is NOT a full spec on its own 🙅♀️. It leans on other standards docs for the juicy details, like packet type-P selection, sampling techniques, vantage selection, and so on. A Fully Specified Targeted IP Diagnostic Suite (FSTIDS) is the version where ALL of those details are nailed down 💅. A TIDS is just the slice of an FSTIDS that's in scope for this doc. Section 3 has the full terminology deets 💜.
Section 4 is where we get into the TCP behavior stuff 🎀, and what that behavior means for what IP packet transfer has to look like. Most of the IP diagnostic tests needed to check that a path has these properties can be built on top of existing IPPM metrics 🔥. You add statistical criteria for judging packet transfer, and in a few cases you need brand-new mechanisms to set up the required temporal structure ⏱️. (One group, the standing queue tests in Section 8.2, doesn't line up with any existing IPPM metric. But new IPPM metrics can be modeled after the existing ones, so it's not a dead end 🙌.)
Figure 1 shows the whole MBM modeling and measurement framework 🧚. The Target Transport Performance at the top is set by what the user or application needs, and that's outside the scope of this doc 🙈. For Bulk Transport Capacity, the star of the show is the Target Data Rate 🌈. BUT TCP's ability to bounce back from less-than-ideal network conditions is seriously shaped by the Round-Trip Time (RTT) and the Maximum Transmission Unit (MTU) of the complete path, so those parameters need to be locked in ahead of time too, based on what the intended application setting looks like 📏. They can represent a specific application running over a real Internet path, or an idealized application and a made-up path standing in for a typical user community 🫶. Section 5 covers the common parameters and the models that come out of the Target Transport Performance 💖.
Target Transport Performance
(Target Data Rate, Target RTT, and Target MTU)
|
________V_________
| mathematical |
| models |
| |
------------------
Traffic parameters | | Statistical criteria
| |
_______V____________V____Targeted IP____
| | * * * | Diagnostic Suite |
_____|_______V____________V________________ |
__|____________V____________V______________ | |
| IP diagnostic tests | | |
| | | | | |
| _____________V__ __V____________ | | |
| | traffic | | Delivery | | | |
| | pattern | | Evaluation | | | |
| | generation | | | | | |
| -------v-------- ------^-------- | | |
| | v test stream via ^ | | |--
| | -->======================>-- | | |
| | subpath under test | |-
----V----------------------------------V--- |
| | | | | |
V V V V V V
fail/inconclusive pass/fail/inconclusive
(traffic generation status) (test result)
Figure 1: Overall Modeling FrameworkSo the math is the mood here 📐✨. Mathematical TCP models figure out the traffic parameters, and then they're used to design traffic patterns that copy TCP's vibe (TCP is super bursty across multiple time scales 💅) or other bulk-data transport protocols. These patterns run at the Target Data Rate, MTU, and RTT across a full range of conditions 🌸. Section 6 explains the techniques for generating them, and they're built from the three Target parameters of the complete path (Target Data Rate, Target RTT, and Target MTU), NOT from the properties of individual subpaths 🧩. As much as possible, the test streams are precomputed (generated deterministically) so that test methodology, measurement points, measurement vantage, and path partitioning don't mess with the details of the measurement traffic 🙅♀️✨.
Section 7 is all about packet transfer statistics and the methods for testing against the statistical criteria the models hand you 📊. The statistical criteria usually apply to the complete path, which is a composition of subpaths [RFC6049] 🔗. So for in situ testing, the end-to-end criteria have to be split up into separate criteria for each subpath 🍰. Subpaths expected to be bottlenecks get to take a bigger slice of the end-to-end packet loss budget 🔥. To balance that out, subpaths that aren't expected to show bottlenecks must be kept to less packet loss 🥺. So the statistical criteria for each subpath in each test of a TIDS is an apportioned share of the end-to-end criteria for the complete path, as determined by the mathematical model 💜.
Section 8 lays out the suite of individual tests needed to verify ALL of the required IP delivery properties 💪. A subpath passes if and only if every single individual IP diagnostic test passes 🎯. If any test fails on a subpath, that means some users will probably fail to reach their Target Transport Performance under some conditions 😬. Tests can also come back inconclusive, not just pass or fail 🤷♀️. Reasons include: the precomputed traffic pattern wasn't generated accurately, the measurement results weren't statistically significant, the test didn't meet some required test preconditions, and so on 🙃. And if everything passes but some tests are inconclusive, the whole suite gets labeled inconclusive too 🫠.
In Section 9, we walk through an example TIDS that could be representative of High Definition (HD) video 📺🍓. It also shows how Model-Based Metrics can handle tricky measurement situations, like confirming that inter-carrier exchanges have enough performance and capacity to deliver HD video between ISPs 🌐✨.
Since the modeling process has some uncertainty baked in, Section 10 describes a validation procedure to diagnose and minimize false positive and false negative results 🔍💖.
3. Terminology 💜
Okay, quick heads-up on the terminology 🎀. Terms with underscores (instead of spaces) show up in equations, and they usually come with algorithmic definitions 🧮. Keep an eye out for those, bestie 👀.
3.1. General Terminology ✨
Target: okay so this is the umbrella term 🎀 Basically any parameter that comes from the user's application or transport performance requirements counts as a Target.
Target Transport Performance: these are the performance target values for the complete path 🛣️ For Bulk Transport Capacity (BTC, defined in this doc), it includes the Target Data Rate, Target RTT, and Target MTU, all explained below 💅
Target Data Rate: the specified application data rate an app needs to run properly 🚀 Conventional BTC metrics are all about the Target Data Rate, but ngl they have little or no predictive value because they ignore the other two parts of Target Transport Performance: the RTT and MTU of the complete paths 💀
Target RTT (Round-Trip Time): this is the specified baseline (minimum) RTT of the longest complete path where the user expects to still hit their target performance 🔁 TCP and other transport protocols are generally better at recovering from path problems when they get more round trips per second. The Target RTT determines both the key parameters of the traffic patterns (e.g., burst sizes) and the thresholds on acceptable IP packet transfer statistics. The Target RTT must be specified considering appropriate packet sizes: MTU-sized packets on the forward path and ACK-sized packets (typically, header_overhead) on the return path. Note that Target RTT is specified and not measured. MBM measurements derived for a given target_RTT will be applicable to any path with a smaller RTT 🧚
Target MTU (Maximum Transmission Unit): this is the specified maximum MTU supported by the complete path where the application expects to hit the target performance 📏 In this document, we assume a 1500-byte MTU unless otherwise specified. If a subpath has a smaller MTU, that smaller one becomes the Target MTU for the complete path, and all model calculations and subpath tests must use the same smaller MTU 🙈
Targeted IP Diagnostic Suite (TIDS): a set of IP diagnostic tests built to figure out if an otherwise ideal complete path (the one containing the subpath under test) can sustain flows at a specific target_data_rate, using packets of size target_MTU, when the RTT of the complete path is target_RTT 🔍
Fully Specified Targeted IP Diagnostic Suite (FSTIDS): a TIDS plus extra specs, like the measurement packet type ("type-p" [RFC2330]). These are out of scope for this document, so they have to come from other standards documents 📚
Bulk Transport Capacity (BTC): BTC metrics check an Internet path's ability to carry bulk data, like large files, streaming (non-real-time) video, and, under some conditions, web images and other content 🎬 Earlier attempts to define BTC metrics were based on [RFC3148], which came out before we really understood TCP and the requirements in Section 4. Generally, "Bulk Transport" means performance is decided by the interplay between the network, cross traffic, and congestion control in the transport protocol. It doesn't include situations where performance is dominated by the RTT alone (e.g., transactions) or where bottlenecks are somewhere else, like in the application itself 🙅♀️
IP diagnostic tests: measurements or diagnostics that figure out if packet transfer statistics meet some precomputed target 📊
traffic patterns: the temporal patterns, or burstiness, of traffic that applications send over transport protocols like TCP 🌊 Several mechanisms cause bursts at different timescales, as described in Section 4.1. The goal here is to mimic the range of common patterns (burst sizes, rates, etc.) without tying the applicability to specific applications, implementations, or technologies, because those are sure to go stale 🥀
Explicit Congestion Notification (ECN): see [RFC3168] 🚦
packet transfer statistics: raw, detailed, or summary stats about the packet transfer properties of the IP layer. That includes packet losses, ECN Congestion Experienced (CE) marks, reordering, or any other properties that might matter for transport performance 📈
packet loss ratio: as defined in [RFC7680] 📉
apportioned: divided and allocated, like budgeting packet loss across multiple subpaths so the losses add up to less than a specified end-to-end loss ratio 💸 Apportioning metrics is basically the inverse of the process described in [RFC5835] 🙃
open loop: a control theory term for techniques where systems with natural circular dependencies can be analyzed by suppressing some of those dependencies, so the resulting dependency graph is acyclic 🔄
3.2. Terminology about Paths 🛤️
For existing terms and definitions, see [RFC2330] and [RFC7398] 📖
data sender: the host sending data and receiving ACKs 📤
data receiver: the host receiving data and sending ACKs 📥
complete path: the end-to-end path from the data sender to the data receiver 🌍
subpath: a portion of the complete path. There's no requirement that subpaths be non-overlapping, and a subpath can be as small as a single device, link, or interface 🔗
measurement point: measurement points as described in [RFC7398] 📍
test path: a path between two measurement points that includes a subpath of the complete path under test. If the measurement points are off path, the test path may include "test leads" between the measurement points and the subpath 🧪
dominant bottleneck: the bottleneck that generally determines most packet transfer statistics for the entire path 👑 It usually sets a flow's self-clock timing, packet loss, and ECN CE marking rate, while other potential bottlenecks have less effect on the packet transfer statistics. See Section 4.1 on TCP properties 🎀
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 the ACKs take from the data receiver back to the data sender 🔙
cross traffic: other, potentially interfering, traffic competing for network resources (like bandwidth and/or queue capacity) 🚗💨
3.3. Properties 🎀
The following properties are determined by the complete path and application. They're described in more detail in Section 5.1 ✨
Application Data Rate: the general term for the data rate as seen by the application above the transport layer, in bytes per second 💕 This is the payload data rate, and it explicitly excludes transport-level and lower-level headers (TCP/IP or other protocols), retransmissions, and any other overhead that isn't part of the total quantity of data delivered to the application 🫶
IP rate: the actual number of IP-layer bytes delivered through a subpath per unit time, including TCP and IP headers, retransmits, and other TCP/IP overhead 📶 This is the same as IP-type-P Link Usage in [RFC5136].
IP capacity: the maximum number of IP-layer bytes that can be transmitted through a subpath per unit time, including TCP and IP headers, retransmits, and other TCP/IP overhead 🧢 This is the same as IP-type-P Link Capacity in [RFC5136].
bottleneck IP capacity: the IP capacity of the dominant bottleneck in the forward path 🍾 All throughput-maximizing protocols estimate this capacity by observing the IP rate delivered through the bottleneck, and most protocols derive their self-clocks from the timing of this data. See Section 4.1 and Appendix B for more details 🔍
implied bottleneck IP capacity: the bottleneck IP capacity implied by the ACKs returning from the receiver 🕵️♀️ It's determined by looking at how much application data the ACK stream at the sender reports as delivered to the data receiver per unit time, at various timescales. If the return path is thinning, batching, or otherwise altering the ACK timing, the implied bottleneck IP capacity over short timescales might be substantially larger than the bottleneck IP capacity averaged over a full RTT 😳 Since TCP derives its clock from the data delivered through the bottleneck, the front path must have sufficient buffering to absorb any data bursts at the dimensions (size and IP rate) implied by the ACK stream. Those bursts are potentially doubled during slowstart 🫠 If the return path is not altering the ACK stream, the implied bottleneck IP capacity will be the same as the bottleneck IP capacity. See Section 4.1 and Appendix B for more details.
sender interface rate: the IP rate that corresponds to the IP capacity of the data sender's interface 🖥️ Thanks to sender efficiency algorithms, including technologies like TCP segmentation offload (TSO), nearly all modern servers deliver data in bursts at full interface link rate. Today, 1 or 10 Gb/s are typical ⚡
header_overhead: the IP and TCP header sizes, which are the portion of each MTU that isn't available for carrying application payload 📦 Without loss of generality, this is assumed to be the size for returning acknowledgments (ACKs). For TCP, the Maximum Segment Size (MSS) is the Target MTU minus the header_overhead 🧮
3.4. Basic Parameters ✨
Okay bestie, these are the basic parameters that show up for both the models and the subpath tests 🎀 The formulas for target_window_size and target_run_length live in Section 5.2. Heads up, these are mixed between application transport performance (excludes headers) and IP performance (includes TCP headers and retransmissions as part of the IP payload) 👀
Network power: The observed data rate divided by the observed RTT. It shows how effectively a transport protocol is filling a network 💅
Window [size]: The total quantity of data carried by packets in-flight plus the data represented by ACKs circulating in the network is what we call the window 🦋 See Section 4.1. People sometimes add other qualifiers, like congestion window (cwnd) or receiver window, to show which mechanism is controlling the window.
pipe size: A general term for the number of packets needed in flight (the window size) to exactly fill a network path or subpath 🌊 It corresponds to the window size that maximizes network power. It's often paired with extra qualifiers to say which path, under what conditions, etc.
target_window_size: The average number of packets in flight (the window size) needed to meet the Target Data Rate for the specified Target RTT and Target MTU. It implies the scale of the bursts that the network might experience 💥
run length: A general term for the observed, measured, or specified number of packets that are (expected to be) delivered between losses or ECN CE marks. Nominally, it is one over the sum of the loss and ECN CE marking probabilities, if they are independently and identically distributed 🎲
target_run_length: An estimate of the minimum number of non-congestion marked packets needed between losses or ECN CE marks to attain the target_data_rate over a path with the specified target_RTT and target_MTU. It's computed by a mathematical model of TCP congestion control 🧮 A reference calculation is shown in Section 5.2, and alternatives are in Appendix A.
reference target_run_length: This is target_run_length computed precisely by the method in Section 5.2 🎯 It's likely to be slightly more conservative than required by modern TCP implementations.
3.5. Ancillary Parameters 🧚
The following ancillary parameters are 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 test parameters, as described in Section 5.3. In exchange, you get more stringent TIDS validation procedures, described in Section 10. Models can be derated by including a multiplicative derating factor to make tests less stringent.
subpath_IP_capacity: The IP capacity of a specific subpath.
test path: A subpath of a complete path under test.
test_path_RTT: The RTT observed between two measurement points using packet sizes that are consistent with the transport protocol. This is generally MTU-sized packets of the forward path and packets with a size of header_overhead on the return path.
test_path_pipe: The pipe size of a test path. Nominally, it is the test_path_RTT times the test path IP_capacity.
test_window: The smallest window sufficient to meet or exceed the target_rate when operating with a pure self-clock over a test path 🌸 The test_window is typically calculated as follows (but see the discussion in Appendix B about the effects of channel scheduling on RTT):
ceiling(target_data_rate * test_path_RTT / (target_MTU -
header_overhead))On some test paths, the test_window may need to be adjusted slightly to compensate for the RTT being inflated by the devices that schedule packets.
3.6. Temporal Patterns for Test Streams 🎶
The terminology below is used to define temporal patterns for test streams. These patterns are designed to mimic TCP behavior, as described in Section 4.1.
packet headway: Time interval between packets, specified from the start of one to the start of the next. For example, if packets are sent with a 1 ms headway, there will be exactly 1000 packets per second ⏱️
burst headway: Time interval between bursts, specified from the start of the first packet of one burst to the start of the first packet of the next burst. For example, if 4 packet bursts are sent with a 1 ms burst headway, there will be exactly 4000 packets per second 🔥
paced single packets: Individual packets sent at the specified rate or packet headway.
paced bursts: Bursts on a timer 🥁 Specify any 3 of the following: average data rate, packet size, burst size (number of packets), and burst headway (burst start to start). By default, the bursts are assumed to occur at full sender interface rate, such that the packet headway within each burst is the minimum supported by the sender's interface. Under some conditions, it is useful to explicitly specify the packet headway within each burst.
slowstart rate: Paced bursts of four packets each, at an average data rate equal to twice the implied bottleneck IP capacity (but not more than the sender interface rate). This mimics TCP slowstart 🐣 It's a two-level burst pattern described in more detail in Section 6.1. If the implied bottleneck IP capacity is more than half of the sender interface rate, the slowstart rate becomes the sender interface rate.
slowstart burst: A specified number of packets in a two-level burst pattern that resembles slowstart. This mimics one round of TCP slowstart.
repeated slowstart bursts: Slowstart bursts repeated once per target_RTT. For TCP, each burst would be twice as large as the prior burst, and the sequence would end at the first ECN CE mark or lost packet. For measurement, all slowstart bursts would be the same size (nominally target_window_size, but other sizes might be specified), and the ECN CE marks and lost packets are counted 📊
3.7. Tests 💜
The tests described in this document can be grouped according to their applicability.
Capacity tests: Capacity tests determine if a network subpath has sufficient capacity to deliver the Target Transport Performance 💪 As long as the test stream is within the proper envelope for the Target Transport Performance, the average packet losses or ECN CE marks must be below the statistical criteria computed by the model. As such, capacity tests reflect parameters that can transition from passing to failing as a consequence of cross traffic, additional presented load, or the actions of other network users. By definition, capacity tests also consume significant network resources (data capacity and/or queue buffer space), so the test schedules must be balanced by their cost.
Monitoring tests: Monitoring tests are designed to capture the most important aspects of a capacity test without presenting excessive ongoing load themselves 🫶 As such, they may miss some details of the network's performance, but they can serve as a useful reduced-cost proxy for a capacity test. For example, they can support continuous production network monitoring.
Engineering tests: Engineering tests evaluate how network algorithms (such as Active Queue Management (AQM) and channel allocation) interact with TCP-style self-clocked protocols and adaptive congestion control based on packet loss and ECN CE marks 🧪 These tests are likely to have complicated interactions with cross traffic and, under some conditions, can be inversely sensitive to load. For example, a test to verify that an AQM algorithm causes ECN CE marks or packet drops early enough to limit queue occupancy may experience a false pass result in the presence of cross traffic 😭 It is important that engineering tests 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 less likely to be useful for engineering tests, although sparse in situ testing might be appropriate.
4. Background ✨
okay bestie, let's talk about why measuring bulk throughput is SO hard 😭. When "Framework for IP Performance Metrics" [RFC2330] came out in 1998, sound Bulk Transport Capacity (BTC) measurement was known to be well beyond our capabilities 💀. Even when "A Framework for Defining Empirical Bulk Transfer Capacity Metrics" [RFC3148] dropped, we knew we didn't really understand the problem 🙈. Now, in hindsight, we get why assessing BTC is such a difficult problem:
- TCP is a control system with circular dependencies 🌀 -- everything affects performance, including components that are explicitly not part of the test (for example, the host processing power is not in-scope of path performance tests). It's giving chaos, no cap.
- Congestion control is a dynamic equilibrium process, kind of like stuff you see in chemistry 🧪. The network and transport protocols find an operating point that balances opposing forces: the transport protocol pushes harder (raising the data rate and/or window) while the network pushes back (raising packet loss ratio, RTT, and/or ECN CE marks). By design, TCP congestion control keeps raising the data rate until the network gives some indication that its capacity has been exceeded by dropping packets or adding ECN CE marks. If a TCP sender accurately fills a path to its IP capacity (e.g., the bottleneck is 100% utilized), then packet losses and ECN CE marks are mostly determined by the TCP sender and how aggressively it seeks additional capacity. They are not determined by the network itself, because the network must send exactly the signals that TCP needs to set its rate. Lowkey iconic, honestly ✨.
- TCP's ability to compensate for network impairments (such as loss, delay, and delay variation, outside of those caused by TCP itself) is directly proportional to the number of send-ACK round-trip exchanges per second (i.e., inversely proportional to the RTT). As a consequence, an impaired subpath may pass a short RTT local test even though it fails when the subpath is extended by an effectively perfect network to some larger RTT. The way that short tests can lie to you 🙃.
- TCP has an extreme form of the Observer Effect (colloquially known as the "Heisenberg Effect") 🦋. 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 bounds on the relative momentum of the measurement and measured particles. In general, for network measurement, you cannot determine even the order of magnitude of the effect. It is possible to construct measurement scenarios where the measurement traffic starves real user traffic, yielding an overly inflated measurement. The inverse is also possible: the user traffic can fill the network, such that the measurement traffic detects only minimal available capacity. In general, you cannot determine which scenario might be in effect, so you cannot gauge the relative magnitude of the uncertainty introduced by interactions with other network traffic. Main character energy for the measurement, but not in a good way 💀.
- As a consequence of the properties listed above, it is difficult, if not impossible, for two independent implementations (hardware or software) of TCP congestion control to produce equivalent performance results [RFC6576] under the same network conditions 😭.
These properties are a consequence of the dynamic equilibrium behavior intrinsic to how all throughput-maximizing protocols interact with the Internet. These protocols rely on control systems based on estimated network metrics to regulate the quantity of data to send into the network. The packet-sending characteristics in turn alter the network properties estimated by the control system metrics, such that there are circular dependencies between every transmission characteristic and every estimated metric. Since some of these dependencies are nonlinear, the entire system is nonlinear, and any change anywhere causes a difficult-to-predict response in network metrics. As a consequence, Bulk Transport Capacity metrics have not fulfilled the analytic framework envisioned in [RFC2330]. Okay, that's a lot, but you get the vibe 🫶.
Model-Based Metrics overcome these problems by making the measurement system open loop 🔓: the packet transfer statistics (akin to the network estimators) do not affect the traffic or traffic patterns (bursts), which are computed on the basis of the Target Transport Performance. A path or subpath meeting the Target Transfer Performance requirements would exhibit packet transfer statistics and estimated metrics that would not cause the control system to slow the traffic below the Target Data Rate. So the fix is to stop the test from messing with itself 💅.
4.1. TCP Properties
TCP and other self-clocked protocols (e.g., the Stream Control Transmission Protocol (SCTP)) carry the vast majority of all Internet data 🌐. Their dominant bulk data transport behavior is to have an approximately fixed quantity of data and acknowledgments (ACKs) circulating in the network. The data receiver reports arriving data by returning ACKs to the data sender, and the data sender typically responds by sending approximately the same quantity of data back into the network. The total 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 window by sending slightly more or less data in response to each ACK. The fundamentally important property of this system is that it is self-clocked 🎀: the data transmissions are a reflection of the ACKs that were delivered by the network, and the ACKs are a reflection of the data arriving from the network. It's a whole love story, honestly 🥹.
A number of protocol features cause bursts of data, even in idealized networks that can be modeled as simple queuing systems 🍓.
During slowstart, the IP rate is doubled on each RTT by sending twice as much data as was delivered to the receiver during the prior RTT. Each returning ACK causes the sender to transmit twice the data the ACK reported arriving at the receiver. For slowstart to be able to fill the pipe, 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 or ECN CE mark. For example, with classic Reno congestion control, an optimal slowstart has to end with a burst that is twice the bottleneck rate for one RTT in duration. This burst causes a queue that is equal to the pipe size (i.e., the window is twice the pipe size), so when the window is halved in response to the first packet loss, the new window will be the pipe size. Math that actually slaps 🧮🔥.
Note that if the bottleneck IP rate is less than half of the capacity of the front path (which is almost always the case), the slowstart bursts will not by themselves cause significant queues anywhere else along the front path. They primarily exercise the queue at the dominant bottleneck 🎯.
Several common efficiency algorithms also cause bursts. The self-clock is typically applied to groups of packets: the receiver's delayed ACK algorithm generally sends only one ACK per two data segments. Furthermore, modern senders use TCP segmentation offload (TSO) to reduce CPU overhead. The sender's software stack builds super-sized TCP segments that the TSO hardware splits into MTU-sized segments on the wire. The net effect of TSO, delayed ACK, and other efficiency algorithms is to send bursts of segments at full sender interface rate 💨.
Note that these efficiency algorithms are almost always in effect, including during slowstart, such that slowstart typically has a two-level burst structure. Section 6.1 describes slowstart in more detail 📖.
Additional sources of bursts include TCP's initial window [RFC6928], application pauses, channel allocation mechanisms, and network devices that schedule ACKs. Appendix B describes these last two items. If the application pauses (e.g., stops reading or writing data) for some fraction of an RTT, many TCP implementations catch up to their earlier window size by sending a burst of data at the full sender interface rate. To fill a network with a realistic application, the network has to be able to tolerate sender interface rate bursts large enough to restore the prior window following application pauses. Even the pauses are chaotic 😭.
Although the sender interface rate bursts are typically smaller than the last burst of a slowstart, they are at a higher IP rate, so they potentially exercise queues at arbitrary points along the front path from the data sender up to and including the queue at the dominant bottleneck. It is known that these bursts can hurt network performance, especially in conjunction with other queue pressure. However, we are not aware of any models for estimating the impact or prescribing limits on the size or frequency of sender rate bursts. So yeah, that one's still a mystery 🫠.
In conclusion, to verify that a path can meet a Target Transport Performance, it is necessary to independently confirm that the path can tolerate bursts at the scales that can be caused by the above mechanisms. Three cases are believed to be sufficient 👑:
- Two-level slowstart bursts sufficient to get connections started properly.
- Ubiquitous sender interface rate bursts caused by efficiency algorithms. We assume four packet bursts to be the most common case, since it matches the effects of delayed ACK during slowstart. These bursts should be assumed not to significantly affect packet transfer statistics.
- Infrequent sender interface rate bursts that are the maximum of the full
target_window_sizeand the initial window size (10 segments in [RFC6928]). Thetarget_run_lengthmay be derated for these large fast bursts.
If a subpath can meet the required packet loss ratio for bursts at all of these scales, then it has sufficient buffering at all potential bottlenecks to tolerate any of the bursts that are likely introduced by TCP or other transport protocols. That's the whole vibe check passed ✅🌸.
4.2. Diagnostic Approach 💅
okay bestie, so a complete path is expected to be able to hit a specified Bulk Transport Capacity 🔥 if its RTT is equal to or smaller than the Target RTT, its MTU is equal to or larger than the Target MTU, and all of these vibes check out:
- The IP capacity is above the Target Data Rate by a sufficient margin to cover all TCP/IP overheads. This can be confirmed by the tests described in Section 8.1 or any number of IP capacity tests adapted to implement MBM.
- The observed packet transfer statistics are better than required by a suitable TCP performance model (for example, fewer packet losses or ECN CE marks). See Section 8.1 or any number of low- or fixed-rate packet loss tests outside of MBM.
- There is sufficient buffering at the dominant bottleneck to absorb a slowstart burst large enough to get the flow out of slowstart at a suitable window size. See Section 8.3.
- 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 any other mechanisms. See Section 8.4.
- When there is a slowly rising standing queue at the bottleneck, the onset of packet loss has to be at an appropriate point (in time or in queue depth) and has to be progressive, for example, by use of Active Queue Management [RFC7567]. See Section 8.2.
- When there is a standing queue at a bottleneck for a shared media subpath (for example, a half-duplex link), there MUST be a suitable bound on the interaction between ACKs and data, for example, due to the channel arbitration mechanism. See Section 8.2.4.
Note that conditions 1 through 4 require capacity tests for validation and thus may need to be monitored on an ongoing basis. Conditions 5 and 6 require engineering tests, which are best performed in controlled environments (for example, bench tests). They won't generally fail due to load but may fail in the field (for example, due to configuration errors, etc.) and thus SHOULD be spot checked. 🧐
A tool that can perform many of the tests is available from [MBMSource]. 🦋
4.3. New Requirements Relative to RFC 2330 ✨
Model-Based Metrics are designed to fulfill some additional requirements that were not recognized at the time RFC 2330 [RFC2330] was published. ngl these missing requirements may have significantly contributed to policy difficulties in the IP measurement space 😭 Some additional requirements are:
- IP 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. 🛠️
- Metrics SHOULD be spatially composable, such that measures of concatenated paths SHOULD be predictable from subpaths. 🧩
- Metrics MUST be vantage point invariant over a significant range of measurement point choices, including off-path measurement points. The only requirements for Measurement Point (MP) selection SHOULD be that the RTT between the MPs is below some reasonable bound and that the effects of the "test leads" connecting MPs to the subpath under test can be calibrated out of the measurements. The latter might be accomplished if the test leads are effectively ideal or their properties can be deducted from the measurements between the MPs. While many tests require that the test leads have at least as much IP capacity as the subpath under test, some do not, for example, the Background Packet Transfer Statistics Tests described in Section 8.1.3. 🎯
- Metric measurements SHOULD be repeatable by multiple parties with no specialized access to MPs or diagnostic infrastructure. It SHOULD be possible for different parties to make the same measurement and observe the same results. In particular, it is important that both a consumer (or the consumer's delegate) and ISP be able to perform the same measurement and get the same result. Note that vantage independence is key to meeting this requirement. 👑
5. Common Models and 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 3. These parameters are determined by the needs of the application or the ultimate end user and the complete Internet path over which the application is expected to operate. The target parameters are in units that make sense to layers above the TCP layer: payload bytes delivered to the application. They exclude overheads associated with TCP and IP headers, retransmits and other protocols (for example, DNS). Note that IP-based network services include TCP headers and retransmissions as part of delivered payload; this difference (header_overhead) is recognized in calculations below. 🍓
Other end-to-end parameters defined in Section 3 include the effective bottleneck data rate, the sender interface data rate, and the TCP and IP header sizes. ✨
The target_data_rate MUST be smaller than all subpath IP capacities by enough headroom to carry the transport protocol overhead, explicitly including retransmissions and an allowance for fluctuations in TCP's actual data rate. Specifying a target_data_rate with insufficient headroom is likely to result in brittle measurements that have little predictive value. 💀
Note that the target parameters can be specified for a hypothetical path (for example, to construct TIDS designed for bench testing in the absence of a real application) or for a live in situ test of production infrastructure. 🧚
The number of concurrent connections is explicitly not a parameter in this model. If a subpath requires multiple connections in order to meet the specified performance, that MUST be stated explicitly, and the procedure described in Section 6.4 applies. 🔗
5.2. Common Model Calculations 🧮
The Target Transport Performance is used to derive the target_window_size and the reference target_run_length. 💖
The target_window_size is the average window size in packets needed to meet the target_rate, for the specified target_RTT and target_MTU. To calculate target_window_size:
target_window_size = ceiling(target_rate * target_RTT / (target_MTU -
header_overhead))The target_run_length is an estimate of the minimum required number of unmarked packets that MUST be delivered between losses or ECN CE marks, as computed by a mathematical model of TCP congestion control. The derivation here is parallel to the derivation in [MSMO97] and, by design, is quite conservative. 🥹
The reference target_run_length is derived as follows. Assume the subpath_IP_capacity is infinitesimally larger than the target_data_rate plus the required header_overhead. Then, target_window_size also predicts the onset of queuing. A larger window will cause a standing queue at the bottleneck. 🌊
Assume the transport protocol is using standard Reno-style Additive Increase Multiplicative Decrease (AIMD) congestion control [RFC5681] (but not Appropriate Byte Counting [RFC3465]) and the receiver is using standard delayed ACKs. Reno increases the window by one packet every pipe size worth of ACKs. With delayed ACKs, this takes two RTTs per increase. To exactly fill the pipe, the spacing of losses must be no closer than when the peak of the AIMD sawtooth reached exactly twice the target_window_size. Otherwise, the multiplicative window reduction triggered by the loss would cause the network to be underfilled. Per [MSMO97] the number of packets between losses MUST be the area under the AIMD sawtooth. They MUST be no more frequent than every 1 in ((3/2)*target_window_size)*(2*target_window_size) packets, which simplifies to:
target_run_length = 3*(target_window_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. If a different model is used, an FSTIDS MUST document the actual method for computing target_run_length and the ratio between alternate target_run_length and the reference target_run_length calculated above, along with a discussion of the rationale for the underlying assumptions. 📚
Most of the individual parameters for the tests in Section 8 are derived from target_window_size and target_run_length. 🎯
5.3. Parameter Derating 🦄
Since some aspects of the models are very conservative, the MBM framework permits some latitude in derating test parameters. 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:
- The FSTIDS MUST document and justify the actual method used to compute the derated metric parameters.
- The validation procedures described in Section 10 MUST be used to demonstrate the feasibility of meeting the Target Transport Performance with infrastructure that just barely passes the derated tests.
- The validation process for an FSTIDS itself MUST be documented in such a way that other researchers can duplicate the validation experiments.
Except as noted, all tests below assume no derating. Tests for which there is not currently a well-established model for the required parameters explicitly include derating as a way to indicate flexibility in the parameters. 🌈
5.4. Test Preconditions 🛡️
Many tests have preconditions that are required to assure their validity. Examples include the presence or non-presence of cross traffic on specific subpaths; negotiating ECN; and a test stream preamble of appropriate length to achieve stable access to network resources in the presence of reactive network elements (as defined in Section 1.1 of [RFC7312]). If preconditions are not properly satisfied for some reason, the tests SHOULD be considered to be inconclusive. In general, it is useful to preserve diagnostic information as to why the preconditions were not met and any test data that was collected even if it is not useful for the intended test. Such diagnostic information and partial test data may be useful for improving the test or test procedures themselves. 🔍
It is important to preserve the record that a test was scheduled; otherwise, precondition enforcement mechanisms can introduce sampling bias. For example, canceling tests due to cross traffic on subscriber access links might introduce sampling bias in tests of the rest of the network by reducing the number of tests during peak network load. 📉
Test preconditions and failure actions MUST be specified in an FSTIDS. 💜
6. Generating Test Streams ✨
Okay bestie, so the big reason so many cool MBM properties work (vantage independence, I'm looking at you 👀) is that the test streams have timing patterns that copy how TCP or other transport protocols behave over a complete path 🦋 Like we covered in Section 4.1, self-clocked protocols naturally come out in bursts tied to the RTT and pipe size of the complete path. Those bursts get bigger (more packets per burst) when the Target RTT or Target Data Rate goes up, or when the Target MTU goes down 📈 So the lowkey tea is that test streams from self-clocked protocols run over short subpaths might not stress the queue at a bottleneck enough to tell whether the subpath can hold the full Target Transport Performance over the complete path 😭
Failing to copy TCP's temporal structure authentically is part of why simple performance tools like iPerf, netperf, nc, etc. have the reputation for giving false pass results over short test paths, even when a subpath has a flaw 💀 Iykyk.
The definitions in Section 3 are sufficient for most test streams 🌸 We're going to describe the slowstart and standing queue test streams in more detail, because they need extra love.
In conventional measurement practice, stochastic processes are used to get rid of lots of unintended correlations and sample biases 🎲 MBM tests, though, are designed to explicitly mimic the temporal correlations caused by network or protocol elements themselves. Some parts of these systems, like traffic arrival (e.g., test scheduling), are naturally stochastic ✨ Other behaviors, like back-to-back packet transmissions, are dominated by implementation-specific deterministic effects. Sure, these behaviors always have non-deterministic elements and might be modeled stochastically, but those details typically don't contribute significantly to the overall system behavior 🙄 Plus, it's known that real protocols can fail when network property estimators suffer from bias due to correlation in their own traffic. For example, TCP's RTT estimator used to determine the Retransmit Timeout (RTO) can be fooled by periodic cross traffic or start-stop applications 🫠 So for these reasons, many details of the test streams are specified deterministically 🔒
It may prove useful to introduce fine-grained noise sources into the models used for generating test streams in an update of Model-Based Metrics, but the complexity isn't warranted at the time this document was written 🧚
6.1. Mimicking Slowstart
TCP slowstart has a two-level burst structure, as shown in Figure 2 🍓 The fine time structure is caused by efficiency algorithms that deliberately batch work (CPU, channel allocation, etc.) to better amortize certain network and host overheads. ACKs passing through the return path typically cause the sender to transmit small bursts of data at the full sender interface rate. For example, TCP Segmentation Offload (TSO) and Delayed Acknowledgment both contribute to this effect 🔥 During slowstart, these bursts are at the same headway as the returning ACKs, but they're typically twice as large (e.g., have twice as much data) as the data the ACK reported was delivered to the receiver. Due to variations in delayed ACK and algorithms such as Appropriate Byte Counting [RFC3465], different pairs of senders and receivers produce slightly different burst patterns 💅 Without loss of generality, we assume each ACK causes four packet sender interface rate bursts at an average headway equal to the ACK headway. This corresponds to sending at an average rate equal to twice the effective bottleneck IP rate ✨ Each slowstart burst consists of a series of four packet sender interface rate bursts, such that the total number of packets is the current window size (as of the last packet in the burst).
The coarse time structure is due to each RTT being a reflection of the prior RTT 🌈 For real transport protocols, each slowstart burst is twice as large (twice the window) as the previous burst, but it's spread out in time by the network bottleneck, so each successive RTT exhibits the same effective bottleneck IP rate. The slowstart phase ends on the first lost packet or ECN mark, which is intended to happen after successive slowstart bursts merge in time: the next burst starts before the bottleneck queue is fully drained and the prior burst is complete 😤
For the diagnostic tests described below, we preserve the fine time structure but manipulate the coarse structure of the slowstart bursts (burst size and headway) to measure the ability of the dominant bottleneck to absorb and smooth slowstart bursts 🎀
Note that a stream of repeated slowstart bursts has three different average rates, depending on the averaging time interval 🤯 At the finest timescale (a few packet times at the sender interface), the peak of the average IP rate is the same as the sender interface rate. At a medium timescale (a few ACK times at the dominant bottleneck), the peak of the average IP rate is twice the implied bottleneck IP capacity. And at timescales longer than the target_RTT, and when the burst size is equal to the target_window_size, the average rate is equal to the target_data_rate 🦋 This pattern corresponds to repeating the last RTT of TCP slowstart when delayed ACK and sender-side byte counting are present, but without the limits specified in Appropriate Byte Counting [RFC3465].
time ==> ( - equals one packet)
Fine time structure of the packet stream:
---- ---- ---- ---- ----
|<>| sender interface rate bursts (typically 3 or 4 packets)
|<===>| burst headway (from the ACK headway)
\____repeating sender______/
rate bursts
Coarse (RTT-level) time structure of the packet stream:
---- ---- ---- ---- ---- ---- ---- ...
|<========================>| slowstart burst size (from the window)
|<==============================================>| slowstart headway
(from the RTT)
\__________________________/ \_________ ...
one slowstart burst Repeated slowstart bursts
Figure 2: Multiple Levels of Slowstart Bursts6.2. Constant Window Pseudo CBR
Pseudo constant bit rate (CBR) is implemented by running a standard self-clocked protocol, such as TCP, with a fixed window size 🎯 If that window size is test_window, the data rate will be slightly above the target_rate.
Since the test_window is constrained to be an integer number of packets, for small RTTs or low data rates there may not be sufficiently precise control over the data rate 😵💫 Rounding the test_window up (as defined above) is likely to result in data rates that are higher than the target rate, but reducing the window by one packet may result in data rates that are too small. Also, cross traffic can raise the RTT, which implicitly reduces the rate. Cross traffic that raises the RTT nearly always makes the test more strenuous (i.e., more demanding for the network path) 💪
Note that Constant Window Pseudo CBR (and Scanned Window Pseudo CBR in the next section) both rely on a self-clock that's at least partially derived from the properties of the subnet under test 🫣 This opens the door to the subnet under test showing behaviors, like extreme RTT fluctuations, that stop these algorithms from accurately controlling data rates.
An FSTIDS specifying a Constant Window Pseudo CBR test must explicitly indicate under what conditions errors in the data rate cause tests to be inconclusive 📝 Conventional paced measurement traffic may be more appropriate for these environments 🌸
6.3. Scanned Window Pseudo CBR
Scanned Window Pseudo CBR is similar to the Constant Window Pseudo CBR described above, except the window is scanned across a range of sizes designed to include two key events: the onset of queuing and the onset of packet loss or ECN CE marks 🔍 The window is scanned by incrementing it by one packet every 2*target_window_size delivered packets. This mimics the additive increase phase of standard Reno TCP congestion avoidance when delayed ACKs are in effect. Normally, the window increases are separated by intervals slightly longer than twice the target_RTT ⏱️
There are two ways to implement this test: 1) applying a window clamp to standard congestion control in a standard protocol such as TCP, and 2) stiffening a non-standard transport protocol 🛠️ When standard congestion control is in effect, any losses or ECN CE marks cause the transport to revert to a window smaller than the clamp, so the scanning clamp loses control of the window size. The NPAD (Network Path and Application Diagnostics) pathdiag tool is an example of this class of algorithms [Pathdiag] 🧪
Alternatively, a non-standard congestion control algorithm can respond to losses by transmitting extra data, so it maintains the specified window size independent of losses or ECN CE marks 😈 Such a stiffened transport explicitly violates mandatory Internet congestion control [RFC5681], and it's not suitable for in situ testing 🚫 It's only appropriate for engineering testing under laboratory conditions. The Windowed Ping tool implements such a test [WPING]. This tool has been updated (see [mpingSource]) 🎉
The test procedures in Section 8.2 describe how to partition the scans into regions and how to interpret the results 🦋
6.4. Concurrent or Channelized Testing 🧚
okay so this whole doc is about single-stream measurement, like one TCP connection or one measurement stream 💖. Ideally we'd ban every performance claim built on multiple concurrent streams, no cap. But that's not practical, for at least two reasons ✨.
First, a lot of super high-rate link technologies are channelized, and they at least partially pin each flow to a channel so packets don't get reordered within a flow 🦋.
Second, TCP itself has scaling limits 😭. The first problem might be solved with different design choices, but this one is way more deeply rooted.
Every congestion control algorithm that's philosophically aligned with [RFC5681] (so anything claiming TCP compatibility, friendliness, or fairness) has scaling limits. As a long fat network (LFN) with a fixed RTT and MTU gets faster, these algorithms get less accurate and have trouble filling the network [CCscaling] 💀. That's a consequence of the original Reno AIMD congestion control design, plus the requirement in [RFC5681] that all transport protocols have similar responses to congestion 🫶.
There are a bunch of reasons you'd want to specify performance across multiple concurrent flows 🎀. However, this approach is not recommended for data rates below several megabits per second, which can be reached with run lengths under 10000 packets on lots of paths. Since the required run length is proportional to the square of the data rate, at higher rates the run lengths get unreasonably large, and multiple flows might be the only feasible option 🌈.
If multiple flows are deemed necessary to meet aggregate performance targets, then this must be stated both in the design of the TIDS and in any claims about network performance 📣. The IP diagnostic tests must be performed concurrently with the specified number of connections. For tests that use bursty test streams, the bursts should be synchronized across streams, unless there is a priori knowledge that the applications have some explicit mechanism to stagger their own bursts. Without an explicit mechanism to stagger bursts, lots of network and application artifacts will sometimes implicitly synchronize bursts anyway 😭. A test that doesn't control burst synchronization may be prone to false pass results for some applications 🙈.
7. Interpreting the Results ✨
7.1. Test Outcomes 💅
To perform an exhaustive test of a complete network path, each test of the TIDS gets applied to each subpath of the complete path 🛣️. If any subpath fails any test, then a standard transport protocol running over the complete path can also be expected to fail to attain the Target Transport Performance under some conditions 💀.
In addition to passing or failing, a test can be deemed inconclusive for a number of reasons 🤷. Proper instrumentation and treatment of inconclusive outcomes is critical to the accuracy and robustness of Model-Based Metrics 🔥. Tests can be inconclusive if the precomputed traffic pattern or data rates weren't accurately generated, if the measurement results weren't statistically significant, if the required preconditions for the test weren't met, or for other causes. See Section 5.4.
For example, consider a test that implements Constant Window Pseudo CBR (Section 6.2) by adding rate controls and detailed IP packet transfer instrumentation to TCP (e.g., using the extended performance statistics for TCP described in [RFC4898]). TCP includes built-in control systems that might interfere with the sending data rate 🧠. If such a test meets the required packet transfer statistics (e.g., run length) while failing to attain the specified data rate, it must be treated as an inconclusive result 🫠. We can't a priori determine whether the reduced data rate was caused by a TCP problem or a network problem, or whether the reduced data rate had a material effect on the observed packet transfer statistics.
Note that for capacity tests, if the observed packet transfer statistics meet the statistical criteria for failing (based on acceptance of hypothesis H1 in Section 7.2), 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 important new properties of MBM, such as vantage independence, are a direct consequence of opening the control loops in the protocols, so the test stream doesn't depend on network conditions or IP packets received 🦋. Any mechanism that introduces feedback between the path's measurements and the test stream generation is at risk of introducing nonlinearities that spoil these properties 😤. Any exceptional event that indicates that such feedback has happened should cause the test to be considered inconclusive 🚩.
Inconclusive tests may be caused by situations in which a test outcome is ambiguous because of network limitations or an unknown limitation on the IP diagnostic test itself, which may have been caused by some uncontrolled feedback from the network 🌀.
Note that procedures that attempt to search the target parameter space to find the limits on a parameter such as target_data_rate are at risk of breaking the location-independent properties of Model-Based Metrics if any part of the boundary between passing, inconclusive, or failing results is sensitive to RTT (which is normally the case) 🫣. For example, the maximum data rate for a marginal link (e.g., exhibiting excess errors) is likely to be sensitive to the test_path_RTT. The maximum observed data rate over the test path has very little value for predicting the maximum rate over a different path 🗺️.
One of the goals for evolving TIDS designs will be to keep sharpening the distinctions between inconclusive, passing, and failing tests ✨. The criteria for inconclusive, passing, and failing tests must be explicitly stated for every test in the TIDS or FSTIDS 📋.
One of the goals for evolving the testing process, procedures, tools, and measurement point selection should be to minimize the number of inconclusive tests 🎯.
It may be useful to keep raw packet transfer statistics and ancillary metrics [RFC3148] for deeper study of the behavior of the network path and to measure the tools themselves 🔍. Raw packet transfer statistics can help drive tool evolution. Under some conditions, it might be possible to re-evaluate the raw data for satisfying alternate Target Transport Performance. However, it is important to guard against sampling bias and other implicit feedback that can cause false results and exhibit measurement point vantage sensitivity 😬. Simply applying different delivery criteria based on a different Target Transport Performance is insufficient if the test traffic patterns (bursts, etc.) do not match the alternate Target Transport Performance 💅.
7.2. Statistical Criteria for Estimating run_length 📊
When evaluating the observed run_length, we need to determine appropriate packet stream sizes and acceptable error levels for efficient measurement. In practice, can we compare the empirically estimated packet loss and ECN CE marking ratios 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 (packet loss ratio, other metric, or any marking we define) 🔁.
As each packet is sent and measured, we have an ongoing estimate of the performance in terms of the ratio of packet loss or ECN CE marks to total packets (i.e., an empirical probability). We continue to send until conditions support a conclusion or a maximum sending limit has been reached ⏳.
We have a target_mark_probability, one mark per target_run_length, where a "mark" is defined as a lost packet, a packet with ECN CE mark, or other signal. This constitutes the null hypothesis:
H0: no more than one mark in target_run_length =
3*(target_window_size)^2 packetsWe can stop sending packets if ongoing measurements support accepting H0 with the specified Type I error = alpha (= 0.05, for example).
We also have an alternative hypothesis to evaluate: is performance significantly lower than the target_mark_probability? Based on analysis of typical values and practical limits on measurement duration, we choose four times the H0 probability:
H1: one or more marks in (target_run_length/4) packetsand 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 this document. While the ongoing measurements don't support either hypothesis, the current status of measurements is inconclusive 🌫️.
The problem above is formulated to match the Sequential Probability Ratio Test (SPRT) [Wald45] [Montgomery90]. Note that as originally framed, the events under consideration were all manufacturing defects. In networking, ECN CE marks and lost packets are not defects but signals, indicating that the transport protocol should slow down 🐢.
The Sequential Probability Ratio Test also starts with a pair of hypotheses specified as above:
H0: p0 = one defect in target_run_length
H1: p1 = one defect in target_run_length/4As 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 + s*n
Rejection line: Xr = h2 + s*nwhere 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) } ]/kfor p0 and p1 as defined in the null and alternative hypotheses statements above, and alpha and beta as the Type I and Type II errors.
The SPRT specifies simple stopping rules:
Xa < defect_count(n) < Xr: continue testingdefect_count(n) <= Xa: Accept H0defect_count(n) >= Xr: Accept H1
The calculations above are implemented in the R-tool for Statistical Analysis [Rtool], in the add-on package for Cross-Validation via Sequential Testing (CVST) [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 + s*n
and n = h1 / sNote that the derivations in [Wald45] and [Montgomery90] differ. Montgomery's simplified derivation of SPRT may assume Bernoulli processes, where the packet loss probabilities are independent and identically distributed, which makes the SPRT more accessible. Wald's seminal paper showed that this assumption is not necessary. It helps to remember that the goal of SPRT is not to estimate the value of the packet loss rate, but only whether or not the packet loss ratio is likely (1) low enough (when we accept the H0 null hypothesis), yielding success, or (2) too high (when we accept the H1 alternate hypothesis), yielding failure 🎯.
7.3. Reordering Tolerance 🦋
okay so every test must be instrumented for packet-level reordering [RFC4737] 🙈 but here's the tea: there is no consensus on how much reordering is acceptable 💀. Over the last two decades the general vibe has been making protocols and apps more tolerant to reordering (see [RFC5827]) 💅, because the network itself has been getting more reordering over time. That's mostly thanks to tech like multithreaded routing lookups and Equal-Cost Multipath (ECMP) routing ✨. These techniques add parallelism to the network, and they're critical to letting overall Internet growth outpace Moore's Law 🚀. Iconic, honestly.
With transport retransmission strategies, there are real trade-offs among reordering tolerance, how quickly losses can be repaired, and overhead from spurious retransmissions 🫠. Before any new retransmission strategies show up, we're proposing a strawman 🍓: transport protocols should be able to adapt to reordering as long as the reordering extent is not more than the maximum of one quarter window or 1 ms, whichever is larger. (These values come from experience prototyping Early Retransmit [RFC5827] and related algorithms. They agree with the values being proposed for "RACK: a time-based fast loss detection algorithm" [RACK].) Within this limit on reorder extent, there should be no bound on reordering density 🌈.
So by implication, reordering that is less than these bounds should not be treated as a network impairment 🫶. But [RFC4737] still applies! Reordering should be instrumented, and the maximum reordering that can be properly characterized by the test (because of the bound on history buffers) should be recorded with the measurement results 📝.
Reordering tolerance and diagnostic limitations, such as the size of the history buffer used to diagnose packets that are way out of order, must be specified in an FSTIDS 💜.
8. IP Diagnostic Tests 🎀
The IP diagnostic tests below are organized according to the technique used to generate the test stream, as described in Section 6. 🧚 All of the results are evaluated in accordance with Section 7, possibly with additional test-specific criteria.
We also introduce some combined tests that are more efficient when networks are expected to pass, but they conflate diagnostic signatures when they fail 🌸.
8.1. Basic Data Rate and Packet Transfer Tests 🍓
We propose several versions of the basic data rate and packet transfer statistics test, and they differ in how the data rate is controlled. The data can be paced on a timer, or it can be window controlled (and self-clocked) ⏰. The first two tests implicitly confirm that sub_path has sufficient raw capacity to carry the target_data_rate. They're recommended for relatively infrequent testing, such as an installation or a periodic auditing process 📋. The third test, Background Packet Transfer Statistics, is a low-rate test designed for ongoing monitoring for changes in subpath quality 👀.
8.1.1. Delivery Statistics at Paced Full Data Rate 🚀
This test confirms that the observed run length is at least the target_run_length, while relying on timer to send data at the target_rate using the procedure described in Section 6.1 with a burst size of 1 (single packets) or 2 (packet pairs) 💫.
The test is considered to be inconclusive if the packet transmission cannot be accurately controlled for any reason 😵💫.
RFC 6673 [RFC6673] is appropriate for measuring packet transfer statistics at full data rate 💖.
8.1.2. Delivery Statistics at Full Data Windowed Rate 🪟
This test confirms that the observed run length is at least the target_run_length while sending at an average rate approximately equal to the target_data_rate. It does this by controlling (or clamping) the window size of a conventional transport protocol to test_window 🎯.
Since losses and ECN CE marks cause transport protocols to reduce their data rates, this test is expected to be less precise about controlling its data rate. It should not be considered inconclusive as long as at least some of the round trips reached the full target_data_rate without incurring losses or ECN CE marks 🌈. To pass this test, the network must deliver target_window_size packets in target_RTT time without any losses or ECN CE marks at least once per two target_window_size round trips, in addition to meeting the run length statistical test 👑.
8.1.3. Background Packet Transfer Statistics Tests 🌙
The Background Packet Transfer Statistics Test is a low-rate version of the target rate test above. It's designed for ongoing, lightweight monitoring for changes in the observed subpath run length without disrupting users 🫧. It should be used in conjunction with one of the above full-rate tests, because it does not confirm that the subpath can support raw data rate. No cap, that's a whole separate job 💅.
RFC 6673 [RFC6673] is appropriate for measuring background packet transfer statistics 🦋.
8.2. Standing Queue Tests 🌊
These engineering tests confirm that the bottleneck is well behaved across the onset of packet loss, which typically follows after the onset of queuing 🧐. Well behaved generally means lossless for transient queues. But once the queue has been sustained for a sufficient period of time (or reaches a sufficient queue depth), there should be a small number of losses or ECN CE marks to signal to the transport protocol that it should reduce its window or data rate 🛑.
Losses that are too early can prevent the transport from averaging at the target_data_rate 😩. Losses that are too late indicate that the queue might not have an appropriate AQM [RFC7567] and, as a consequence, be subject to bufferbloat [wikiBloat]. Queues without AQM have the potential to inflict excess delays on all flows sharing the bottleneck 🐢. Excess losses (more than half of the window) at the onset of loss make loss recovery problematic for the transport protocol 🥴. Non-linear, erratic, or excessive RTT increases suggest poor interactions between the channel acquisition algorithms and the transport self-clock 📉.
All of the tests in this section use the same basic scanning algorithm, described here, but they score the link or subpath on how well it avoids each of these problems 🔍.
Some network technologies rely on virtual queues or other techniques to meter traffic without adding any queuing delay. In that case the data rate will vary with the window size all the way up to the onset of load-induced packet loss or ECN CE marks 🪄. For these technologies, the discussion of queuing in Section 6.3 does not apply, but it is still necessary to confirm that the onset of losses or ECN CE marks is at an appropriate point and is progressive. If the network bottleneck does not introduce significant queuing delay, modify the procedure described in Section 6.3 to start the scan at a window equal to or slightly smaller than the test_window 🎀.
Use the procedure in Section 6.3 to sweep the window across the onset of queuing and the onset of loss 🧹. The tests below all assume that the scan emulates standard additive increase and delayed ACK, by incrementing the window by one packet for every 2*target_window_size packets delivered. A scan can typically be divided into three regions: below the onset of queuing, a standing queue, and at or beyond the onset of loss 🍰.
Below the onset of queuing, the RTT is typically fairly constant, and the data rate varies in proportion to the window size 📈. Once the data rate reaches the subpath IP rate, the data rate becomes fairly constant, and the RTT increases in proportion to the increase in window size 🧃. The precise transition across the start of queuing can be identified by the maximum network power, which is defined to be the ratio of data rate over the RTT. The network power can be computed at each window size, and the window with the maximum is taken as the start of the queuing region ✨.
If there is random background loss (e.g., bit errors), precise determination of the onset of queue-induced packet loss may require multiple scans 🔁. At window sizes large enough to cause loss in queues, all transport protocols are expected to experience periodic losses. Those are determined by the interaction between the congestion control and AQM algorithms 🤝. For standard congestion control algorithms, the periodic losses are likely to be relatively widely spaced, and the details are typically dominated by the behavior of the transport protocol itself 🌷. For the case of stiffened transport protocols (with non-standard, aggressive congestion control algorithms), the details of periodic losses will be dominated by how the window increase function responds to loss 😤.
8.2.1. Congestion Avoidance 🕊️
A subpath passes the congestion avoidance standing queue test if more than target_run_length packets are delivered between the onset of queuing (as determined by the window with the maximum network power, as described above) and the first loss or ECN CE mark 🥇. If this test is implemented using a standard congestion control algorithm with a clamp, it can be performed in situ in the production Internet as a capacity test. For an example of such a test, see [Pathdiag] 🧪.
For technologies that do not have conventional queues, use the test_window in place of the onset of queuing. That is, a subpath passes the congestion avoidance standing queue test if more than target_run_length packets are delivered between the start of the scan at test_window and the first loss or ECN CE mark 🫧.
8.2.2. Bufferbloat 🫧
This test confirms that there is some mechanism to limit buffer occupancy (e.g., one that prevents bufferbloat) 🧼. Note that this is not strictly a requirement for single-stream bulk transport capacity. However, if there is no mechanism to limit buffer queue occupancy, then a single stream with sufficient data to deliver is likely to cause the problems described in [RFC7567] and [wikiBloat]. This may cause only minor symptoms for the dominant flow, but it has the potential to make the subpath unusable for other flows and applications 😭.
The test will pass if the onset of loss occurs before a standing queue has introduced delay greater than twice the target_RTT, or another well-defined and specified limit 💅. Note that there is not yet a model for how much standing queue is acceptable. The factor of two chosen here reflects a rule of thumb 🤷♀️. In conjunction with the previous test, this test implies that the first loss should occur at a queuing delay that is between one and two times the target_RTT 💜.
Specified RTT limits that are larger than twice the target_RTT must be fully justified in the FSTIDS 📝🌈.
8.2.3. Non-excessive Loss 💅
okay so this test is all about making sure the loss doesn't start getting out of hand 🙈 ✨ The test passes if the losses are equal to or less than the increase in cross traffic plus the test stream window increase since the previous RTT. Another way to say it: the total throughput of the subpath shouldn't go down when loss first starts 💜 (Note that if there's a quick dip in subpath throughput and there isn't already a standing queue, a subpath that passes the other queue tests in this document will have enough queue space to hold one full RTT's worth of data 🌸)
Token bucket policers will not pass this test, and that's totally intentional 🙅♀️💀 TCP often faceplants when more than a small fraction of packets get dropped in one RTT 😭 Lots of TCP implementations will need a timeout and slowstart just to recover their self-clock. Even when they do recover from the massive losses, the sudden change in available capacity at the bottleneck wastes serving and front-path capacity until TCP can adjust to the new rate [Policing] 🫠
8.2.4. Duplex Self-Interference 🦋
this one is an engineering test that checks the limits on how the forward data path and the ACK return path interact when they share a half-duplex link 🔥
some older half-duplex technologies had the vibe where each direction held the channel until it fully drained its queue 😩 When a self-clocked transport protocol like TCP has data and ACKs going in opposite directions through a link like that, the behavior often falls back to stop-and-wait 💀 Every extra packet added to the window bumps the observed RTT up by two packet times: once for the extra packet passing through the data path, and once for the extra delay the ACK spends waiting on the return path 🙃
The Duplex Self-Interference Test fails if the RTT goes up by more than a fixed bound above the expected queuing time. That expected queuing time is computed as the excess window divided by the subpath IP capacity 📊 This bound must be smaller than target_RTT/2 to avoid falling back to stop-and-wait behavior (e.g., data packets and ACKs both have to be released at least twice per RTT) ✨
8.3. Slowstart Tests 🍓
these tests copy slowstart: data gets sent at twice the effective bottleneck rate so the queue at the dominant bottleneck gets a real workout 🏋️♀️💖
8.3.1. Full Window Slowstart Test 👑
this is a capacity test that checks whether slowstart is unlikely to exit too early 🧚 To run it, send slowstart bursts that are target_window_size total packets, and collect packet transfer statistics as described in Section 7.2 to score the outcome. The test passes if it's statistically significant that the observed number of good packets delivered between losses or ECN CE marks is larger than the target_run_length. The test fails if it's statistically significant that the observed interval between losses or ECN CE marks is smaller than the target_run_length 😭
The test is inconclusive if the time to send the data burst is not less than half of the time to receive the ACKs. (So it's fine to send data too fast, but sending it slower than twice the actual bottleneck rate, as the ACKs show it, counts as inconclusive 🤷♀️) The headway for the slowstart bursts should be the target_RTT. 💅
Note that these are the same parameters used for the Sustained Full-Rate Bursts Test, except the burst rate is at slowstart rate instead of the sender interface rate 🦋
8.3.2. Slowstart AQM Test 🌈
to run this one, do a continuous slowstart (send data nonstop at twice the implied IP bottleneck capacity) until the first loss 🌊 Then stop, let the network drain, and repeat. Gather statistics on how many packets got delivered before the loss, the pattern of losses, the maximum observed RTT, and the window size, and then justify your results 📝 There isn't currently enough theory to justify requiring any particular result 🤓 But design decisions that affect the outcome of this test also affect how the network balances between long and short flows (the "mice vs. elephants" problem) 🐭🐘 The queue sojourn time for the first packet delivered after the first loss should be at least one half of the target_RTT. 🫶
This engineering test should be run on a quiescent network or testbed, since cross traffic can change the results in ill-defined ways 🙈
8.4. Sender Rate Burst Tests 🔥
these tests figure out how well the network can deliver bursts sent at the sender's interface rate 💥 Heads up: this test really exercises the front path and might include infrastructure that's out of scope for an access ISP, even though the bursts could be caused by ACK compression, thinning, or channel arbitration in the access ISP itself. See Appendix B. 👀
there are also a bunch of details about sender interface rate bursts that aren't fully defined here 😬 Stuff like the assumed sender interface rate should be explicitly stated in an FSTIDS. 📋
Current standards let TCP send full window bursts after an application pause. (Congestion Window Validation [RFC2861] and updates to support Rate-Limited Traffic [RFC7661] are not required.) Since full window bursts match standard behavior, it's desirable for the network to be able to deliver them. Otherwise, application pauses will cause unwarranted losses 😤 Also, the AIMD sawtooth needs a peak window that's twice the target_window_size, so the worst-case burst may be 2*target_window_size 🧠
The application and serving community also gets that interface rate bursts cost the network something, and that cost has to be balanced against other costs inside the servers themselves 🖥️ For example, TCP Segmentation Offload (TSO) cuts server CPU in exchange for larger network bursts, which crank up the stress on network buffer memory 😵 Some newer TCP implementations can pace traffic at scale [TSO_pacing] [TSO_fq_pacing]. It's still TBD if and how quickly these changes get deployed ⏳
There's no theory yet that unifies these costs or gives a framework for optimizing global efficiency 🫠 We also don't have a model for how many server rate bursts the network should tolerate. Some bursts have to be tolerated by the network, but it's probably unreasonable to expect the network to efficiently deliver all data as a series of bursts 🌧️
So this is the only test where we encourage derating ✨ A TIDS could include a table with pairs of derating parameters: burst sizes, and how much each burst size is allowed to reduce the run length, relative to the target_run_length 📊💜
8.5. Combined and Implicit Tests 🧩
combined tests check multiple network properties all at once, in a single test, and they can even happen as a side effect of normal content delivery 🎀 They need less measurement traffic than other strategies, but when they fail, the diagnostic signatures get mixed together 🌀 These are by far the most efficient option for monitoring networks that are supposed to pass all tests 👑
8.5.1. Sustained Full-Rate Bursts Test 🔥
The Sustained Full-Rate Bursts Test is a combined, worst-case version of all the capacity tests above 💪 To run it, send target_window_size bursts of packets at the server interface rate, with a target_RTT burst headway (burst start to next burst start), and verify that the observed packet transfer statistics meet the target_run_length. ✨
Key observations:
- The subpath under test is expected to go idle for some fraction of the time. That fraction is determined by the difference between the time to drain the queue at the
subpath_IP_capacityand thetarget_RTT. If the queue doesn't drain completely, that may mean the subpath has insufficient IP capacity, or that there's some other problem with the test (e.g., it's inconclusive) 🤔 - The burst sensitivity can be derated by sending smaller bursts more frequently (e.g., by sending
target_window_size*deratepacket bursts everytarget_RTT*derate, where "derate" is less than one). 🍬 - When not derated, this test is the most strenuous capacity test 🏆
- A subpath that passes this test is likely to be able to sustain higher rates (close to
subpath_IP_capacity) for paths with RTTs significantly smaller than thetarget_RTT. 🚀 - This test can be implemented with instrumented TCP [RFC4898], using a specialized measurement application at one end (e.g., [MBMSource]) and a minimal service at the other end (e.g., [RFC863] and [RFC864]). 🛠️
- This test is efficient to implement, since it doesn't need per-packet timers, and it can make use of TSO in modern network interfaces. 💅
- If a subpath is known to pass the standing queue engineering tests (particularly that it has a progressive onset of loss at an appropriate queue depth), then the Sustained Full-Rate Bursts Test is sufficient to assure that the subpath under test won't impair Bulk Transport Capacity at the target performance under all conditions. See Section 8.2 for a discussion of the standing queue tests. 🫶
Note that this test is clearly independent of the subpath RTT or other details of the measurement infrastructure, as long as the measurement infrastructure can accurately and reliably deliver the required bursts to the subpath under test. 🌸
8.5.2. Passive Measurements ✨
Okay so any non-throughput-maximizing app 💅, like fixed-rate streaming media, can be used to run passive or hybrid (defined in [RFC7799]) versions of Model-Based Metrics 🦋. You just need some extra instrumentation, and maybe a traffic shaper or other controls on the servers. The big essential thing is that the data transmission has to be constrained, so that even with random app pauses and bursts, the data rate and burst sizes stay inside the envelope set by each individual test above 🫶
If the app's serving data rate can be kept at or below target_data_rate, and the serving_RTT (the RTT between the sender and client) is less than target_RTT, the easiest way to pull this off is clamping the transport window size to serving_window_clamp 🎀. That clamp is set to test_window and computed for the actual serving path.
Under those constraints, the serving_window_clamp will cap both the serving data rate and the burst sizes, so they're no larger than the parameters from Section 8.1.2, 8.4, or 8.5.1 🔥. Since the serving RTT is smaller than the target_RTT, the worst-case bursts that could show up are smaller than what Section 8.4 calls for. The sender rate burst sizes get implicitly derated by serving_window_clamp divided by target_window_size at the very least 💜. (Depending on the app's behavior, the data might be way smoother than anything the burst tests specify 🌸.)
In an alternative setup, the data rate and bursts could be controlled explicitly, either by a programmable traffic shaper or by pacing at the sender ✨. This gives you better control over transmissions, but it's more complicated to build, even though the required technology is available [TSO_pacing] [TSO_fq_pacing] 🧚.
Also, these techniques work for any content delivery that can run at a constrained data rate to inhibit TCP equilibrium behavior 🌈.
Now, Dynamic Adaptive Streaming over HTTP (DASH) is generally in conflict with passive Model-Based Metrics measurement, because it's a rate-maximizing protocol 💀. It can still meet the requirement here if the rate can be capped. For example, you could know a priori the maximum rate needed to deliver a particular piece of content 🙈.
9. Example 🍓
In this section, we look at a TIDS designed to confirm that an access ISP can reliably deliver HD video from multiple content providers to all of its customers 📺. With modern codecs, minimal HD video (720p) generally fits in 2.5 Mb/s. Because of the ISP's geographical size, network topology, and modem characteristics, the ISP figures most content is within a 50 ms RTT of its users 🗺️. (This example RTT is enough to cover the propagation delay to continental Europe or to either coast of the United States with low-delay modems. It covers somewhat smaller regions if the modems need extra delay for advanced compression and error recovery 🫠.)
+----------------------+-------+---------+
| End-to-End Parameter | value | units |
+----------------------+-------+---------+
| target_rate | 2.5 | Mb/s |
| target_RTT | 50 | ms |
| target_MTU | 1500 | bytes |
| header_overhead | 64 | bytes |
| | | |
| target_window_size | 11 | packets |
| target_run_length | 363 | packets |
+----------------------+-------+---------+
Table 1: 2.5 Mb/s over a 50 ms PathTable 1 shows the default TCP model with no derating, so it's pretty conservative 🥹. The simplest TIDS would be the Sustained Full-Rate Bursts Test, described in Section 8.5.1 🏃♀️. That test would send 11 packet bursts every 50 ms and confirm there's no more than 1 packet loss per 33 bursts (363 total packets in 1.650 seconds) 👑.
Since this number is the entire end-to-end loss budget, you can run independent subpath tests by splitting the packet loss ratio across subpaths 🧩. For example, 50% of the losses might go to the access or last mile link to the user, 40% to the network interconnections with other ISPs, and 1% to each internal hop (assuming no more than 10 internal hops). Then all of the subpaths can be tested independently, and the spatial composition of the passing subpaths would be expected to stay within the end-to-end loss budget 💖.
9.1. Observations about Applicability 🌷
Guidance on deploying and using MBM belongs in a future document 📚. Still, the example above shows some of the issues that might need to be considered 👀.
Another ISP with different geographical coverage, topology, or modem technology may need a different target_RTT. That means a different target_window_size and target_run_length, even for the same target_data_rate 🙃. One consequence is that shared infrastructure, like Internet Exchange Points (IXPs) and other interconnects, may need to be evaluated using the most stringent target_window_size and target_run_length of any participating ISP 🚦. One way to handle this is to choose target parameters for evaluating shared infrastructure based on a hypothetical reference path that doesn't necessarily match any actual path 🦄.
Testing interconnects has generally been a mess 😭. Conventional performance tests run between measurement points adjacent to either side of the interconnect usually aren't useful. Unconstrained TCP tests, like iPerf [iPerf], are usually overly aggressive because the RTT is so small (often less than 1 ms) 🏎️. With a short RTT, these tools can report inflated data rates, since they can tolerate very high packet loss ratios and can push other cross traffic off the network 🥴. So these measurements are useless for predicting actual user performance over longer paths, and they may be pretty disruptive themselves. Model-Based Metrics solves this problem ✨. The interconnect can be evaluated with the same TIDS as the other subpaths. Continuing our example, if the interconnect gets 40% of the losses, then 11 packet bursts sent every 50 ms should have fewer than one loss per 82 bursts (902 packets) 🎉.
10. Validation 🫧
Some aspects of the models are likely to be too conservative. So Section 5.2 allows alternate protocol models, and Section 5.3 allows test parameter derating 🧁. If either of these techniques is used, we need demonstrations that the TIDS can reliably detect subpaths that would stop authentic applications running state-of-the-art protocol implementations from meeting the specified Target Transport Performance 🔍. This correctness criteria is potentially hard to prove, because it implicitly requires validating a TIDS against every possible path and subpath 😵💫. The procedures described here are still experimental 🧪.
We suggest two approaches, and both should be applied 💅. First, publish a fully open description of the TIDS. That includes what assumptions were used and how it was derived, so the research community can evaluate the design decisions, test them, and comment on their applicability 📢. Second, demonstrate that applications do meet the Target Transport Performance when running over a network testbed with the tightest possible constraints that still let the TIDS tests pass 🏋️♀️.
This procedure is kind of like an epsilon-delta proof in calculus 🤓. You build a test network where all of the individual TIDS tests pass by only small (infinitesimal) margins. Then you show that a variety of authentic applications, running over real TCP implementations (or other protocols, as appropriate), meet the Target Transport Performance over that network 🌟. The workloads should include multiple types of streaming media and transaction-oriented short flows (e.g., synthetic web traffic) 🛜.
For example, for the HD streaming video TIDS described in Section 9, the IP capacity should be exactly header_overhead above 2.5 Mb/s 📏. The per packet random background loss ratio should be 1/363 (for a run length of 363 packets). The bottleneck queue should be 11 packets, and the front path should have just enough buffering to withstand 11 packet interface rate bursts 🪣. We want every one of the TIDS tests to fail if we slightly increase the relevant test parameter. So, for example, sending a 12-packet burst should cause excess (possibly deterministic) packet drops at the dominant queue at the bottleneck 💥. This network has the tightest possible constraints that can be expected to pass the TIDS, yet a real application using a stock TCP implementation in the vendor's default configuration should be able to attain 2.5 Mb/s over a 50 ms path 🥳.
The hardest part of setting this up is arranging for the testbed to have the tightest possible constraints that still let it pass the individual tests 😤. Two approaches are suggested:
- constraining (configuring) the network devices so they don't use all available resources (e.g., by limiting available buffer space or data rate)
- pre-loading subpaths with cross traffic
It's important that a single tightly constrained environment just barely passes all the tests 🎯. Otherwise, there's a chance TCP can exploit extra latitude in some parameters (such as data rate) to partially make up for constraints in other parameters (e.g., queue space) 🙅♀️. This effect can go both ways: extra latitude in the queue space tests can let TCP compensate for insufficient data-rate headroom 🔄.
To the extent a TIDS is used to inform public dialog, it should be fully documented publicly. That includes the details of the tests, what assumptions were used, and how it was derived 📣. All of the details of the validation experiment should also be published with enough detail for other researchers to replicate it 🔬. All components should be either open source or fully described proprietary implementations that are available to the research community 🫶
11. Security Considerations 🔒
Okay bestie, this part is SERIOUS 😭 because measurement is often used to make business and policy decisions, which means it can get manipulated. Model-Based Metrics are expected to be a huge step forward because equivalent measurements can be run from multiple vantage points. That means performance claims can be independently validated by multiple parties. Iconic 💅
Honestly, a lot of the drama in the Net Neutrality debate comes from the fact that there were no good vantage-independent tools to characterize network performance. Traditional ways of measuring Bulk Transport Capacity are super sensitive to RTT 🙈 So they often give totally different results when you run them locally to an ISP or interconnect versus over a customer's complete path. Neither the ISP nor the customer can repeat the other's measurements, and that leads to major distrust and acrimony. Model-Based Metrics are expected to greatly improve this situation, and I'm obsessed with that for them 🫶
Heads up: in situ measurements sometimes need to send synthetic measurement traffic between arbitrary locations in the network. That makes them potentially attractive platforms for launching DDoS attacks 💀 All active measurement tools and protocols must be designed to minimize the opportunities for these misuses. See the discussion in Section 7 of [RFC7594].
Some of the tests in this document are not intended for frequent network monitoring 🌸 They have the potential to cause high network loads and might adversely affect other traffic, so please don't run them on repeat like it's your fave song 😅
Also, this document only describes a framework for designing a Fully Specified Targeted IP Diagnostic Suite (FSTIDS) ✨ Each FSTIDS must include its own security section. So every suite needs its own security glow-up 💖
12. IANA Considerations 🧚
Good news, this one's quick: this document has no IANA actions 🎀
13. Informative References 📚
Here's the full reading list, bestie. Every single one of these is cited, and I've kept every detail exactly as it was 🌈
[RFC863] Postel, J., "Discard Protocol", STD 21, RFC 863,
DOI 10.17487/RFC0863, May 1983,
<https://www.rfc-editor.org/info/rfc863>.
[RFC864] Postel, J., "Character Generator Protocol", STD 22,
RFC 864, DOI 10.17487/RFC0864, May 1983,
<https://www.rfc-editor.org/info/rfc864>.
[RFC2330] Paxson, V., Almes, G., Mahdavi, J., and M. Mathis,
"Framework for IP Performance Metrics", RFC 2330,
DOI 10.17487/RFC2330, May 1998,
<https://www.rfc-editor.org/info/rfc2330>.
[RFC2861] Handley, M., Padhye, J., and S. Floyd, "TCP Congestion
Window Validation", RFC 2861, DOI 10.17487/RFC2861, June
2000, <https://www.rfc-editor.org/info/rfc2861>.
[RFC3148] Mathis, M. and M. Allman, "A Framework for Defining
Empirical Bulk Transfer Capacity Metrics", RFC 3148,
DOI 10.17487/RFC3148, July 2001,
<https://www.rfc-editor.org/info/rfc3148>.
[RFC3168] Ramakrishnan, K., Floyd, S., and D. Black, "The Addition
of Explicit Congestion Notification (ECN) to IP",
RFC 3168, DOI 10.17487/RFC3168, September 2001,
<https://www.rfc-editor.org/info/rfc3168>.
[RFC3465] Allman, M., "TCP Congestion Control with Appropriate Byte
Counting (ABC)", RFC 3465, DOI 10.17487/RFC3465, February
2003, <https://www.rfc-editor.org/info/rfc3465>.
[RFC4737] Morton, A., Ciavattone, L., Ramachandran, G., Shalunov,
S., and J. Perser, "Packet Reordering Metrics", RFC 4737,
DOI 10.17487/RFC4737, November 2006,
<https://www.rfc-editor.org/info/rfc4737>.
[RFC4898] Mathis, M., Heffner, J., and R. Raghunarayan, "TCP
Extended Statistics MIB", RFC 4898, DOI 10.17487/RFC4898,
May 2007, <https://www.rfc-editor.org/info/rfc4898>.
[RFC5136] Chimento, P. and J. Ishac, "Defining Network Capacity",
RFC 5136, DOI 10.17487/RFC5136, February 2008,
<https://www.rfc-editor.org/info/rfc5136>.
[RFC5681] Allman, M., Paxson, V., and E. Blanton, "TCP Congestion
Control", RFC 5681, DOI 10.17487/RFC5681, September 2009,
<https://www.rfc-editor.org/info/rfc5681>.
[RFC5827] Allman, M., Avrachenkov, K., Ayesta, U., Blanton, J., and
P. Hurtig, "Early Retransmit for TCP and Stream Control
Transmission Protocol (SCTP)", RFC 5827,
DOI 10.17487/RFC5827, May 2010,
<https://www.rfc-editor.org/info/rfc5827>.
[RFC5835] Morton, A., Ed. and S. Van den Berghe, Ed., "Framework for
Metric Composition", RFC 5835, DOI 10.17487/RFC5835, April
2010, <https://www.rfc-editor.org/info/rfc5835>.
[RFC6049] Morton, A. and E. Stephan, "Spatial Composition of
Metrics", RFC 6049, DOI 10.17487/RFC6049, January 2011,
<https://www.rfc-editor.org/info/rfc6049>.
[RFC6576] Geib, R., Ed., Morton, A., Fardid, R., and A. Steinmitz,
"IP Performance Metrics (IPPM) Standard Advancement
Testing", BCP 176, RFC 6576, DOI 10.17487/RFC6576, March
2012, <https://www.rfc-editor.org/info/rfc6576>.
[RFC6673] Morton, A., "Round-Trip Packet Loss Metrics", RFC 6673,
DOI 10.17487/RFC6673, August 2012,
<https://www.rfc-editor.org/info/rfc6673>.
[RFC6928] Chu, J., Dukkipati, N., Cheng, Y., and M. Mathis,
"Increasing TCP's Initial Window", RFC 6928,
DOI 10.17487/RFC6928, April 2013,
<https://www.rfc-editor.org/info/rfc6928>.
[RFC7312] Fabini, J. and A. Morton, "Advanced Stream and Sampling
Framework for IP Performance Metrics (IPPM)", RFC 7312,
DOI 10.17487/RFC7312, August 2014,
<https://www.rfc-editor.org/info/rfc7312>.
[RFC7398] Bagnulo, M., Burbridge, T., Crawford, S., Eardley, P., and
A. Morton, "A Reference Path and Measurement Points for
Large-Scale Measurement of Broadband Performance",
RFC 7398, DOI 10.17487/RFC7398, February 2015,
<https://www.rfc-editor.org/info/rfc7398>.
[RFC7567] Baker, F., Ed. and G. Fairhurst, Ed., "IETF
Recommendations Regarding Active Queue Management",
BCP 197, RFC 7567, DOI 10.17487/RFC7567, July 2015,
<https://www.rfc-editor.org/info/rfc7567>.
[RFC7594] Eardley, P., Morton, A., Bagnulo, M., Burbridge, T.,
Aitken, P., and A. Akhter, "A Framework for Large-Scale
Measurement of Broadband Performance (LMAP)", RFC 7594,
DOI 10.17487/RFC7594, September 2015,
<https://www.rfc-editor.org/info/rfc7594>.
[RFC7661] Fairhurst, G., Sathiaseelan, A., and R. Secchi, "Updating
TCP to Support Rate-Limited Traffic", RFC 7661,
DOI 10.17487/RFC7661, October 2015,
<https://www.rfc-editor.org/info/rfc7661>.
[RFC7680] Almes, G., Kalidindi, S., Zekauskas, M., and A. Morton,
Ed., "A One-Way Loss Metric for IP Performance Metrics
(IPPM)", STD 82, RFC 7680, DOI 10.17487/RFC7680, January
2016, <https://www.rfc-editor.org/info/rfc7680>.
[RFC7799] Morton, A., "Active and Passive Metrics and Methods (with
Hybrid Types In-Between)", RFC 7799, DOI 10.17487/RFC7799,
May 2016, <https://www.rfc-editor.org/info/rfc7799>.
[AFD] Pan, R., Breslau, L., Prabhakar, B., and S. Shenker,
"Approximate fairness through differential dropping", ACM
SIGCOMM Computer Communication Review, Volume 33, Issue 2,
DOI 10.1145/956981.956985, April 2003.
[CCscaling]
Paganini, F., Doyle, J., and S. Low, "Scalable laws for
stable network congestion control", Proceedings of IEEE
Conference on Decision and Control,,
DOI 10.1109/CDC.2001.980095, December 2001.
[CVST] Krueger, T. and M. Braun, "R package: Fast Cross-
Validation via Sequential Testing", version 0.1, 11 2012.
[iPerf] Wikipedia, "iPerf", November 2017,
<https://en.wikipedia.org/w/
index.php?title=Iperf&oldid=810583885>.
[MBMSource]
"mbm", July 2016, <https://github.com/m-lab/MBM>.
[Montgomery90]
Montgomery, D., "Introduction to Statistical Quality
Control", 2nd Edition, ISBN 0-471-51988-X, 1990.
[mpingSource]
"mping", July 2016, <https://github.com/m-lab/mping>.
[MSMO97] Mathis, M., Semke, J., Mahdavi, J., and T. Ott, "The
Macroscopic Behavior of the TCP Congestion Avoidance
Algorithm", Computer Communications Review, Volume 27,
Issue 3, DOI 10.1145/263932.264023, July 1997.
[Pathdiag] Mathis, M., Heffner, J., O'Neil, P., and P. Siemsen,
"Pathdiag: Automated TCP Diagnosis", Passive and Active
Network Measurement, Lecture Notes in Computer Science,
Volume 4979, DOI 10.1007/978-3-540-79232-1_16, 2008.
[Policing] Flach, T., Papageorge, P., Terzis, A., Pedrosa, L., Cheng,
Y., Karim, T., Katz-Bassett, E., and R. Govindan, "An
Internet-Wide Analysis of Traffic Policing", Proceedings
of ACM SIGCOMM, DOI 10.1145/2934872.2934873, August 2016.
[RACK] Cheng, Y., Cardwell, N., Dukkipati, N., and P. Jha, "RACK:
a time-based fast loss detection algorithm for TCP", Work
in Progress, draft-ietf-tcpm-rack-03, March 2018.
[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, 2011,
<http://www.R-project.org/>.
[TSO_fq_pacing]
Dumazet, E. and Y. Chen, "TSO, fair queuing, pacing:
three's a charm", Proceedings of IETF 88, TCPM WG,
November 2013,
<https://www.ietf.org/proceedings/88/slides/
slides-88-tcpm-9.pdf>.
[TSO_pacing]
Corbet, J., "TSO sizing and the FQ scheduler", August
2013, <https://lwn.net/Articles/564978/>.
[Wald45] Wald, A., "Sequential Tests of Statistical Hypotheses",
The Annals of Mathematical Statistics, Volume 16, Number
2, pp. 117-186, June 1945,
<http://www.jstor.org/stable/2235829>.
[wikiBloat]
Wikipedia, "Bufferbloat", January 2018,
<https://en.wikipedia.org/w/
index.php?title=Bufferbloat&oldid=819293377>.
[WPING] Mathis, M., "Windowed Ping: An IP Level Performance
Diagnostic", Computer Networks and ISDN Systems, Volume
27, Issue 3, DOI 10.1016/0169-7552(94)90119-8, June 1994.Appendix A. Model Derivations 🎀
okay so the reference target_run_length from Section 5.2 is built on really conservative assumptions 💅 It assumes that ALL the excess data in flight (aka the window size) above the target_window_size piles up into a standing queue that bumps up the RTT, and that classic Reno congestion control with delayed ACKs is running the show. In this section we give two alternative calculations with different assumptions ✨
It might feel kinda random to give a measurement method this much wiggle room, but this section comes with offsetting requirements 🫶
The estimates here make the most sense when you look at network performance logarithmically 📈 In the operational Internet, data rates span more than eight orders of magnitude, RTT spans more than three orders of magnitude, and packet loss ratio spans at least eight orders of magnitude if not more. When you view that logarithmically (like in decibels), it adds up to 80 dB of dynamic range 🤯 On an 80 dB scale, a 3 dB error is less than 4% of the scale, even though it's a factor of 2 in the untransformed parameter. iconic, honestly 😭
This document gives a lot of latitude for calculating target_run_length, but people designing a TIDS should think about how their choices feed into the ongoing tussle about whether "TCP friendliness" is the right model for Internet capacity allocation 👑 Picking a target_run_length that's substantially smaller than the reference target_run_length in Section 5.2 strengthens the argument that it may be appropriate to abandon "TCP friendliness" as the Internet fairness model. That gives developers incentive and permission to build even more aggressive applications and protocols, for example by opening more connections concurrently. the drama is real 💀
A.1. Queueless Reno
Back in Section 5.2, the models assumed that the subpath IP rate matches the target rate plus overhead, so the excess window needed for the AIMD sawtooth causes a fluctuating queue at the bottleneck 🌊
But there's another situation: a bottleneck with no significant queue, where losses come from some mechanism that doesn't add extra delay, like a virtual queue as done in Approximate Fair Dropping [AFD] 🦋 A flow controlled by that kind of bottleneck has a constant RTT and a data rate that fluctuates in a sawtooth because of AIMD congestion control. Assume the losses are being controlled so the average data rate meets some goal that is equal to or greater than the target_rate. The necessary run length to meet the target_rate can be worked out like this:
For some value of Wmin, the window sweeps from Wmin packets to 2*Wmin packets in 2*Wmin RTT (because of delayed ACK). Unlike the queuing case, where Wmin = target_window_size, we want the average of Wmin and 2*Wmin to be the target_window_size, so the average data rate is the target rate. Thus, we want Wmin = (2/3)*target_window_size ✨
Between losses, each sawtooth delivers (1/2)(Wmin+2*Wmin)(2Wmin) packets in 2*Wmin RTTs.
Substituting these together, we get:
target_run_length = (4/3)(target_window_size^2)Note that this is 44% of the reference_run_length computed earlier. That makes sense, because under the assumptions in Section 5.2, the AIMD sawtooth caused a queue at the bottleneck, which raised the effective RTT by 50% 📈
Appendix B. The Effects of ACK Scheduling 🎀
For many network technologies, simple queuing models don't apply 🙈 The network schedules, thins, or otherwise changes the timing of ACKs and data, usually to make the channel allocation algorithms more efficient when they're dealing with relatively widely spaced small ACKs. These efficiency strategies are everywhere for half-duplex, wireless, and broadcast media.
Messing with the ACK stream by holding or thinning ACKs usually has two consequences. First, it raises the implied bottleneck IP capacity, making the fine-grained slowstart bursts either faster or larger. Second, it raises the effective RTT by the average time the ACKs and data are delayed. The first effect can be partly mitigated by re-clocking ACKs once they're past the bottleneck on the return path to the sender, but that further raises the effective RTT 😩
The most extreme example would be a half-duplex channel that isn't released as long as the endpoint holding it has more traffic (data or ACKs) to send. In environments like that, self-clocked protocols under full load revert to extremely inefficient stop-and-wait behavior. The channel forces the protocol to send an entire window of data as one contiguous burst on the forward path, followed by the entire window of ACKs on the return path. (A channel with this behavior would fail the Duplex Self-Interference Test described in Section 8.2.4.) main character behavior, but not in a good way 💀
If a particular return path contains a subpath or device that alters the timing of the ACK stream, then the entire front path from the sender up to the bottleneck must be tested at the burst parameters implied by the ACK scheduling algorithm. The most important parameter is the implied bottleneck IP capacity, which is the average rate at which the ACKs advance snd.una. Also, thinning the ACK stream (relying on the cumulative nature of seg.ack to permit discarding some ACKs) causes most TCP implementations to send interface rate bursts to offset the longer gaps between ACKs and keep the average data rate up 🍓
Because of ubiquitous self-clocking in Internet protocols, ill-conceived channel allocation mechanisms are likely to increase the queuing stress on the front path, since they cause larger full sender rate data bursts. not it, causing queue chaos 🌪️
Holding data or ACKs for channel allocation or other reasons (such as forward error correction) always raises the effective RTT relative to the minimum delay for the path 😭 So it may be necessary to replace target_RTT in the calculation in Section 5.2 with an effective_RTT, which includes the target_RTT plus a term accounting for the extra delays these mechanisms introduce 🫶
Acknowledgments 💜
Ganga Maguluri suggested the statistical test for measuring loss probability in the target run length. Alex Gilgur and Merry Mou helped with the statistics 📊
Meredith Whittaker improved the clarity of the communications 🌸
Ruediger Geib provided feedback that greatly improved the document 🧚
This work was inspired by Measurement Lab: open tools running on an open platform, using open tools to collect open data. See http://www.measurementlab.net/. the open data era, bestie 🌈
Authors' Addresses 🦋
Matt Mathis
Google, Inc
1600 Amphitheatre Parkway
Mountain View, CA 94043
United States of America
Email: mattmathis@google.comAl Morton
AT&T Labs
200 Laurel Avenue South
Middletown, NJ 07748
United States of America
Phone: +1 732 420 1571
Email: acmorton@att.com📦 where this page comes from
- ietf: rfc8337 2018-03
- 💖 girlified by claude-haiku-5-5 on : a language model's rewrite of the original, under its terms