Network Working Group T. Mizrahi
Internet-Draft Huawei
Intended status: Informational G. Fioccola
Expires: April 28, 2022 Huawei Technologies
M. Cociglio
Telecom Italia
M. Chen
Huawei Technologies
G. Mirsky
Ericsson
October 25, 2021
Marking Methods for Performance Measurement
draft-mizrahi-ippm-marking-00
Abstract
This memo presents a summary of marking methods for performance
measurements, and discusses the tradeoffs among them. These marking
methods enable to measure various performance metrics such as packet
loss and delay, and require a low overhead in the header of data
packets, as low as one or two bits per packet, or in some cases even
zero bits per packet. The target audience of this document is
network protocol designers; this document is intended to help
protocol designers choose the best marking method(s) based on the
protocol's constraints and requirements.
Status of This Memo
This Internet-Draft is submitted in full conformance with the
provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF). Note that other groups may also distribute
working documents as Internet-Drafts. The list of current Internet-
Drafts is at https://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any
time. It is inappropriate to use Internet-Drafts as reference
material or to cite them other than as "work in progress."
This Internet-Draft will expire on April 28, 2022.
Mizrahi, et al. Expires April 28, 2022 [Page 1]
Internet-Draft Marking Methods for Performance Measurement October 2021
Copyright Notice
Copyright (c) 2021 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
(https://trustee.ietf.org/license-info) in effect on the date of
publication of this document. Please review these documents
carefully, as they describe your rights and restrictions with respect
to this document. Code Components extracted from this document must
include Simplified BSD License text as described in Section 4.e of
the Trust Legal Provisions and are provided without warranty as
described in the Simplified BSD License.
Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1. Background . . . . . . . . . . . . . . . . . . . . . . . 3
1.2. Scope . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1. Abbreviations . . . . . . . . . . . . . . . . . . . . . . 5
3. Marking Abstractions . . . . . . . . . . . . . . . . . . . . 5
4. Double Marking . . . . . . . . . . . . . . . . . . . . . . . 6
5. Single-bit Marking . . . . . . . . . . . . . . . . . . . . . 7
5.1. Single Marking Using the First Packet . . . . . . . . . . 7
5.2. Single Marking using the Mean Delay . . . . . . . . . . . 8
5.3. Single Marking using a Multiplexed Marking Bit . . . . . 8
5.3.1. Overview . . . . . . . . . . . . . . . . . . . . . . 8
5.4. Pulse Marking . . . . . . . . . . . . . . . . . . . . . . 9
6. Zero Marking: Hashed . . . . . . . . . . . . . . . . . . . . 10
6.1. Hash-based Sampling . . . . . . . . . . . . . . . . . . . 10
6.1.1. Hashed Pulse Marking . . . . . . . . . . . . . . . . 11
6.1.2. Hashed Step Marking . . . . . . . . . . . . . . . . . 11
7. Single Marking: Hashed . . . . . . . . . . . . . . . . . . . 11
8. Timing and Synchronization Aspects . . . . . . . . . . . . . 12
8.1. Synchronization Aspects in Multiplexed Marking . . . . . 13
9. Multipoint Marking Methods . . . . . . . . . . . . . . . . . 14
10. Summary of Marking Methods . . . . . . . . . . . . . . . . . 15
11. Alternate Marking using Reserved Values . . . . . . . . . . . 19
12. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 20
13. Security Considerations . . . . . . . . . . . . . . . . . . . 20
14. References . . . . . . . . . . . . . . . . . . . . . . . . . 20
14.1. Normative References . . . . . . . . . . . . . . . . . . 20
14.2. Informative References . . . . . . . . . . . . . . . . . 20
Appendix A. Ongoing Marking Work in the IETF . . . . . . . . . . 23
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 24
Mizrahi, et al. Expires April 28, 2022 [Page 2]
Internet-Draft Marking Methods for Performance Measurement October 2021
1. Introduction
1.1. Background
Performance measurement using packet marking, defined in [RFC8321],
is a method for measuring performance metrics such as packet loss and
packet delay. Typical delay and loss measurement protocols require
the two measurement points (MPs) to exchange timestamps and/or
counters which are carried over test packets or embedded in the
header of data packets. In contrast, marking methods do not require
timestamps or counters to be exchanged. Instead, every data packet
carries one or more marking bits used for triggering measurement
events. Note that the frequency of these measurement events is
dependent on the users' application(s) and the node characteristics.
One of the most notable marking methods is Alternate Marking
[RFC8321], in which the marking bit is used as a color indication
that is toggled periodically. This approach is illustrated in
Figure 1.
A: packet with color 0
B: packet with color 1
Packets AAAAAAAAAA BBBBBBBBBB AAAAAAAAAA BBBBBBBBBB AAAAAAAAAA
Time ---------------------------------------------------------->
| | | | |
| Block 1 | Block 2 | Block 3 | Block 4 | Block 5 ...
| | | | |
Color 0000000000 1111111111 0000000000 1111111111 0000000000
Figure 1: Alternate marking: packets are monitored on a per-color
basis.
Alternate marking is used between two MPs, the initiating MP and the
monitoring MP. The initiating MP incorporates the marking field into
en-route packets, allowing the monitoring MP to use the marking field
in order to bind each packet to the corresponding block.
Alternate marking can be used for loss measurement and/or delay
measurement. For example, loss measurement can be performed by
having each of the MPs maintains two counters, one per color. At the
end of each block, the counter values can be collected by a central
management system and analyzed; the packet loss can be computed by
comparing the counter values of the two MPs.
Alternate marking, as described above, requires a single marking bit
per packet. Double marking is an approach that uses an additional
Mizrahi, et al. Expires April 28, 2022 [Page 3]
Internet-Draft Marking Methods for Performance Measurement October 2021
marking bit, thus simplifying the measurement method. Double marking
is further described in this document.
Allocating one or two bits in the header of every packet is not
necessarily possible in every encapsulation. For example, if marking
is implemented over IPv4, allocating two marking bits in the IPv4
header is challenging, as every bit in the 20-octet header is costly;
one of the possible approaches discussed in [RFC8321] is to use one
or two bits from the DSCP field for marking. In this case, every
marking bit comes at the expense of reducing the DSCP range by a
factor of two. Thus, there is a high motivation to use marking
methods that use a small number of bits: either a single marking bit
or no marking bits at all.
This memo presents an overview of double marking methods as well as
more efficient methods that require a single marking bit, or zero
marking bits.
Several single-bit marking methods are described, and specifically
multiplexed marking and pulse marking. These two methods were
introduced in [I-D.mizrahi-ippm-multiplexed-alternate-marking] and
[IEEE-Network-PNPM]. In multiplexed marking, the color indicator and
the timestamp indicator are multiplexed into a single bit, providing
the advantages of the double marking method while using a single bit
in the packet header. In pulse marking, both delay and loss
measurements are triggered by a 'pulse' value in a single marking
field.
1.2. Scope
This document also discusses zero-bit marking methods that leverage
well-known hash-based selection approaches ([RFC5474], [RFC5475]).
Marking methods are discussed in this memo as using a single bit or
two bits. However, these methods can similarly be applied to larger
fields, such as an IPv6 Flow Label or an MPLS Label; single-bit
marking can be applied using two reserved values, and two-bit marking
can be applied using four reserved values. Marking based on reserved
values is further discussed in this document, including its
application to MPLS and IPv6.
This memo presents a summary of the various marking methods, and
discusses the tradeoffs among them. It is expected that different
network protocols will have different constraints, and therefore may
choose to use different alternate marking methods. In some cases it
may be preferable to support more than one marking method; in this
case the particular marking method may be signaled through the
control plane.
Mizrahi, et al. Expires April 28, 2022 [Page 4]
Internet-Draft Marking Methods for Performance Measurement October 2021
Note (to be removed before publication): this draft is partially
based on [I-D.mizrahi-ippm-compact-alternate-marking] (expired).
2. Terminology
2.1. Abbreviations
The following abbreviations are used in this document:
DSCP Differentiated Services Code Point
DM Delay Measurement
LM Loss Measurement
LSP Label Switched Path
MP Measurement Point
MPLS Multiprotocol Label Switching
SFL Synonymous Flow Label [RFC8957]
3. Marking Abstractions
The marking methods that are discussed in this document use two basic
abstractions, pulse detection, and step detection.
The common thread along the various marking methods is that one or
two marking bits are used by the MPs to signal a measurement event.
The value of the marking bit indicates when the event takes place, in
one of two ways:
Pulse An event is detected when the value of the marking bit
is toggled in a single packet.
Step An event is detected when the value of the marking bit
is toggled and remains at the new value.
Double marking (Section 4) uses pulse-based detection for DM and
step-based detection for LM.
Pulse-based detection affects the processing of a single packet; the
packet that indicates the pulse is processed differently than the
packets around it. For example, in the double marking method, the
marked packet is timestamped for DM, without affecting the packets
before or after it. Note that if the marked packet is lost, no pulse
is detected, yielding a missing measurement (see Figure 2).
Mizrahi, et al. Expires April 28, 2022 [Page 5]
Internet-Draft Marking Methods for Performance Measurement October 2021
P: indicates a packet
Packets PPPPPPPPPP PPPPPPPPPP PPPPPPPPPP PPPPPPPPPP PPPPPPPPPP
Time ---------------------------------------------------------->
Marking bit 0000010000 0000010000 0000010000 0000010000 00000 0000
^ ^ ^ ^ ^
Pulse-based | | | | |
detection | | | | |
Dropped packet:
no detection
Figure 2: Pulse-based Detection.
In step-based detection, the event is detected by observing a value
change in a stream of packets. Specifically, when the step approach
is used for LM (as in the double marking method), two counters are
used per flow; each MP decides which counter to use based on the
value of the marking bit. Thus, the step-based approach allows
accurate counting even when packets arrive out-of-order (see
Figure 3). When the step approach is used for DM (e.g., single
marking using the first packet), out-of-order causes the delay
measurement to be false, without any indication to the management
system.
P: indicates a packet
Packets PPPPPPPPPP PPPPPPPPPP PPPPPPPPPP PPPPPPPPPP PPPPPPPPPP
Time ---------------------------------------------------------->
Marking bit 0000000000 1111111111 000000000 10111111111 0000000000
^ ^ ^ ^
Step-based | | | |
detection | | | |
out-of-order
Figure 3: Step-based Detection.
4. Double Marking
The two-bit marking method of [RFC8321] uses two marking bits: a
color indicator and a delay measurement indicator. The color bit is
used for step-based LM, while the delay bit is used as a pulse-based
DM trigger. This double marking approach is the most straightforward
of the approaches discussed in this memo, as it allows accurate
measurement, is resilient to out-of-order delivery and is relatively
Mizrahi, et al. Expires April 28, 2022 [Page 6]
Internet-Draft Marking Methods for Performance Measurement October 2021
simple to implement. The main drawback is that it requires two bits,
which are not always available.
Figure 4 illustrates the double marking method: each block of packets
includes a packet that is marked for timestamping, and therefore has
its delay bit set.
A: packet with color 0
B: packet with color 1
Packets AAAAAAAAAA BBBBBBBBBB AAAAAAAAAA BBBBBBBBBB AAAAAAAAAA
Time ---------------------------------------------------------->
| | | | |
| Block 1 | Block 2 | Block 3 | Block 4 | Block 5 ...
| | | | |
Color bit 0000000000 1111111111 0000000000 1111111111 0000000000
Delay bit 0000100000 0000100000 0000100000 0000100000 0001000000
^ ^ ^ ^ ^
Packets | | | | |
marked for | | | | |
timestamping | | | | |
Figure 4: The double marking method.
5. Single-bit Marking
5.1. Single Marking Using the First Packet
This method uses a single marking bit that indicates the color, as
described in [RFC8321]. Both LM and DM are implemented using a step-
based approach; LM is implemented using two color-based counters per
flow. The first packet of every period is used by the two MPs as the
reference for measuring the delay. As denoted above, the delay
computed in this method may be erroneous when packets are delivered
out-of-order.
Mizrahi, et al. Expires April 28, 2022 [Page 7]
Internet-Draft Marking Methods for Performance Measurement October 2021
A: packet with color 0
B: packet with color 1
Packets AAAAAAAAAA BBBBBBBBBB AAAAAAAAAA BBBBBBBBBB AAAAAAAAAA
Time ---------------------------------------------------------->
| | | | |
| Block 1 | Block 2 | Block 3 | Block 4 | Block 5 ...
| | | | |
Color bit 0000000000 1111111111 0000000000 1111111111 0000000000
^ ^ ^ ^ ^
Packets | | | | |
used for DM | | | | |
Figure 5: Single marking using the first packet of the block.
5.2. Single Marking using the Mean Delay
As in the first-packet approach, in the mean delay approach
([RFC8321]) a single marking bit is used to indicate the color,
enabling step-based loss measurement. Delay is measured in each
period by averaging the measured delay over all the packets in the
period. As discussed above, this approach is not sensitive to out-
of-order delivery, but may be heavy from a computational perspective.
5.3. Single Marking using a Multiplexed Marking Bit
5.3.1. Overview
This section introduces a method that uses a single marking bit that
serves two purposes: a color indicator and a timestamp indicator.
The double marking method that was discussed in the previous section
uses two 1-bit values: a color indicator C, and a timestamp indicator
T. The multiplexed marking bit, denoted by M, is an exclusive or
between these two values: M = C XOR T.
An example of the use of the multiplexed marking bit is depicted in
Figure 6. The example considers two routers, R1 and R2, that use the
multiplexed bit method to measure traffic from R1 to R2. In each
block, R1 designates one of the packets for delay measurement. In
each of these designated packets, the value of the multiplexed bit is
reversed compared to the other packets in the same block, allowing R2
to distinguish the designated packets from the other packets.
Mizrahi, et al. Expires April 28, 2022 [Page 8]
Internet-Draft Marking Methods for Performance Measurement October 2021
A: packet with color 0
B: packet with color 1
Packets AAAAAAAAAA BBBBBBBBBB AAAAAAAAAA BBBBBBBBBB AAAAAAAAAA
Time ---------------------------------------------------------->
| | | | |
| Block 1 | Block 2 | Block 3 | Block 4 | Block 5 ...
| | | | |
Color 0000000000 1111111111 0000000000 1111111111 0000000000
^ ^ ^ ^ ^
Packets | | | | |
marked for | | | | |
timestamping | | | | |
v v v v v
Muxed bit 0000100000 1111011111 0000100000 1111101111 0001000000
Figure 6: Alternate marking with a multiplexed bit.
5.4. Pulse Marking
Pulse marking uses a single marking bit that is used as a trigger for
both LM and DM. In this method, the two MPs maintain a single per-
flow counter for LM, in contrast to the color-based methods, which
require two counters per flow. In each block, one of the packets is
marked. The marked packet triggers two actions in each of MPs:
o The timestamp is captured for DM.
o The value of the counter is captured for LM.
In each period, each of the MPs exports the timestamp and counter-
stamp to the management system, which can then compute the loss and
delay in that period. It should be noted that as in [RFC8321], if
the length of the measurement period is L time units, then all
network devices must be synchronized to the same clock reference with
an accuracy of +/- L/2 time units.
The pulse marking approach is illustrated in Figure 7. Since both LM
and DM use a pulse-based trigger, if the marked packet is lost, then
no measurement is available in this period. Moreover, the LM
accuracy may be affected by out-of-order delivery.
Mizrahi, et al. Expires April 28, 2022 [Page 9]
Internet-Draft Marking Methods for Performance Measurement October 2021
P: packet - all packets have the same color
Packets PPPPPPPPPP PPPPPPPPP PPPPPPPPPP PPPPPPPPPP PPPPPPPPPP
Time ---------------------------------------------------------->
| | | | |
| Block 1 | Block 2 | Block 3 | Block 4 | Block 5 ...
| | | | |
^ ^ ^ ^ ^
Packets | | | | |
marked for | | | | |
DM and LM | | | | |
v v v v v
Marking bit 0000100000 0000100000 0000100000 0000010000 0001000000
Figure 7: Pulse marking method.
6. Zero Marking: Hashed
6.1. Hash-based Sampling
Hash-based selection [RFC5475] is a well-known method for sampling a
subset of packets. As defined in [RFC5475]:
A Hash Function h maps the Packet Content c, or some portion of
it, onto a Hash Range R. The packet is selected if h(c) is an
element of S, which is a subset of R called the Hash Selection
Range.
Hash-based selection can be leveraged as a marking method, allowing a
zero-bit marking approach. Specifically, the pulse and step
abstractions can be implemented using hashed selection:
o Hashed pulse-based trigger: in this approach, a packet is selected
if h(c) is an element of S, which is a strict subset of the hash
range R. When |S|<<|R|, the average sampling period is long,
reducing the probability of ambiguity between consecutive
packets. |S| and |R| denote the number of elements in S and R,
respectively.
o Hashed step-based trigger: the hash values of a given traffic flow
are said to be monotonically increasing if for two packets p1 and
p2, if p1 is sent before p2 then h(p1) <= h(p2). If it is
guaranteed that the hash values of a flow are monotonically
increasing, then a step-based approach can be used on the range R.
For example, in an IPv4 flow, the Identification field can be used
as the hash value of each packet. Since the Identification field
is monotonically increasing, the step-based trigger can be
Mizrahi, et al. Expires April 28, 2022 [Page 10]
Internet-Draft Marking Methods for Performance Measurement October 2021
implemented using consecutive ranges of the Identification value.
For example, the fourth bit of the Identification field is toggled
every eight packets. Thus, a possible hash function simply takes
the fourth bit of the Identification field as the hash value.
This hash value is toggled every eight packets, simulating the
alternate marking behavior of Section 4.
Note that as opposed to the double marking and single marking
methods, hashed sampling is not based on fixed time intervals, as the
duration between sampled packets depends only on the hash value.
It is also important to note that all methods that use hash-based
marking require the hash function and the set S to be configured
consistently across the MPs.
6.1.1. Hashed Pulse Marking
In this approach, a hash is computed over the packet content, and
both LM and DM are triggered based on the pulse-based trigger
(Section 6.1). A pulse is detected when the hash value h(c) is equal
to one of the values in S. The hash function h and the set S
determine the probability (or frequency) of the pulse event.
6.1.2. Hashed Step Marking
As in the previous approach, hashed step marking also uses a hash
that is computed over the packet content. In this approach, DM is
performed using a pulse-based trigger, whereas the LM trigger is
step-based (Section 6.1). The main drawback of this method is that
the step-based trigger is possible only under the assumption that the
hash function is monotonically increasing, which is not necessarily
possible in all cases. Specifically, a measured flow is not
necessarily an IPv4 5-tuple. For example, a measured flow may
include multiple IPv4 5-tuple flows, and in this case, the
Identification field is not monotonically increasing.
7. Single Marking: Hashed
Mixed hashed marking combines the single marking approach with hash-
based sampling. A single marking bit is used in the packet header as
a color indicator, while a hash-based pulse is used to trigger DM.
Although this method requires a single bit, it is described in this
section as it is closely related to the other hash-based methods that
require zero marking bits.
The hash-based selection for DM can be applied in one of two possible
approaches: the basic approach and the dynamic approach. In the
basic approach, packets forwarded between two MPs, MP1 and MP2, are
Mizrahi, et al. Expires April 28, 2022 [Page 11]
Internet-Draft Marking Methods for Performance Measurement October 2021
selected using a hash function, as described above. One of the
challenges is that the frequency of the sampled packets may vary
considerably, making it difficult for the management system to
correlate samples from the two MPs. Thus, the dynamic approach can
be used.
In the dynamic hash-based sampling, alternate marking is used to
create divide time into periods, so that hash-based samples are
divided into batches, allowing to anchor the selected samples to
their period. Moreover, by dynamically adapting the length of the
hash value, the number of samples is bounded in each marking period.
This can be realized by choosing first the maximum number of samples
(NMAX) to be used with the initial hash length. The algorithm starts
with only a few hash bits, which permit the selection of a greater
percentage of packets (e.g., with 1 bit of hash, half of the packets
are sampled). When the number of selected packets reaches NMAX, a
hashing bit is added. Consequently, the sampling proceeds at half of
the original rate, and the packets already selected that do not match
the new hash are discarded. This step can be repeated iteratively.
It is assumed that each sample includes the timestamp (used for DM)
and the hash value, allowing the management system to match the
samples received from the two MPs.
The dynamic process statistically converges at the end of a marking
period, and the number of selected samples beyond the initial NMAX
samples mentioned above is between NMAX/2 and NMAX. Therefore, the
dynamic approach paces the sampling rate, allowing to bound the
number of sampled packets per sampling period.
8. Timing and Synchronization Aspects
As pointed out in [RFC8321], it is assumed that all MPs are
synchronized to a common reference time with an accuracy of +/- L/2,
where L is the periodic measurement interval. Thus, the difference
between the clock values of any two MPs is bounded by L. Note that
this is a relatively relaxed synchronization requirement that does
not require complex means of synchronization. Clocks can be
synchronized, for example, using NTP [RFC5905], PTP [IEEE1588], or by
other means.
In the step-based approaches, the common reference time is used for
dividing the time domain into equal-sized measurement periods, such
that all packets forwarded during a measurement period have the same
color, and consecutive periods have alternating colors. In the
pulse-based approaches the synchronization helps the management
system to correlate measurements from multiple measurement points
without ambiguity.
Mizrahi, et al. Expires April 28, 2022 [Page 12]
Internet-Draft Marking Methods for Performance Measurement October 2021
8.1. Synchronization Aspects in Multiplexed Marking
The single marking bit incorporates two multiplexed values. From the
monitoring MP's perspective, the two values are Time-Division
Multiplexed (TDM), as depicted in Figure 8. It is assumed that the
start time of every measurement period is known to both the
initiating MP and the monitoring MP. If the measurement period is L,
then during the first and the last L/4 time units of each block the
marking bit is interpreted by the monitoring MP as a color indicator.
During the middle part of the block, the marking bit is interpreted
as a timestamp indicator; if the value of this bit is different than
the color value, the corresponding packet is used as a reference for
delay measurement.
+--- Beginning of measurement period
|
v
...BBBBBBBBB | AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA | BBBBBBBBB...
|<======================================>|
| L |
<========>|<========><==================><========>|<========>
L/4 L/4 L/2 L/4 L/4
<===================><==================><===================>
Detect color Detect timestamping Detect color
change indication change
Figure 8: Multiplexed marking field interpretation at the receiving
measurement point.
In order to prevent ambiguity in the receiver's interpretation of the
marking field, the initiating MP is permitted to set the timestamp
indication only during a specific interval, as depicted in Figure 9.
Since the receiver is willing to receive the timestamp indication
during the middle L/2 time units of the block, the sender refrains
from sending the timestamp indication during a guardband interval of
d time units at the beginning and end of the L/2-period.
Mizrahi, et al. Expires April 28, 2022 [Page 13]
Internet-Draft Marking Methods for Performance Measurement October 2021
+--- Beginning of measurement period
|
v
...BBBBBBBBB | AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA | BBBBBBBBB...
|<======================================>|
| L |
<========>|<========>|<================>|<========>|
L/4 L/4 | L/2 | L/4
<=>|<=> <=>|<=>
d d d d
<==========>
permissible
timestamping
indication
interval
Figure 9: A time domain view.
The guardband d is given by d = A + D_max - D_min, where A is the
clock accuracy, D_max is an upper bound on the network delay between
the MPs, and D_min is a lower bound on the delay. It is
straightforward from Figure 9 that d < L/4 must be satisfied. The
latter implies a minimal requirement on the synchronization accuracy.
All MPs must be synchronized to the same reference time with an
accuracy of +/- L/8. Depending on the system topology, in some
systems, the accuracy requirement will be even more stringent,
subject to d < L/4. Note that the accuracy requirement of the
conventional alternate marking method [RFC8321] is +/- L/2, while the
multiplexed marking method requires an accuracy of +/- L/8.
Note that we assume that the middle L/2-period is designated as the
timestamp indication period, allowing a sufficiently long guardband
between the transitions. However, a system may be configured to use
a longer timestamp indication period or a shorter one, if it is
guaranteed that the synchronization accuracy meets the guardband
requirements (i.e., the constraints on d).
9. Multipoint Marking Methods
It should be noted that most of the marking methods that were
presented in this memo are intended for point-to-point measurements,
e.g., from MP1 to MP2 in Figure 10. In point-to-multipoint
measurements, the mean delay method can be used to measure the loss
and delay of the entire point-to-multipoint flow (which includes all
the traffic from MP3 to either MP4 or MP5), while other methods such
as double marking can be used to measure the point-to-point
Mizrahi, et al. Expires April 28, 2022 [Page 14]
Internet-Draft Marking Methods for Performance Measurement October 2021
performance, for example, from MP3 to MP5. Alternate marking in
multipoint scenarios is discussed in detail in
[I-D.ietf-ippm-multipoint-alt-mark].
MP1 MP2 MP3 MP4
+--+ +--+ +--+ +--+ +--+
| |---------->| | | |----->| |----->| |
+--+ +--+ +--+ +--+ +--+
|
| MP5
| +--+
+------>| |
+--+
Point-to-point measurement Point-to-multipoint measurement
Figure 10: Point-to-point and point-to-multipoint measurements.
10. Summary of Marking Methods
This section summarizes the marking methods described in this memo.
Each row in the table of Figure 11 represents a marking method. For
each method, the table specifies the number of bits required in the
header, the number of counters per flow for LM, the methods used for
LM and DM (pulse or step), and also the resilience to disturbances.
Mizrahi, et al. Expires April 28, 2022 [Page 15]
Internet-Draft Marking Methods for Performance Measurement October 2021
+--------------+----+----+------+------+-------------+-------------+
| Method |# of|# of|LM |DM |Resilience to|Resilience to|
| |bits|coun|Method|Method|Reordering |Packet drops |
| | |ters| | +------+------+------+------+
| | | | | | LM | DM | LM | DM |
+--------------+----+----+------+------+------+------+------+------+
|Single marking| 1 | 2 |Step |Step | + | -- | + | -- |
|- 1st packet | | | | | | | | |
+--------------+----+----+------+------+------+------+------+------+
|Single marking| 1 | 2 |Step |Mean | + | + | + | - |
|- mean delay | | | | | | | | |
+--------------+----+----+------+------+------+------+------+------+
|Double marking| 2 | 2 |Step |Pulse | + | + | + | = |
+--------------+----+----+------+------+------+------+------+------+
|Single marking| 1 | 2 |Step |Pulse | + | + | + | = |
|multiplexed | | | | | | | | |
+--------------+----+----+------+------+------+------+------+------+
|Pulse marking | 1 | 1 |Pulse |Pulse | -- | + | - | = |
+--------------+----+----+------+------+------+------+------+------+
|Zero marking | 0 | 1 |Hashed|Hashed| -- | + | - | + |
|- hashed | |(2) |pulse |pulse | (-) | | | |
| | | |(step)| | | | | |
+--------------+----+----+------+------+------+------+------+------+
|Single marking| 1 | 2 |Step |Hashed| + | + | + | + |
|- hashed | | | |pulse | | | | |
+--------------+----+----+------+------+------+------+------+------+
+ Accurate measurement.
= Invalidate only if a measured packet is lost (detectable)
- No measurement in case of disturbance (detectable).
-- False measurement in case of disturbance (not detectable).
Figure 11: Detailed Summary of Marking Methods
In the context of this comparison, two possible disturbances are
considered: out-of-order delivery and packet drops. Generally
speaking, pulse-based methods are sensitive to packet drops since if
the marked packet is dropped, no measurement is recorded in the
current period. Notably, a missing measurement is detectable by the
management system, and is not as severe as a false measurement.
Step-based triggers are generally resilient to out-of-order delivery
for LM, but are not resilient to out-of-order delivery for DM.
Notably, a step-based trigger may yield a false delay measurement
when packets are delivered out-of-order, and this inaccuracy is not
detectable.
As mentioned above, the double marking method is the most
straightforward approach and is resilient to most of the disturbances
Mizrahi, et al. Expires April 28, 2022 [Page 16]
Internet-Draft Marking Methods for Performance Measurement October 2021
that were analyzed. Its obvious drawback is that it requires two
marking bits.
Several single marking methods are discussed in this memo. In this
case, there is no clear verdict on which method is the optimal one.
The first packet method may be simple to implement, but may present
erroneous delay measurements in case of dropped or reordered packets.
Arguably, the mean delay approach and the multiplexed approach may be
more difficult to implement (depending on the underlying platform),
but are more resilient to the disturbances that were considered here.
Note that the computational complexity of the mean delay approach can
be reduced by combining it with a hashed approach, i.e., by computing
the mean delay over a hash-based subset of the packets. The pulse
marking method requires only a single counter per flow, while the
other methods require two counters per flow.
The hash-based sampling approaches reduce the overhead to zero bits,
which is a significant advantage. However, the sampling period in
these approaches is not associated with a fixed time interval.
Therefore, in some cases, adjacent packets may be selected for the
sampling, potentially causing measurement errors. Furthermore, when
the traffic rate is low, measurements may become significantly
infrequent.
It is clear from the previous table that packet loss measurement can
be considered resilient to both reordering and packet drops if at
least one bit is used with a step-based approach. Thus, since the
packet loss can be considered obvious, the previous table can be
simplified into Figure 12, where only the characteristics of delay
measurements are highlighted. This more compact table allows room
for an additional column referring to multipoint-to-multipoint
(Section 9) delay measurement compatibility.
Mizrahi, et al. Expires April 28, 2022 [Page 17]
Internet-Draft Marking Methods for Performance Measurement October 2021
+--------------+----+--------+------------+------------+-----------+
| Marking |# of|LM |DM |DM |DM |
| Method |bits|on |Resilience |Resilience |Multipoint |
| | |All |to |to |compatible |
| | |Packets |Reordering |Packet drops| |
+--------------+----+--------+------------+------------+-----------+
|Single marking| 1 | Yes | -- | - | No |
|- 1st packet | | | | | |
+--------------+----+--------+------------+------------+-----------+
|Single marking| 1 | Yes | + | - | Yes |
|- mean delay | | | | | |
+--------------+----+--------+------------+------------+-----------+
|Double marking| 2 | Yes | + | = | No |
+--------------+----+--------+------------+------------+-----------+
|Single marking| 1 | Yes | + | = | No |
|multiplexed | | | | | |
+--------------+----+--------+------------+------------+-----------+
|Pulse marking | 1 | No | + | = | No |
+--------------+----+--------+------------+------------+-----------+
|Zero marking | 0 | No | + | + | Yes |
|- hashed | | | | | |
| | | | | | |
+--------------+----+--------+------------+------------+-----------+
|Single marking| 1 | Yes | + | + | Yes |
|- hashed | | | | | |
+--------------+----+--------+------------+------------+-----------+
+ Accurate measurement.
= Invalidate only if a measured packet is lost (detectable)
- No measurement in case of disturbance (detectable).
-- False measurement in case of disturbance (not detectable).
Figure 12: Summary of Marking Methods: focus on Delay Measurement
In the context of delay measurement, both zero marking hashed and
single marking hashed are resilient to packet drops. Using double
marking it could also be possible to perform an accurate measurement
in the case of packet drops, as long as the packet that is marked for
DM is not dropped.
The single marking hashed method seems the most complete approach,
especially because it is also compatible with multipoint-to-
multipoint measurements.
Mizrahi, et al. Expires April 28, 2022 [Page 18]
Internet-Draft Marking Methods for Performance Measurement October 2021
11. Alternate Marking using Reserved Values
As mentioned in Section 1, a marking bit is not necessarily a single
bit, but may be implemented by using two well-known values in one of
the header fields. Similarly, two-bit marking can be implemented
using four reserved values.
A notable example is MPLS Synonymous Flow Labels (SFL), as defined in
[I-D.ietf-mpls-rfc6374-sfl]. Two MPLS Label values can be used to
indicate the two colors of a given LSP: the original Label value, and
an SFL value. A similar approach can be applied to IPv6 using the
Flow Label field.
The following example illustrates how alternate marking can be
implemented using reserved values. The bit multiplexing approach of
Section 5.3 is applicable not only to single-bit color indicators,
but also to two-value indicators; instead of using a single bit that
is toggled between '0' and '1', two values of the indicator field, U
and W, can be used in the same manner, allowing both loss and delay
measurement to be performed using only two reserved values. Thus,
the multiplexing approach of Figure 6 can be illustrated more
generally with two values, U and W, as depicted in Figure 13.
A: packet with color 0
B: packet with color 1
Packets AAAAAAAAAA BBBBBBBBBB AAAAAAAAAA BBBBBBBBBB AAAAAAAAAA
Time ---------------------------------------------------------->
| | | | |
| Block 1 | Block 2 | Block 3 | Block 4 | Block 5 ...
| | | | |
Color 0000000000 1111111111 0000000000 1111111111 0000000000
^ ^ ^ ^ ^
Packets | | | | |
marked for | | | | |
timestamping | | | | |
v v v v v
Muxed UUUUWUUUUU WWWWUWWWWW UUUUWUUUUU WWWWWUWWWW UUUWUUUUUU
marking
values
Figure 13: Alternate marking with two multiplexed marking values, U
and W.
Mizrahi, et al. Expires April 28, 2022 [Page 19]
Internet-Draft Marking Methods for Performance Measurement October 2021
12. IANA Considerations
This memo includes no requests from IANA.
13. Security Considerations
The security considerations of the alternate marking method are
discussed in [RFC8321]. The analysis of Section 10 emphasizes the
sensitivity of some of the alternate marking methods to packet drops
and to packet reordering. Thus, a malicious attacker may attempt to
tamper with the measurements by either selectively dropping packets,
or by selectively reordering specific packets. The multiplexed
marking method Section 5.3 that is defined in this document requires
slightly more stringent synchronization than the conventional marking
method, potentially making the method more vulnerable to attacks on
the time synchronization protocol. A detailed discussion about the
threats against time synchronization protocols and how to mitigate
them is presented in [RFC7384].
14. References
14.1. Normative References
[RFC8321] Fioccola, G., Ed., Capello, A., Cociglio, M., Castaldelli,
L., Chen, M., Zheng, L., Mirsky, G., and T. Mizrahi,
"Alternate-Marking Method for Passive and Hybrid
Performance Monitoring", RFC 8321, DOI 10.17487/RFC8321,
January 2018, <https://www.rfc-editor.org/info/rfc8321>.
14.2. Informative References
[I-D.cnbf-ippm-user-devices-explicit-monitoring]
Cociglio, M., Nilo, M., Bulgarella, F., and G. Fioccola,
"User Devices Explicit Monitoring", draft-cnbf-ippm-user-
devices-explicit-monitoring-02 (work in progress), July
2021.
[I-D.fz-spring-srv6-alt-mark]
Fioccola, G., Zhou, T., and M. Cociglio, "Segment Routing
Header encapsulation for Alternate Marking Method", draft-
fz-spring-srv6-alt-mark-01 (work in progress), July 2021.
[I-D.ietf-6man-ipv6-alt-mark]
Fioccola, G., Zhou, T., Cociglio, M., Qin, F., and R.
Pang, "IPv6 Application of the Alternate Marking Method",
draft-ietf-6man-ipv6-alt-mark-12 (work in progress),
October 2021.
Mizrahi, et al. Expires April 28, 2022 [Page 20]
Internet-Draft Marking Methods for Performance Measurement October 2021
[I-D.ietf-bier-pmmm-oam]
Mirsky, G., Zheng, L., Chen, M., and G. Fioccola,
"Performance Measurement (PM) with Marking Method in Bit
Index Explicit Replication (BIER) Layer", draft-ietf-bier-
pmmm-oam-11 (work in progress), October 2021.
[I-D.ietf-ippm-multipoint-alt-mark]
Fioccola, G., Cociglio, M., Sapio, A., and R. Sisto,
"Multipoint Alternate-Marking Method for Passive and
Hybrid Performance Monitoring", draft-ietf-ippm-
multipoint-alt-mark-09 (work in progress), March 2020.
[I-D.ietf-mpls-rfc6374-sfl]
Bryant, S., Swallow, G., Chen, M., Fioccola, G., and G.
Mirsky, "RFC6374 Synonymous Flow Labels", draft-ietf-mpls-
rfc6374-sfl-10 (work in progress), March 2021.
[I-D.mdt-ippm-explicit-flow-measurements]
Cociglio, M., Ferrieux, A., Fioccola, G., Lubashev, I.,
Bulgarella, F., Hamchaoui, I., Nilo, M., Sisto, R., and D.
Tikhonov, "Explicit Flow Measurements Techniques", draft-
mdt-ippm-explicit-flow-measurements-02 (work in progress),
July 2021.
[I-D.mirsky-sfc-pmamm]
Mirsky, G., Fioccola, G., and T. Mizrahi, "Performance
Measurement (PM) with Alternate Marking Method in Service
Function Chaining (SFC) Domain", draft-mirsky-sfc-pmamm-14
(work in progress), September 2021.
[I-D.mizrahi-ippm-compact-alternate-marking]
Mizrahi, T., Arad, C., Fioccola, G., Cociglio, M., Chen,
M., Zheng, L., and G. Mirsky, "Compact Alternate Marking
Methods for Passive and Hybrid Performance Monitoring",
draft-mizrahi-ippm-compact-alternate-marking-05 (work in
progress), July 2019.
[I-D.mizrahi-ippm-multiplexed-alternate-marking]
Mizrahi, T., Arad, C., Fioccola, G., Cociglio, M., Chen,
M., Zheng, L., and G. Mirsky, "Compact Alternate Marking
Methods for Passive Performance Monitoring", draft-
mizrahi-ippm-multiplexed-alternate-marking-02 (work in
progress), June 2017.
Mizrahi, et al. Expires April 28, 2022 [Page 21]
Internet-Draft Marking Methods for Performance Measurement October 2021
[I-D.zhou-ippm-enhanced-alternate-marking]
Zhou, T., Fioccola, G., Liu, Y., Lee, S., Cociglio, M.,
and W. Li, "Enhanced Alternate Marking Method", draft-
zhou-ippm-enhanced-alternate-marking-07 (work in
progress), July 2021.
[IEEE-Network-PNPM]
Mizrahi, T., Navon, G., Fioccola, G., Cociglio, M., Chen,
M., and G. Mirsky, "AM-PM: Efficient Network Telemetry
using Alternate Marking", IEEE Network vol. 33, no. 4,
pp. 155-161, DOI 10.1109/MNET.2019.1800152, July 2019.
[IEEE1588]
IEEE, "IEEE 1588 Standard for a Precision Clock
Synchronization Protocol for Networked Measurement and
Control Systems Version 2", 2008.
[RFC5474] Duffield, N., Ed., Chiou, D., Claise, B., Greenberg, A.,
Grossglauser, M., and J. Rexford, "A Framework for Packet
Selection and Reporting", RFC 5474, DOI 10.17487/RFC5474,
March 2009, <https://www.rfc-editor.org/info/rfc5474>.
[RFC5475] Zseby, T., Molina, M., Duffield, N., Niccolini, S., and F.
Raspall, "Sampling and Filtering Techniques for IP Packet
Selection", RFC 5475, DOI 10.17487/RFC5475, March 2009,
<https://www.rfc-editor.org/info/rfc5475>.
[RFC5905] Mills, D., Martin, J., Ed., Burbank, J., and W. Kasch,
"Network Time Protocol Version 4: Protocol and Algorithms
Specification", RFC 5905, DOI 10.17487/RFC5905, June 2010,
<https://www.rfc-editor.org/info/rfc5905>.
[RFC7384] Mizrahi, T., "Security Requirements of Time Protocols in
Packet Switched Networks", RFC 7384, DOI 10.17487/RFC7384,
October 2014, <https://www.rfc-editor.org/info/rfc7384>.
[RFC8957] Bryant, S., Chen, M., Swallow, G., Sivabalan, S., and G.
Mirsky, "Synonymous Flow Label Framework", RFC 8957,
DOI 10.17487/RFC8957, January 2021,
<https://www.rfc-editor.org/info/rfc8957>.
[RFC9000] Iyengar, J., Ed. and M. Thomson, Ed., "QUIC: A UDP-Based
Multiplexed and Secure Transport", RFC 9000,
DOI 10.17487/RFC9000, May 2021,
<https://www.rfc-editor.org/info/rfc9000>.
Mizrahi, et al. Expires April 28, 2022 [Page 22]
Internet-Draft Marking Methods for Performance Measurement October 2021
Appendix A. Ongoing Marking Work in the IETF
The marking methods that are described in this document are used in
several proposed solutions that are currently under discussion in the
IETF.
IPv6 packet marking is defined in [I-D.ietf-6man-ipv6-alt-mark] using
a new option called the AltMark option, which can be incorporated
either into a Hop-by-Hop or into a Destination Extension Header. A
proposed enhancement of the AltMark option is presented in
[I-D.zhou-ippm-enhanced-alternate-marking].
A similar option was proposed for SRv6 [I-D.fz-spring-srv6-alt-mark],
defined as a Type-Length-Value (TLV) field for the Segment Routing
Header (SRH).
As described in the previous section, MPLS Synonymous Flow Labels
(SFLs) can be used for marking in MPLS [I-D.ietf-mpls-rfc6374-sfl].
This draft discusses both single and double marking, where instead of
using one or two bits, SFLs are used to represent the different
marking values.
Double marking has also been defined in the BIER header in
[I-D.ietf-bier-pmmm-oam].
In the context of Service Function Chains (SFC), the proposal in
[I-D.mirsky-sfc-pmamm] defines a single-bit marking in the Network
Service Header (NSH), with a few different options of how to use the
single marking bit.
It should be noted that the current draft is focused on marking
methods that are unidirectional and connectionless by nature. Other
marking methods that are connection-oriented by nature are used in
the Transport Layer, such as the spin bit in QUIC [RFC9000]. A
generalization of this approach is discussed in
[I-D.cnbf-ippm-user-devices-explicit-monitoring] and
[I-D.mdt-ippm-explicit-flow-measurements]. These Transport Layer
marking methods are not within the scope of the current document.
The following table summarizes the proposed marking solutions that
are currently under discussion, and for each solution the table
specifies whether the solution uses double marking or single marking.
Note that solutions that use double marking can implicitly support
the ability to use single marking as well. In cases where the
solution explicitly includes two separate options, one for single
marking and one for double marking, both columns are marked in the
table.
Mizrahi, et al. Expires April 28, 2022 [Page 23]
Internet-Draft Marking Methods for Performance Measurement October 2021
+--------------------------------------------------+-------+-------+
| Proposed Solution |Double |Single |
| |Marking|Marking|
+--------------------------------------------------+-------+-------+
| [I-D.ietf-6man-ipv6-alt-mark] | + | |
| [I-D.zhou-ippm-enhanced-alternate-marking] | | |
+--------------------------------------------------+-------+-------+
| [I-D.fz-spring-srv6-alt-mark] | + | |
+--------------------------------------------------+-------+-------+
| [I-D.ietf-mpls-rfc6374-sfl] | + | + |
+--------------------------------------------------+-------+-------+
| [I-D.ietf-bier-pmmm-oam] | + | |
+--------------------------------------------------+-------+-------+
| [I-D.mirsky-sfc-pmamm] | | + |
+--------------------------------------------------+-------+-------+
Figure 14: Summary of Ongoing Work on Marking Solutions in the IETF
Authors' Addresses
Tal Mizrahi
Huawei
Israel
Email: tal.mizrahi.phd@gmail.com
Giuseppe Fioccola
Huawei Technologies
Email: giuseppe.fioccola@huawei.com
Mauro Cociglio
Telecom Italia
Via Reiss Romoli, 274
Torino 10148
Italy
Email: mauro.cociglio@telecomitalia.it
Mach(Guoyi) Chen
Huawei Technologies
Email: mach.chen@huawei.com
Mizrahi, et al. Expires April 28, 2022 [Page 24]
Internet-Draft Marking Methods for Performance Measurement October 2021
Greg Mirsky
Ericsson
Email: gregimirsky@gmail.com
Mizrahi, et al. Expires April 28, 2022 [Page 25]