1 .. This work is licensed under a Creative Commons Attribution 4.0 International License.
2 .. http://creativecommons.org/licenses/by/4.0
3 .. (c) OPNFV, Intel Corporation, AT&T and others.
5 ******************************
6 VSPERF LEVEL TEST DESIGN (LTD)
7 ******************************
15 The intention of this Level Test Design (LTD) document is to specify the set of
16 tests to carry out in order to objectively measure the current characteristics
17 of a virtual switch in the Network Function Virtualization Infrastructure
18 (NFVI) as well as the test pass criteria. The detailed test cases will be
19 defined in details-of-LTD_, preceded by the doc-id-of-LTD_ and the scope-of-LTD_.
21 This document is currently in draft form.
31 The document id will be used to uniquely
32 identify versions of the LTD. The format for the document id will be:
33 OPNFV\_vswitchperf\_LTD\_REL\_STATUS, where by the
34 status is one of: draft, reviewed, corrected or final. The document id
35 for this version of the LTD is:
36 OPNFV\_vswitchperf\_LTD\_Brahmaputra\_REVIEWED.
45 The main purpose of this project is to specify a suite of
46 performance tests in order to objectively measure the current packet
47 transfer characteristics of a virtual switch in the NFVI. The intent of
48 the project is to facilitate testing of any virtual switch. Thus, a
49 generic suite of tests shall be developed, with no hard dependencies to
50 a single implementation. In addition, the test case suite shall be
51 architecture independent.
53 The test cases developed in this project shall not form part of a
54 separate test framework, all of these tests may be inserted into the
55 Continuous Integration Test Framework and/or the Platform Functionality
56 Test Framework - if a vSwitch becomes a standard component of an OPNFV
64 * `RFC 1242 Benchmarking Terminology for Network Interconnection
65 Devices <http://www.ietf.org/rfc/rfc1242.txt>`__
66 * `RFC 2544 Benchmarking Methodology for Network Interconnect
67 Devices <http://www.ietf.org/rfc/rfc2544.txt>`__
68 * `RFC 2285 Benchmarking Terminology for LAN Switching
69 Devices <http://www.ietf.org/rfc/rfc2285.txt>`__
70 * `RFC 2889 Benchmarking Methodology for LAN Switching
71 Devices <http://www.ietf.org/rfc/rfc2889.txt>`__
72 * `RFC 3918 Methodology for IP Multicast
73 Benchmarking <http://www.ietf.org/rfc/rfc3918.txt>`__
74 * `RFC 4737 Packet Reordering
75 Metrics <http://www.ietf.org/rfc/rfc4737.txt>`__
76 * `RFC 5481 Packet Delay Variation Applicability
77 Statement <http://www.ietf.org/rfc/rfc5481.txt>`__
78 * `RFC 6201 Device Reset
79 Characterization <http://tools.ietf.org/html/rfc6201>`__
85 ================================
86 Details of the Level Test Design
87 ================================
89 This section describes the features to be tested (FeaturesToBeTested-of-LTD_), and
90 identifies the sets of test cases or scenarios (TestIdentification-of-LTD_).
94 .. _FeaturesToBeTested-of-LTD:
99 Characterizing virtual switches (i.e. Device Under Test (DUT) in this document)
100 includes measuring the following performance metrics:
104 - Packet delay variation
109 - Availability and capacity of the DUT
113 .. _TestIdentification-of-LTD:
123 The following tests aim to determine the maximum forwarding rate that
124 can be achieved with a virtual switch. The list is not exhaustive but
125 should indicate the type of tests that should be required. It is
126 expected that more will be added.
132 Test ID: LTD.Throughput.RFC2544.PacketLossRatio
133 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
135 **Title**: RFC 2544 X% packet loss ratio Throughput and Latency Test
137 **Prerequisite Test**: N/A
143 This test determines the DUT's maximum forwarding rate with X% traffic
144 loss for a constant load (fixed length frames at a fixed interval time).
145 The default loss percentages to be tested are: - X = 0% - X = 10^-7%
147 Note: Other values can be tested if required by the user.
149 The selected frame sizes are those previously defined under
150 :ref:`default-test-parameters`.
151 The test can also be used to determine the average latency of the traffic.
153 Under the `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__
154 test methodology, the test duration will
155 include a number of trials; each trial should run for a minimum period
156 of 60 seconds. A binary search methodology must be applied for each
157 trial to obtain the final result.
159 **Expected Result**: At the end of each trial, the presence or absence
160 of loss determines the modification of offered load for the next trial,
161 converging on a maximum rate, or
162 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__ Throughput with X%
164 The Throughput load is re-used in related
165 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__ tests and other
168 **Metrics Collected**:
170 The following are the metrics collected for this test:
172 - The maximum forwarding rate in Frames Per Second (FPS) and Mbps of
173 the DUT for each frame size with X% packet loss.
174 - The average latency of the traffic flow when passing through the DUT
175 (if testing for latency, note that this average is different from the
176 test specified in Section 26.3 of
177 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__).
178 - CPU and memory utilization may also be collected as part of this
179 test, to determine the vSwitch's performance footprint on the system.
183 .. _PacketLossRatioFrameModification:
185 Test ID: LTD.Throughput.RFC2544.PacketLossRatioFrameModification
186 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
188 **Title**: RFC 2544 X% packet loss Throughput and Latency Test with
191 **Prerequisite Test**: N/A
197 This test determines the DUT's maximum forwarding rate with X% traffic
198 loss for a constant load (fixed length frames at a fixed interval time).
199 The default loss percentages to be tested are: - X = 0% - X = 10^-7%
201 Note: Other values can be tested if required by the user.
203 The selected frame sizes are those previously defined under
204 :ref:`default-test-parameters`.
205 The test can also be used to determine the average latency of the traffic.
207 Under the `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__
208 test methodology, the test duration will
209 include a number of trials; each trial should run for a minimum period
210 of 60 seconds. A binary search methodology must be applied for each
211 trial to obtain the final result.
213 During this test, the DUT must perform the following operations on the
216 - Perform packet parsing on the DUT's ingress port.
217 - Perform any relevant address look-ups on the DUT's ingress ports.
218 - Modify the packet header before forwarding the packet to the DUT's
219 egress port. Packet modifications include:
221 - Modifying the Ethernet source or destination MAC address.
222 - Modifying/adding a VLAN tag. (**Recommended**).
223 - Modifying/adding a MPLS tag.
224 - Modifying the source or destination ip address.
225 - Modifying the TOS/DSCP field.
226 - Modifying the source or destination ports for UDP/TCP/SCTP.
229 **Expected Result**: The Packet parsing/modifications require some
230 additional degree of processing resource, therefore the
231 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__
232 Throughput is expected to be somewhat lower than the Throughput level
233 measured without additional steps. The reduction is expected to be
234 greatest on tests with the smallest packet sizes (greatest header
237 **Metrics Collected**:
239 The following are the metrics collected for this test:
241 - The maximum forwarding rate in Frames Per Second (FPS) and Mbps of
242 the DUT for each frame size with X% packet loss and packet
243 modification operations being performed by the DUT.
244 - The average latency of the traffic flow when passing through the DUT
245 (if testing for latency, note that this average is different from the
246 test specified in Section 26.3 of
247 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__).
248 - The `RFC5481 <https://www.rfc-editor.org/rfc/rfc5481.txt>`__
249 PDV form of delay variation on the traffic flow,
250 using the 99th percentile.
251 - CPU and memory utilization may also be collected as part of this
252 test, to determine the vSwitch's performance footprint on the system.
256 Test ID: LTD.Throughput.RFC2544.Profile
257 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
259 **Title**: RFC 2544 Throughput and Latency Profile
261 **Prerequisite Test**: N/A
267 This test reveals how throughput and latency degrades as the offered
268 rate varies in the region of the DUT's maximum forwarding rate as
269 determined by LTD.Throughput.RFC2544.PacketLossRatio (0% Packet Loss).
270 For example it can be used to determine if the degradation of throughput
271 and latency as the offered rate increases is slow and graceful or sudden
274 The selected frame sizes are those previously defined under
275 :ref:`default-test-parameters`.
277 The offered traffic rate is described as a percentage delta with respect
278 to the DUT's RFC 2544 Throughput as determined by
279 LTD.Throughput.RFC2544.PacketLoss Ratio (0% Packet Loss case). A delta
280 of 0% is equivalent to an offered traffic rate equal to the RFC 2544
281 Maximum Throughput; A delta of +50% indicates an offered rate half-way
282 between the Maximum RFC2544 Throughput and line-rate, whereas a delta of
283 -50% indicates an offered rate of half the RFC 2544 Maximum Throughput.
284 Therefore the range of the delta figure is natuarlly bounded at -100%
285 (zero offered traffic) and +100% (traffic offered at line rate).
287 The following deltas to the maximum forwarding rate should be applied:
289 - -50%, -10%, 0%, +10% & +50%
291 **Expected Result**: For each packet size a profile should be produced
292 of how throughput and latency vary with offered rate.
294 **Metrics Collected**:
296 The following are the metrics collected for this test:
298 - The forwarding rate in Frames Per Second (FPS) and Mbps of the DUT
299 for each delta to the maximum forwarding rate and for each frame
301 - The average latency for each delta to the maximum forwarding rate and
303 - CPU and memory utilization may also be collected as part of this
304 test, to determine the vSwitch's performance footprint on the system.
305 - Any failures experienced (for example if the vSwitch crashes, stops
306 processing packets, restarts or becomes unresponsive to commands)
307 when the offered load is above Maximum Throughput MUST be recorded
308 and reported with the results.
312 Test ID: LTD.Throughput.RFC2544.SystemRecoveryTime
313 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
315 **Title**: RFC 2544 System Recovery Time Test
317 **Prerequisite Test** LTD.Throughput.RFC2544.PacketLossRatio
323 The aim of this test is to determine the length of time it takes the DUT
324 to recover from an overload condition for a constant load (fixed length
325 frames at a fixed interval time). The selected frame sizes are those
326 previously defined under :ref:`default-test-parameters`,
327 traffic should be sent to the DUT under normal conditions. During the
328 duration of the test and while the traffic flows are passing though the
329 DUT, at least one situation leading to an overload condition for the DUT
330 should occur. The time from the end of the overload condition to when
331 the DUT returns to normal operations should be measured to determine
332 recovery time. Prior to overloading the DUT, one should record the
333 average latency for 10,000 packets forwarded through the DUT.
335 The overload condition SHOULD be to transmit traffic at a very high
336 frame rate to the DUT (150% of the maximum 0% packet loss rate as
337 determined by LTD.Throughput.RFC2544.PacketLossRatio or line-rate
338 whichever is lower), for at least 60 seconds, then reduce the frame rate
339 to 75% of the maximum 0% packet loss rate. A number of time-stamps
340 should be recorded: - Record the time-stamp at which the frame rate was
341 reduced and record a second time-stamp at the time of the last frame
342 lost. The recovery time is the difference between the two timestamps. -
343 Record the average latency for 10,000 frames after the last frame loss
344 and continue to record average latency measurements for every 10,000
345 frames, when latency returns to within 10% of pre-overload levels record
350 **Metrics collected**
352 The following are the metrics collected for this test:
354 - The length of time it takes the DUT to recover from an overload
356 - The length of time it takes the DUT to recover the average latency to
357 pre-overload conditions.
359 **Deployment scenario**:
361 - Physical → virtual switch → physical.
365 .. _BackToBackFrames:
367 Test ID: LTD.Throughput.RFC2544.BackToBackFrames
368 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
370 **Title**: RFC2544 Back To Back Frames Test
372 **Prerequisite Test**: N
378 The aim of this test is to characterize the ability of the DUT to
379 process back-to-back frames. For each frame size previously defined
380 under :ref:`default-test-parameters`, a burst of traffic
381 is sent to the DUT with the minimum inter-frame gap between each frame.
382 If the number of received frames equals the number of frames that were
383 transmitted, the burst size should be increased and traffic is sent to
384 the DUT again. The value measured is the back-to-back value, that is the
385 maximum burst size the DUT can handle without any frame loss. Please note
386 a trial must run for a minimum of 2 seconds and should be repeated 50
387 times (at a minimum).
391 Tests of back-to-back frames with physical devices have produced
392 unstable results in some cases. All tests should be repeated in multiple
393 test sessions and results stability should be examined.
395 **Metrics collected**
397 The following are the metrics collected for this test:
399 - The average back-to-back value across the trials, which is
400 the number of frames in the longest burst that the DUT will
401 handle without the loss of any frames.
402 - CPU and memory utilization may also be collected as part of this
403 test, to determine the vSwitch's performance footprint on the system.
405 **Deployment scenario**:
407 - Physical → virtual switch → physical.
411 Test ID: LTD.Throughput.RFC2889.MaxForwardingRateSoak
412 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
414 **Title**: RFC 2889 X% packet loss Max Forwarding Rate Soak Test
416 **Prerequisite Test** LTD.Throughput.RFC2544.PacketLossRatio
422 The aim of this test is to understand the Max Forwarding Rate stability
423 over an extended test duration in order to uncover any outliers. To allow
424 for an extended test duration, the test should ideally run for 24 hours
425 or, if this is not possible, for at least 6 hours. For this test, each frame
426 size must be sent at the highest Throughput rate with X% packet loss, as
427 determined in the prerequisite test. The default loss percentages to be
428 tested are: - X = 0% - X = 10^-7%
430 Note: Other values can be tested if required by the user.
434 **Metrics Collected**:
436 The following are the metrics collected for this test:
438 - Max Forwarding Rate stability of the DUT.
440 - This means reporting the number of packets lost per time interval
441 and reporting any time intervals with packet loss. The
442 `RFC2889 <https://www.rfc-editor.org/rfc/rfc2289.txt>`__
443 Forwarding Rate shall be measured in each interval.
444 An interval of 60s is suggested.
446 - CPU and memory utilization may also be collected as part of this
447 test, to determine the vSwitch's performance footprint on the system.
448 - The `RFC5481 <https://www.rfc-editor.org/rfc/rfc5481.txt>`__
449 PDV form of delay variation on the traffic flow,
450 using the 99th percentile.
454 Test ID: LTD.Throughput.RFC2889.MaxForwardingRateSoakFrameModification
455 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
457 **Title**: RFC 2889 Max Forwarding Rate Soak Test with Frame Modification
459 **Prerequisite Test**:
460 LTD.Throughput.RFC2544.PacketLossRatioFrameModification (0% Packet Loss)
466 The aim of this test is to understand the Max Forwarding Rate stability over an
467 extended test duration in order to uncover any outliers. To allow for an
468 extended test duration, the test should ideally run for 24 hours or, if
469 this is not possible, for at least 6 hour. For this test, each frame
470 size must be sent at the highest Throughput rate with 0% packet loss, as
471 determined in the prerequisite test.
473 During this test, the DUT must perform the following operations on the
476 - Perform packet parsing on the DUT's ingress port.
477 - Perform any relevant address look-ups on the DUT's ingress ports.
478 - Modify the packet header before forwarding the packet to the DUT's
479 egress port. Packet modifications include:
481 - Modifying the Ethernet source or destination MAC address.
482 - Modifying/adding a VLAN tag (**Recommended**).
483 - Modifying/adding a MPLS tag.
484 - Modifying the source or destination ip address.
485 - Modifying the TOS/DSCP field.
486 - Modifying the source or destination ports for UDP/TCP/SCTP.
491 **Metrics Collected**:
493 The following are the metrics collected for this test:
495 - Max Forwarding Rate stability of the DUT.
497 - This means reporting the number of packets lost per time interval
498 and reporting any time intervals with packet loss. The
499 `RFC2889 <https://www.rfc-editor.org/rfc/rfc2289.txt>`__
500 Forwarding Rate shall be measured in each interval.
501 An interval of 60s is suggested.
503 - CPU and memory utilization may also be collected as part of this
504 test, to determine the vSwitch's performance footprint on the system.
505 - The `RFC5481 <https://www.rfc-editor.org/rfc/rfc5481.txt>`__
506 PDV form of delay variation on the traffic flow, using the 99th
511 Test ID: LTD.Throughput.RFC6201.ResetTime
512 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
514 **Title**: RFC 6201 Reset Time Test
516 **Prerequisite Test**: N/A
522 The aim of this test is to determine the length of time it takes the DUT
523 to recover from a reset.
525 Two reset methods are defined - planned and unplanned. A planned reset
526 requires stopping and restarting the virtual switch by the usual
527 'graceful' method defined by it's documentation. An unplanned reset
528 requires simulating a fatal internal fault in the virtual switch - for
529 example by using kill -SIGKILL on a Linux environment.
531 Both reset methods SHOULD be exercised.
533 For each frame size previously defined under :ref:`default-test-parameters`,
534 traffic should be sent to the DUT under
535 normal conditions. During the duration of the test and while the traffic
536 flows are passing through the DUT, the DUT should be reset and the Reset
537 time measured. The Reset time is the total time that a device is
538 determined to be out of operation and includes the time to perform the
539 reset and the time to recover from it (cf. `RFC6201
540 <https://www.rfc-editor.org/rfc/rfc6201.txt>`__).
542 `RFC6201 <https://www.rfc-editor.org/rfc/rfc6201.txt>`__ defines two methods
543 to measure the Reset time:
545 - Frame-Loss Method: which requires the monitoring of the number of
546 lost frames and calculates the Reset time based on the number of
547 frames lost and the offered rate according to the following
550 .. code-block:: console
552 Frames_lost (packets)
553 Reset_time = -------------------------------------
554 Offered_rate (packets per second)
556 - Timestamp Method: which measures the time from which the last frame
557 is forwarded from the DUT to the time the first frame is forwarded
558 after the reset. This involves time-stamping all transmitted frames
559 and recording the timestamp of the last frame that was received prior
560 to the reset and also measuring the timestamp of the first frame that
561 is received after the reset. The Reset time is the difference between
562 these two timestamps.
564 According to `RFC6201 <https://www.rfc-editor.org/rfc/rfc6201.txt>`__ the
565 choice of method depends on the test tool's capability; the Frame-Loss
566 method SHOULD be used if the test tool supports:
568 * Counting the number of lost frames per stream.
569 * Transmitting test frame despite the physical link status.
571 whereas the Timestamp method SHOULD be used if the test tool supports:
573 * Timestamping each frame.
574 * Monitoring received frame's timestamp.
575 * Transmitting frames only if the physical link status is up.
579 **Metrics collected**
581 The following are the metrics collected for this test:
583 * Average Reset Time over the number of trials performed.
585 Results of this test should include the following information:
587 * The reset method used.
588 * Throughput in Fps and Mbps.
589 * Average Frame Loss over the number of trials performed.
590 * Average Reset Time in milliseconds over the number of trials performed.
591 * Number of trials performed.
592 * Protocol: IPv4, IPv6, MPLS, etc.
593 * Frame Size in Octets
594 * Port Media: Ethernet, Gigabit Ethernet (GbE), etc.
595 * Port Speed: 10 Gbps, 40 Gbps etc.
596 * Interface Encapsulation: Ethernet, Ethernet VLAN, etc.
598 **Deployment scenario**:
600 * Physical → virtual switch → physical.
604 Test ID: LTD.Throughput.RFC2889.MaxForwardingRate
605 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
607 **Title**: RFC2889 Forwarding Rate Test
609 **Prerequisite Test**: LTD.Throughput.RFC2544.PacketLossRatio
615 This test measures the DUT's Max Forwarding Rate when the Offered Load
616 is varied between the throughput and the Maximum Offered Load for fixed
617 length frames at a fixed time interval. The selected frame sizes are
618 those previously defined under :ref:`default-test-parameters`.
619 The throughput is the maximum offered
620 load with 0% frame loss (measured by the prerequisite test), and the
621 Maximum Offered Load (as defined by
622 `RFC2285 <https://www.rfc-editor.org/rfc/rfc2285.txt>`__) is *"the highest
623 number of frames per second that an external source can transmit to a
624 DUT/SUT for forwarding to a specified output interface or interfaces"*.
626 Traffic should be sent to the DUT at a particular rate (TX rate)
627 starting with TX rate equal to the throughput rate. The rate of
628 successfully received frames at the destination counted (in FPS). If the
629 RX rate is equal to the TX rate, the TX rate should be increased by a
630 fixed step size and the RX rate measured again until the Max Forwarding
633 The trial duration for each iteration should last for the period of time
634 needed for the system to reach steady state for the frame size being
635 tested. Under `RFC2889 <https://www.rfc-editor.org/rfc/rfc2289.txt>`__
636 (Sec. 5.6.3.1) test methodology, the test
637 duration should run for a minimum period of 30 seconds, regardless
638 whether the system reaches steady state before the minimum duration
641 **Expected Result**: According to
642 `RFC2889 <https://www.rfc-editor.org/rfc/rfc2289.txt>`__ The Max Forwarding
643 Rate is the highest forwarding rate of a DUT taken from an iterative set of
644 forwarding rate measurements. The iterative set of forwarding rate measurements
645 are made by setting the intended load transmitted from an external source and
646 measuring the offered load (i.e what the DUT is capable of forwarding). If the
647 Throughput == the Maximum Offered Load, it follows that Max Forwarding Rate is
648 equal to the Maximum Offered Load.
650 **Metrics Collected**:
652 The following are the metrics collected for this test:
654 - The Max Forwarding Rate for the DUT for each packet size.
655 - CPU and memory utilization may also be collected as part of this
656 test, to determine the vSwitch's performance footprint on the system.
658 **Deployment scenario**:
660 - Physical → virtual switch → physical. Note: Full mesh tests with
661 multiple ingress and egress ports are a key aspect of RFC 2889
662 benchmarks, and scenarios with both 2 and 4 ports should be tested.
663 In any case, the number of ports used must be reported.
667 Test ID: LTD.Throughput.RFC2889.ForwardPressure
668 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
670 **Title**: RFC2889 Forward Pressure Test
672 **Prerequisite Test**: LTD.Throughput.RFC2889.MaxForwardingRate
678 The aim of this test is to determine if the DUT transmits frames with an
679 inter-frame gap that is less than 12 bytes. This test overloads the DUT
680 and measures the output for forward pressure. Traffic should be
681 transmitted to the DUT with an inter-frame gap of 11 bytes, this will
682 overload the DUT by 1 byte per frame. The forwarding rate of the DUT
685 **Expected Result**: The forwarding rate should not exceed the maximum
686 forwarding rate of the DUT collected by
687 LTD.Throughput.RFC2889.MaxForwardingRate.
689 **Metrics collected**
691 The following are the metrics collected for this test:
693 - Forwarding rate of the DUT in FPS or Mbps.
694 - CPU and memory utilization may also be collected as part of this
695 test, to determine the vSwitch's performance footprint on the system.
697 **Deployment scenario**:
699 - Physical → virtual switch → physical.
703 Test ID: LTD.Throughput.RFC2889.ErrorFramesFiltering
704 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
706 **Title**: RFC2889 Error Frames Filtering Test
708 **Prerequisite Test**: N/A
714 The aim of this test is to determine whether the DUT will propagate any
715 erroneous frames it receives or whether it is capable of filtering out
716 the erroneous frames. Traffic should be sent with erroneous frames
717 included within the flow at random intervals. Illegal frames that must
718 be tested include: - Oversize Frames. - Undersize Frames. - CRC Errored
719 Frames. - Dribble Bit Errored Frames - Alignment Errored Frames
721 The traffic flow exiting the DUT should be recorded and checked to
722 determine if the erroneous frames where passed through the DUT.
724 **Expected Result**: Broken frames are not passed!
726 **Metrics collected**
728 No Metrics are collected in this test, instead it determines:
730 - Whether the DUT will propagate erroneous frames.
731 - Or whether the DUT will correctly filter out any erroneous frames
732 from traffic flow with out removing correct frames.
734 **Deployment scenario**:
736 - Physical → virtual switch → physical.
740 Test ID: LTD.Throughput.RFC2889.BroadcastFrameForwarding
741 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
743 **Title**: RFC2889 Broadcast Frame Forwarding Test
745 **Prerequisite Test**: N
751 The aim of this test is to determine the maximum forwarding rate of the
752 DUT when forwarding broadcast traffic. For each frame previously defined
753 under :ref:`default-test-parameters`, the traffic should
754 be set up as broadcast traffic. The traffic throughput of the DUT should
757 The test should be conducted with at least 4 physical ports on the DUT.
758 The number of ports used MUST be recorded.
760 As broadcast involves forwarding a single incoming packet to several
761 destinations, the latency of a single packet is defined as the average
762 of the latencies for each of the broadcast destinations.
764 The incoming packet is transmitted on each of the other physical ports,
765 it is not transmitted on the port on which it was received. The test MAY
766 be conducted using different broadcasting ports to uncover any
767 performance differences.
771 **Metrics collected**:
773 The following are the metrics collected for this test:
775 - The forwarding rate of the DUT when forwarding broadcast traffic.
776 - The minimum, average & maximum packets latencies observed.
778 **Deployment scenario**:
780 - Physical → virtual switch 3x physical. In the Broadcast rate testing,
781 four test ports are required. One of the ports is connected to the test
782 device, so it can send broadcast frames and listen for miss-routed frames.
786 Test ID: LTD.Throughput.RFC2544.WorstN-BestN
787 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
789 **Title**: Modified RFC 2544 X% packet loss ratio Throughput and Latency Test
791 **Prerequisite Test**: N/A
797 This test determines the DUT's maximum forwarding rate with X% traffic
798 loss for a constant load (fixed length frames at a fixed interval time).
799 The default loss percentages to be tested are: X = 0%, X = 10^-7%
801 Modified RFC 2544 throughput benchmarking methodology aims to quantify
802 the throughput measurement variations observed during standard RFC 2544
803 benchmarking measurements of virtual switches and VNFs. The RFC2544
804 binary search algorithm is modified to use more samples per test trial
805 to drive the binary search and yield statistically more meaningful
806 results. This keeps the heart of the RFC2544 methodology, still relying
807 on the binary search of throughput at specified loss tolerance, while
808 providing more useful information about the range of results seen in
809 testing. Instead of using a single traffic trial per iteration step,
810 each traffic trial is repeated N times and the success/failure of the
811 iteration step is based on these N traffic trials. Two types of revised
812 tests are defined - *Worst-of-N* and *Best-of-N*.
816 *Worst-of-N* indicates the lowest expected maximum throughput for (
817 packet size, loss tolerance) when repeating the test.
819 1. Repeat the same test run N times at a set packet rate, record each
821 2. Take the WORST result (highest packet loss) out of N result samples,
822 called the Worst-of-N sample.
823 3. If Worst-of-N sample has loss less than the set loss tolerance, then
824 the step is successful - increase the test traffic rate.
825 4. If Worst-of-N sample has loss greater than the set loss tolerance
826 then the step failed - decrease the test traffic rate.
831 *Best-of-N* indicates the highest expected maximum throughput for (
832 packet size, loss tolerance) when repeating the test.
834 1. Repeat the same traffic run N times at a set packet rate, record
836 2. Take the BEST result (least packet loss) out of N result samples,
837 called the Best-of-N sample.
838 3. If Best-of-N sample has loss less than the set loss tolerance, then
839 the step is successful - increase the test traffic rate.
840 4. If Best-of-N sample has loss greater than the set loss tolerance,
841 then the step failed - decrease the test traffic rate.
844 Performing both Worst-of-N and Best-of-N benchmark tests yields lower
845 and upper bounds of expected maximum throughput under the operating
846 conditions, giving a very good indication to the user of the
847 deterministic performance range for the tested setup.
849 **Expected Result**: At the end of each trial series, the presence or
850 absence of loss determines the modification of offered load for the
851 next trial series, converging on a maximum rate, or
852 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__ Throughput
854 The Throughput load is re-used in related
855 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__ tests and other
858 **Metrics Collected**:
860 The following are the metrics collected for this test:
862 - The maximum forwarding rate in Frames Per Second (FPS) and Mbps of
863 the DUT for each frame size with X% packet loss.
864 - The average latency of the traffic flow when passing through the DUT
865 (if testing for latency, note that this average is different from the
866 test specified in Section 26.3 of
867 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__).
868 - Following may also be collected as part of this test, to determine
869 the vSwitch's performance footprint on the system:
871 - CPU core utilization.
872 - CPU cache utilization.
874 - System bus (QPI, PCI, ...) utilization.
875 - CPU cycles consumed per packet.
879 Test ID: LTD.Throughput.Overlay.Network.<tech>.RFC2544.PacketLossRatio
880 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
882 **Title**: <tech> Overlay Network RFC 2544 X% packet loss ratio Throughput and Latency Test
885 NOTE: Throughout this test, four interchangeable overlay technologies are covered by the
886 same test description. They are: VXLAN, GRE, NVGRE and GENEVE.
888 **Prerequisite Test**: N/A
893 This test evaluates standard switch performance benchmarks for the scenario where an
894 Overlay Network is deployed for all paths through the vSwitch. Overlay Technologies covered
895 (replacing <tech> in the test name) include:
902 Performance will be assessed for each of the following overlay network functions:
905 - De-encapsulation only
906 - Both Encapsulation and De-encapsulation
908 For each native packet, the DUT must perform the following operations:
910 - Examine the packet and classify its correct overlay net (tunnel) assignment
911 - Encapsulate the packet
912 - Switch the packet to the correct port
914 For each encapsulated packet, the DUT must perform the following operations:
916 - Examine the packet and classify its correct native network assignment
917 - De-encapsulate the packet, if required
918 - Switch the packet to the correct port
920 The selected frame sizes are those previously defined under
921 :ref:`default-test-parameters`.
923 Thus, each test comprises an overlay technology, a network function,
924 and a packet size *with* overlay network overhead included
925 (but see also the discussion at
926 https://etherpad.opnfv.org/p/vSwitchTestsDrafts ).
928 The test can also be used to determine the average latency of the traffic.
930 Under the `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__
931 test methodology, the test duration will
932 include a number of trials; each trial should run for a minimum period
933 of 60 seconds. A binary search methodology must be applied for each
934 trial to obtain the final result for Throughput.
936 **Expected Result**: At the end of each trial, the presence or absence
937 of loss determines the modification of offered load for the next trial,
938 converging on a maximum rate, or
939 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__ Throughput with X%
940 loss (where the value of X is typically equal to zero).
941 The Throughput load is re-used in related
942 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__ tests and other
945 **Metrics Collected**:
946 The following are the metrics collected for this test:
948 - The maximum Throughput in Frames Per Second (FPS) and Mbps of
949 the DUT for each frame size with X% packet loss.
950 - The average latency of the traffic flow when passing through the DUT
951 and VNFs (if testing for latency, note that this average is different from the
952 test specified in Section 26.3 of
953 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__).
954 - CPU and memory utilization may also be collected as part of this
955 test, to determine the vSwitch's performance footprint on the system.
959 Test ID: LTD.Throughput.RFC2544.MatchAction.PacketLossRatio
960 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
962 **Title**: RFC 2544 X% packet loss ratio match action Throughput and Latency Test
964 **Prerequisite Test**: LTD.Throughput.RFC2544.PacketLossRatio
970 The aim of this test is to determine the cost of carrying out match
971 action(s) on the DUT’s RFC2544 Throughput with X% traffic loss for
972 a constant load (fixed length frames at a fixed interval time).
974 Each test case requires:
976 * selection of a specific match action(s),
977 * specifying a percentage of total traffic that is elligible
978 for the match action,
979 * determination of the specific test configuration (number
980 of flows, number of test ports, presence of an external
981 controller, etc.), and
982 * measurement of the RFC 2544 Throughput level with X% packet
983 loss: Traffic shall be bi-directional and symmetric.
985 Note: It would be ideal to verify that all match action-elligible
986 traffic was forwarded to the correct port, and if forwarded to
987 an unintended port it should be considered lost.
989 A match action is an action that is typically carried on a frame
990 or packet that matches a set of flow classification parameters
991 (typically frame/packet header fields). A match action may or may
992 not modify a packet/frame. Match actions include [1]:
994 * output : outputs a packet to a particular port.
995 * normal: Subjects the packet to traditional L2/L3 processing
997 * flood: Outputs the packet on all switch physical ports
998 other than the port on which it was received and any ports
999 on which flooding is disabled.
1000 * all: Outputs the packet on all switch physical ports other
1001 than the port on which it was received.
1002 * local: Outputs the packet on the ``local port``, which
1003 corresponds to the network device that has the same name as
1005 * in_port: Outputs the packet on the port from which it was
1007 * Controller: Sends the packet and its metadata to the
1008 OpenFlow controller as a ``packet in`` message.
1009 * enqueue: Enqueues the packet on the specified queue
1011 * drop: discard the packet.
1013 Modifications include [1]:
1015 * mod vlan: covered by LTD.Throughput.RFC2544.PacketLossRatioFrameModification
1016 * mod_dl_src: Sets the source Ethernet address.
1017 * mod_dl_dst: Sets the destination Ethernet address.
1018 * mod_nw_src: Sets the IPv4 source address.
1019 * mod_nw_dst: Sets the IPv4 destination address.
1020 * mod_tp_src: Sets the TCP or UDP or SCTP source port.
1021 * mod_tp_dst: Sets the TCP or UDP or SCTP destination port.
1022 * mod_nw_tos: Sets the DSCP bits in the IPv4 ToS/DSCP or
1023 IPv6 traffic class field.
1024 * mod_nw_ecn: Sets the ECN bits in the appropriate IPv4 or
1026 * mod_nw_ttl: Sets the IPv4 TTL or IPv6 hop limit field.
1028 Note: This comprehensive list requires extensive traffic generator
1031 The match action(s) that were applied as part of the test should be
1032 reported in the final test report.
1034 During this test, the DUT must perform the following operations on
1037 * Perform packet parsing on the DUT’s ingress port.
1038 * Perform any relevant address look-ups on the DUT’s ingress
1040 * Carry out one or more of the match actions specified above.
1042 The default loss percentages to be tested are: - X = 0% - X = 10^-7%
1043 Other values can be tested if required by the user. The selected
1044 frame sizes are those previously defined under
1045 :ref:`default-test-parameters`.
1047 The test can also be used to determine the average latency of the
1048 traffic when a match action is applied to packets in a flow. Under
1049 the RFC2544 test methodology, the test duration will include a
1050 number of trials; each trial should run for a minimum period of 60
1051 seconds. A binary search methodology must be applied for each
1052 trial to obtain the final result.
1054 **Expected Result:**
1056 At the end of each trial, the presence or absence of loss
1057 determines the modification of offered load for the next trial,
1058 converging on a maximum rate, or RFC2544Throughput with X% loss.
1059 The Throughput load is re-used in related RFC2544 tests and other
1062 **Metrics Collected:**
1064 The following are the metrics collected for this test:
1066 * The RFC 2544 Throughput in Frames Per Second (FPS) and Mbps
1067 of the DUT for each frame size with X% packet loss.
1068 * The average latency of the traffic flow when passing through
1069 the DUT (if testing for latency, note that this average is
1070 different from the test specified in Section 26.3 ofRFC2544).
1071 * CPU and memory utilization may also be collected as part of
1072 this test, to determine the vSwitch’s performance footprint
1075 The metrics collected can be compared to that of the prerequisite
1076 test to determine the cost of the match action(s) in the pipeline.
1078 **Deployment scenario**:
1080 - Physical → virtual switch → physical (and others are possible)
1082 [1] ovs-ofctl - administer OpenFlow switches
1083 [http://openvswitch.org/support/dist-docs/ovs-ofctl.8.txt ]
1088 Packet Latency tests
1089 --------------------
1091 These tests will measure the store and forward latency as well as the packet
1092 delay variation for various packet types through the virtual switch. The
1093 following list is not exhaustive but should indicate the type of tests
1094 that should be required. It is expected that more will be added.
1098 Test ID: LTD.PacketLatency.InitialPacketProcessingLatency
1099 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1101 **Title**: Initial Packet Processing Latency
1103 **Prerequisite Test**: N/A
1109 In some virtual switch architectures, the first packets of a flow will
1110 take the system longer to process than subsequent packets in the flow.
1111 This test determines the latency for these packets. The test will
1112 measure the latency of the packets as they are processed by the
1113 flow-setup-path of the DUT. There are two methods for this test, a
1114 recommended method and a nalternative method that can be used if it is
1115 possible to disable the fastpath of the virtual switch.
1117 Recommended method: This test will send 64,000 packets to the DUT, each
1118 belonging to a different flow. Average packet latency will be determined
1119 over the 64,000 packets.
1121 Alternative method: This test will send a single packet to the DUT after
1122 a fixed interval of time. The time interval will be equivalent to the
1123 amount of time it takes for a flow to time out in the virtual switch
1124 plus 10%. Average packet latency will be determined over 1,000,000
1127 This test is intended only for non-learning virtual switches; For learning
1128 virtual switches use RFC2889.
1130 For this test, only unidirectional traffic is required.
1132 **Expected Result**: The average latency for the initial packet of all
1133 flows should be greater than the latency of subsequent traffic.
1135 **Metrics Collected**:
1137 The following are the metrics collected for this test:
1139 - Average latency of the initial packets of all flows that are
1140 processed by the DUT.
1142 **Deployment scenario**:
1144 - Physical → Virtual Switch → Physical.
1148 Test ID: LTD.PacketDelayVariation.RFC3393.Soak
1149 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1151 **Title**: Packet Delay Variation Soak Test
1153 **Prerequisite Tests**: LTD.Throughput.RFC2544.PacketLossRatio (0% Packet Loss)
1159 The aim of this test is to understand the distribution of packet delay
1160 variation for different frame sizes over an extended test duration and
1161 to determine if there are any outliers. To allow for an extended test
1162 duration, the test should ideally run for 24 hours or, if this is not
1163 possible, for at least 6 hour. For this test, each frame size must be
1164 sent at the highest possible throughput with 0% packet loss, as
1165 determined in the prerequisite test.
1167 **Expected Result**:
1169 **Metrics Collected**:
1171 The following are the metrics collected for this test:
1173 - The packet delay variation value for traffic passing through the DUT.
1174 - The `RFC5481 <https://www.rfc-editor.org/rfc/rfc5481.txt>`__
1175 PDV form of delay variation on the traffic flow,
1176 using the 99th percentile, for each 60s interval during the test.
1177 - CPU and memory utilization may also be collected as part of this
1178 test, to determine the vSwitch's performance footprint on the system.
1185 The general aim of these tests is to understand the impact of large flow
1186 table size and flow lookups on throughput. The following list is not
1187 exhaustive but should indicate the type of tests that should be required.
1188 It is expected that more will be added.
1192 .. _Scalability0PacketLoss:
1194 Test ID: LTD.Scalability.Flows.RFC2544.0PacketLoss
1195 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1197 **Title**: RFC 2544 0% loss Flow Scalability throughput test
1199 **Prerequisite Test**: LTD.Throughput.RFC2544.PacketLossRatio, IF the
1200 delta Throughput between the single-flow RFC2544 test and this test with
1201 a variable number of flows is desired.
1207 The aim of this test is to measure how throughput changes as the number
1208 of flows in the DUT increases. The test will measure the throughput
1209 through the fastpath, as such the flows need to be installed on the DUT
1210 before passing traffic.
1212 For each frame size previously defined under :ref:`default-test-parameters`
1213 and for each of the following number of flows:
1222 - Max supported number of flows.
1224 This test will be conducted under two conditions following the
1225 establishment of all flows as required by RFC 2544, regarding the flow
1226 expiration time-out:
1228 1) The time-out never expires during each trial.
1230 2) The time-out expires for all flows periodically. This would require a
1231 short time-out compared with flow re-appearance for a small number of
1232 flows, and may not be possible for all flow conditions.
1234 The maximum 0% packet loss Throughput should be determined in a manner
1235 identical to LTD.Throughput.RFC2544.PacketLossRatio.
1237 **Expected Result**:
1239 **Metrics Collected**:
1241 The following are the metrics collected for this test:
1243 - The maximum number of frames per second that can be forwarded at the
1244 specified number of flows and the specified frame size, with zero
1249 Test ID: LTD.MemoryBandwidth.RFC2544.0PacketLoss.Scalability
1250 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1252 **Title**: RFC 2544 0% loss Memory Bandwidth Scalability test
1254 **Prerequisite Tests**: LTD.Throughput.RFC2544.PacketLossRatio, IF the
1255 delta Throughput between an undisturbed RFC2544 test and this test with
1256 the Throughput affected by cache and memory bandwidth contention is desired.
1262 The aim of this test is to understand how the DUT's performance is
1263 affected by cache sharing and memory bandwidth between processes.
1265 During the test all cores not used by the vSwitch should be running a
1266 memory intensive application. This application should read and write
1267 random data to random addresses in unused physical memory. The random
1268 nature of the data and addresses is intended to consume cache, exercise
1269 main memory access (as opposed to cache) and exercise all memory buses
1270 equally. Furthermore:
1272 - the ratio of reads to writes should be recorded. A ratio of 1:1
1274 - the reads and writes MUST be of cache-line size and be cache-line aligned.
1275 - in NUMA architectures memory access SHOULD be local to the core's node.
1276 Whether only local memory or a mix of local and remote memory is used
1278 - the memory bandwidth (reads plus writes) used per-core MUST be recorded;
1279 the test MUST be run with a per-core memory bandwidth equal to half the
1280 maximum system memory bandwidth divided by the number of cores. The test
1281 MAY be run with other values for the per-core memory bandwidth.
1282 - the test MAY also be run with the memory intensive application running
1285 Under these conditions the DUT's 0% packet loss throughput is determined
1286 as per LTD.Throughput.RFC2544.PacketLossRatio.
1288 **Expected Result**:
1290 **Metrics Collected**:
1292 The following are the metrics collected for this test:
1294 - The DUT's 0% packet loss throughput in the presence of cache sharing and
1295 memory bandwidth between processes.
1299 Test ID: LTD.Scalability.VNF.RFC2544.PacketLossRatio
1300 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1302 **Title**: VNF Scalability RFC 2544 X% packet loss ratio Throughput and
1305 **Prerequisite Test**: N/A
1311 This test determines the DUT's throughput rate with X% traffic loss for
1312 a constant load (fixed length frames at a fixed interval time) when the
1313 number of VNFs on the DUT increases. The default loss percentages
1314 to be tested are: - X = 0% - X = 10^-7% . The minimum number of
1315 VNFs to be tested are 3.
1317 Flow classification should be conducted with L2, L3 and L4 matching
1318 to understand the matching and scaling capability of the vSwitch. The
1319 matching fields which were used as part of the test should be reported
1320 as part of the benchmark report.
1322 The vSwitch is responsible for forwarding frames between the VNFs
1324 The SUT (vSwitch and VNF daisy chain) operation should be validated
1325 before running the test. This may be completed by running a burst or
1326 continuous stream of traffic through the SUT to ensure proper operation
1329 **Note**: The traffic rate used to validate SUT operation should be low
1330 enough not to stress the SUT.
1332 **Note**: Other values can be tested if required by the user.
1334 **Note**: The same VNF should be used in the "daisy chain" formation.
1335 Each addition of a VNF should be conducted in a new test setup (The DUT
1336 is brought down, then the DUT is brought up again). An atlernative approach
1337 would be to continue to add VNFs without bringing down the DUT. The
1338 approach used needs to be documented as part of the test report.
1340 The selected frame sizes are those previously defined under
1341 :ref:`default-test-parameters`.
1342 The test can also be used to determine the average latency of the traffic.
1344 Under the `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__
1345 test methodology, the test duration will
1346 include a number of trials; each trial should run for a minimum period
1347 of 60 seconds. A binary search methodology must be applied for each
1348 trial to obtain the final result for Throughput.
1350 **Expected Result**: At the end of each trial, the presence or absence
1351 of loss determines the modification of offered load for the next trial,
1352 converging on a maximum rate, or
1353 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__ Throughput with X%
1355 The Throughput load is re-used in related
1356 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__ tests and other
1359 If the test VNFs are rather light-weight in terms of processing, the test
1360 provides a view of multiple passes through the vswitch on logical
1361 interfaces. In other words, the test produces an optimistic count of
1362 daisy-chained VNFs, but the cumulative effect of traffic on the vSwitch is
1363 "real" (assuming that the vSwitch has some dedicated resources, and the
1364 effects on shared resources is understood).
1367 **Metrics Collected**:
1368 The following are the metrics collected for this test:
1370 - The maximum Throughput in Frames Per Second (FPS) and Mbps of
1371 the DUT for each frame size with X% packet loss.
1372 - The average latency of the traffic flow when passing through the DUT
1373 and VNFs (if testing for latency, note that this average is different from the
1374 test specified in Section 26.3 of
1375 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__).
1376 - CPU and memory utilization may also be collected as part of this
1377 test, to determine the vSwitch's performance footprint on the system.
1381 Test ID: LTD.Scalability.VNF.RFC2544.PacketLossProfile
1382 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1384 **Title**: VNF Scalability RFC 2544 Throughput and Latency Profile
1386 **Prerequisite Test**: N/A
1392 This test reveals how throughput and latency degrades as the number
1393 of VNFs increases and offered rate varies in the region of the DUT's
1394 maximum forwarding rate as determined by
1395 LTD.Throughput.RFC2544.PacketLossRatio (0% Packet Loss).
1396 For example it can be used to determine if the degradation of throughput
1397 and latency as the number of VNFs and offered rate increases is slow
1398 and graceful, or sudden and severe. The minimum number of VNFs to
1401 The selected frame sizes are those previously defined under
1402 :ref:`default-test-parameters`.
1404 The offered traffic rate is described as a percentage delta with respect
1405 to the DUT's RFC 2544 Throughput as determined by
1406 LTD.Throughput.RFC2544.PacketLoss Ratio (0% Packet Loss case). A delta
1407 of 0% is equivalent to an offered traffic rate equal to the RFC 2544
1408 Throughput; A delta of +50% indicates an offered rate half-way
1409 between the Throughput and line-rate, whereas a delta of
1410 -50% indicates an offered rate of half the maximum rate. Therefore the
1411 range of the delta figure is natuarlly bounded at -100% (zero offered
1412 traffic) and +100% (traffic offered at line rate).
1414 The following deltas to the maximum forwarding rate should be applied:
1416 - -50%, -10%, 0%, +10% & +50%
1418 **Note**: Other values can be tested if required by the user.
1420 **Note**: The same VNF should be used in the "daisy chain" formation.
1421 Each addition of a VNF should be conducted in a new test setup (The DUT
1422 is brought down, then the DUT is brought up again). An atlernative approach
1423 would be to continue to add VNFs without bringing down the DUT. The
1424 approach used needs to be documented as part of the test report.
1426 Flow classification should be conducted with L2, L3 and L4 matching
1427 to understand the matching and scaling capability of the vSwitch. The
1428 matching fields which were used as part of the test should be reported
1429 as part of the benchmark report.
1431 The SUT (vSwitch and VNF daisy chain) operation should be validated
1432 before running the test. This may be completed by running a burst or
1433 continuous stream of traffic through the SUT to ensure proper operation
1436 **Note**: the traffic rate used to validate SUT operation should be low
1437 enough not to stress the SUT
1439 **Expected Result**: For each packet size a profile should be produced
1440 of how throughput and latency vary with offered rate.
1442 **Metrics Collected**:
1444 The following are the metrics collected for this test:
1446 - The forwarding rate in Frames Per Second (FPS) and Mbps of the DUT
1447 for each delta to the maximum forwarding rate and for each frame
1449 - The average latency for each delta to the maximum forwarding rate and
1450 for each frame size.
1451 - CPU and memory utilization may also be collected as part of this
1452 test, to determine the vSwitch's performance footprint on the system.
1453 - Any failures experienced (for example if the vSwitch crashes, stops
1454 processing packets, restarts or becomes unresponsive to commands)
1455 when the offered load is above Maximum Throughput MUST be recorded
1456 and reported with the results.
1463 The general aim of these tests is to understand the capacity of the
1464 and speed with which the vswitch can accommodate new flows.
1468 Test ID: LTD.Activation.RFC2889.AddressCachingCapacity
1469 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1471 **Title**: RFC2889 Address Caching Capacity Test
1473 **Prerequisite Test**: N/A
1479 Please note this test is only applicable to virtual switches that are capable of
1480 MAC learning. The aim of this test is to determine the address caching
1481 capacity of the DUT for a constant load (fixed length frames at a fixed
1482 interval time). The selected frame sizes are those previously defined
1483 under :ref:`default-test-parameters`.
1485 In order to run this test the aging time, that is the maximum time the
1486 DUT will keep a learned address in its flow table, and a set of initial
1487 addresses, whose value should be >= 1 and <= the max number supported by
1488 the implementation must be known. Please note that if the aging time is
1489 configurable it must be longer than the time necessary to produce frames
1490 from the external source at the specified rate. If the aging time is
1491 fixed the frame rate must be brought down to a value that the external
1492 source can produce in a time that is less than the aging time.
1494 Learning Frames should be sent from an external source to the DUT to
1495 install a number of flows. The Learning Frames must have a fixed
1496 destination address and must vary the source address of the frames. The
1497 DUT should install flows in its flow table based on the varying source
1498 addresses. Frames should then be transmitted from an external source at
1499 a suitable frame rate to see if the DUT has properly learned all of the
1500 addresses. If there is no frame loss and no flooding, the number of
1501 addresses sent to the DUT should be increased and the test is repeated
1502 until the max number of cached addresses supported by the DUT
1505 **Expected Result**:
1507 **Metrics collected**:
1509 The following are the metrics collected for this test:
1511 - Number of cached addresses supported by the DUT.
1512 - CPU and memory utilization may also be collected as part of this
1513 test, to determine the vSwitch's performance footprint on the system.
1515 **Deployment scenario**:
1517 - Physical → virtual switch → 2 x physical (one receiving, one listening).
1521 Test ID: LTD.Activation.RFC2889.AddressLearningRate
1522 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1524 **Title**: RFC2889 Address Learning Rate Test
1526 **Prerequisite Test**: LTD.Memory.RFC2889.AddressCachingCapacity
1532 Please note this test is only applicable to virtual switches that are capable of
1533 MAC learning. The aim of this test is to determine the rate of address
1534 learning of the DUT for a constant load (fixed length frames at a fixed
1535 interval time). The selected frame sizes are those previously defined
1536 under :ref:`default-test-parameters`, traffic should be
1537 sent with each IPv4/IPv6 address incremented by one. The rate at which
1538 the DUT learns a new address should be measured. The maximum caching
1539 capacity from LTD.Memory.RFC2889.AddressCachingCapacity should be taken
1540 into consideration as the maximum number of addresses for which the
1541 learning rate can be obtained.
1543 **Expected Result**: It may be worthwhile to report the behaviour when
1544 operating beyond address capacity - some DUTs may be more friendly to
1545 new addresses than others.
1547 **Metrics collected**:
1549 The following are the metrics collected for this test:
1551 - The address learning rate of the DUT.
1553 **Deployment scenario**:
1555 - Physical → virtual switch → 2 x physical (one receiving, one listening).
1559 Coupling between control path and datapath Tests
1560 ------------------------------------------------
1562 The following tests aim to determine how tightly coupled the datapath
1563 and the control path are within a virtual switch. The following list
1564 is not exhaustive but should indicate the type of tests that should be
1565 required. It is expected that more will be added.
1569 Test ID: LTD.CPDPCouplingFlowAddition
1570 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1572 **Title**: Control Path and Datapath Coupling
1574 **Prerequisite Test**:
1580 The aim of this test is to understand how exercising the DUT's control
1581 path affects datapath performance.
1583 Initially a certain number of flow table entries are installed in the
1584 vSwitch. Then over the duration of an RFC2544 throughput test
1585 flow-entries are added and removed at the rates specified below. No
1586 traffic is 'hitting' these flow-entries, they are simply added and
1589 The test MUST be repeated with the following initial number of
1590 flow-entries installed: - < 10 - 1000 - 100,000 - 10,000,000 (or the
1591 maximum supported number of flow-entries)
1593 The test MUST be repeated with the following rates of flow-entry
1594 addition and deletion per second: - 0 - 1 (i.e. 1 addition plus 1
1595 deletion) - 100 - 10,000
1597 **Expected Result**:
1599 **Metrics Collected**:
1601 The following are the metrics collected for this test:
1603 - The maximum forwarding rate in Frames Per Second (FPS) and Mbps of
1605 - The average latency of the traffic flow when passing through the DUT
1606 (if testing for latency, note that this average is different from the
1607 test specified in Section 26.3 of
1608 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__).
1609 - CPU and memory utilization may also be collected as part of this
1610 test, to determine the vSwitch's performance footprint on the system.
1612 **Deployment scenario**:
1614 - Physical → virtual switch → physical.
1618 CPU and memory consumption
1619 --------------------------
1621 The following tests will profile a virtual switch's CPU and memory
1622 utilization under various loads and circumstances. The following
1623 list is not exhaustive but should indicate the type of tests that
1624 should be required. It is expected that more will be added.
1630 Test ID: LTD.Stress.RFC2544.0PacketLoss
1631 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1633 **Title**: RFC 2544 0% Loss CPU OR Memory Stress Test
1635 **Prerequisite Test**:
1641 The aim of this test is to understand the overall performance of the
1642 system when a CPU or Memory intensive application is run on the same DUT as
1643 the Virtual Switch. For each frame size, an
1644 LTD.Throughput.RFC2544.PacketLossRatio (0% Packet Loss) test should be
1645 performed. Throughout the entire test a CPU or Memory intensive application
1646 should be run on all cores on the system not in use by the Virtual Switch.
1647 For NUMA system only cores on the same NUMA node are loaded.
1649 It is recommended that stress-ng be used for loading the non-Virtual
1650 Switch cores but any stress tool MAY be used.
1652 **Expected Result**:
1654 **Metrics Collected**:
1656 The following are the metrics collected for this test:
1658 - Memory and CPU utilization of the cores running the Virtual Switch.
1659 - The number of identity of the cores allocated to the Virtual Switch.
1660 - The configuration of the stress tool (for example the command line
1661 parameters used to start it.)
1663 **Note:** Stress in the test ID can be replaced with the name of the
1664 component being stressed, when reporting the results:
1665 LTD.CPU.RFC2544.0PacketLoss or LTD.Memory.RFC2544.0PacketLoss
1669 Summary List of Tests
1670 ---------------------
1674 - Test ID: LTD.Throughput.RFC2544.PacketLossRatio
1675 - Test ID: LTD.Throughput.RFC2544.PacketLossRatioFrameModification
1676 - Test ID: LTD.Throughput.RFC2544.Profile
1677 - Test ID: LTD.Throughput.RFC2544.SystemRecoveryTime
1678 - Test ID: LTD.Throughput.RFC2544.BackToBackFrames
1679 - Test ID: LTD.Throughput.RFC2889.Soak
1680 - Test ID: LTD.Throughput.RFC2889.SoakFrameModification
1681 - Test ID: LTD.Throughput.RFC6201.ResetTime
1682 - Test ID: LTD.Throughput.RFC2889.MaxForwardingRate
1683 - Test ID: LTD.Throughput.RFC2889.ForwardPressure
1684 - Test ID: LTD.Throughput.RFC2889.ErrorFramesFiltering
1685 - Test ID: LTD.Throughput.RFC2889.BroadcastFrameForwarding
1686 - Test ID: LTD.Throughput.RFC2544.WorstN-BestN
1687 - Test ID: LTD.Throughput.Overlay.Network.<tech>.RFC2544.PacketLossRatio
1689 2. Packet Latency tests
1691 - Test ID: LTD.PacketLatency.InitialPacketProcessingLatency
1692 - Test ID: LTD.PacketDelayVariation.RFC3393.Soak
1694 3. Scalability tests
1696 - Test ID: LTD.Scalability.Flows.RFC2544.0PacketLoss
1697 - Test ID: LTD.MemoryBandwidth.RFC2544.0PacketLoss.Scalability
1698 - LTD.Scalability.VNF.RFC2544.PacketLossProfile
1699 - LTD.Scalability.VNF.RFC2544.PacketLossRatio
1703 - Test ID: LTD.Activation.RFC2889.AddressCachingCapacity
1704 - Test ID: LTD.Activation.RFC2889.AddressLearningRate
1706 5. Coupling between control path and datapath Tests
1708 - Test ID: LTD.CPDPCouplingFlowAddition
1710 6. CPU and memory consumption
1712 - Test ID: LTD.Stress.RFC2544.0PacketLoss