1 CHARACTERIZE VSWITCH PERFORMANCE FOR TELCO NFV USE CASES LEVEL TEST DESIGN
2 ==========================================================================
4 .. contents:: Table of Contents
9 The objective of the OPNFV project titled
10 **“Characterize vSwitch Performance for Telco NFV Use Cases”**, is to
11 evaluate a virtual switch to identify its suitability for a Telco
12 Network Function Virtualization (NFV) environment. The intention of this
13 Level Test Design (LTD) document is to specify the set of tests to carry
14 out in order to objectively measure the current characteristics of a
15 virtual switch in the Network Function Virtualization Infrastructure
16 (NFVI) as well as the test pass criteria. The detailed test cases will
17 be defined in `Section 2 <#DetailsOfTheLevelTestDesign>`__, preceded by
18 the `Document identifier <#DocId>`__ and the `Scope <#Scope>`__.
20 This document is currently in draft form.
22 1.1. Document identifier
23 ------------------------
25 The document id will be used to uniquely
26 identify versions of the LTD. The format for the document id will be:
27 OPNFV\_vswitchperf\_LTD\_ver\_NUM\_MONTH\_YEAR\_STATUS, where by the
28 status is one of: draft, reviewed, corrected or final. The document id
29 for this version of the LTD is:
30 OPNFV\_vswitchperf\_LTD\_ver\_1.6\_Jan\_15\_DRAFT.
35 The main purpose of this project is to specify a suite of
36 performance tests in order to objectively measure the current packet
37 transfer characteristics of a virtual switch in the NFVI. The intent of
38 the project is to facilitate testing of any virtual switch. Thus, a
39 generic suite of tests shall be developed, with no hard dependencies to
40 a single implementation. In addition, the test case suite shall be
41 architecture independent.
43 The test cases developed in this project shall not form part of a
44 separate test framework, all of these tests may be inserted into the
45 Continuous Integration Test Framework and/or the Platform Functionality
46 Test Framework - if a vSwitch becomes a standard component of an OPNFV
52 * `RFC 1242 Benchmarking Terminology for Network Interconnection
53 Devices <http://www.ietf.org/rfc/rfc1242.txt>`__
54 * `RFC 2544 Benchmarking Methodology for Network Interconnect
55 Devices <http://www.ietf.org/rfc/rfc2544.txt>`__
56 * `RFC 2285 Benchmarking Terminology for LAN Switching
57 Devices <http://www.ietf.org/rfc/rfc2285.txt>`__
58 * `RFC 2889 Benchmarking Methodology for LAN Switching
59 Devices <http://www.ietf.org/rfc/rfc2889.txt>`__
60 * `RFC 3918 Methodology for IP Multicast
61 Benchmarking <http://www.ietf.org/rfc/rfc3918.txt>`__
62 * `RFC 4737 Packet Reordering
63 Metrics <http://www.ietf.org/rfc/rfc4737.txt>`__
64 * `RFC 5481 Packet Delay Variation Applicability
65 Statement <http://www.ietf.org/rfc/rfc5481.txt>`__
66 * `RFC 6201 Device Reset
67 Characterization <http://tools.ietf.org/html/rfc6201>`__
69 2. Details of the Level Test Design
70 ===================================
72 This section describes the features to be tested (`cf. 2.1
73 <#FeaturesToBeTested>`__), the test approach (`cf. 2.2 <#Approach>`__);
74 it also identifies the sets of test cases or scenarios (`cf. 2.3
75 <#TestIdentification>`__) along with the pass/fail criteria (`cf. 2.4
76 <#PassFail>`__) and the test deliverables (`cf. 2.5 <#TestDeliverables>`__).
78 2.1. Features to be tested
79 --------------------------
81 Characterizing virtual switches (i.e. Device Under Test (DUT) in this document)
82 includes measuring the following performance metrics:
84 - **Throughput** as defined by `RFC1242
85 <https://www.rfc-editor.org/rfc/rfc1242.txt>`__: The maximum rate at which
86 **none** of the offered frames are dropped by the DUT. The maximum frame
87 rate and bit rate that can be transmitted by the DUT without any error
88 should be recorded. Note there is an equivalent bit rate and a specific
89 layer at which the payloads contribute to the bits. Errors and
90 improperly formed frames or packets are dropped.
91 - **Packet delay** introduced by the DUT and its cumulative effect on
92 E2E networks. Frame delay can be measured equivalently.
93 - **Packet delay variation**: measured from the perspective of the
94 VNF/application. Packet delay variation is sometimes called "jitter".
95 However, we will avoid the term "jitter" as the term holds different
96 meaning to different groups of people. In this document we will
97 simply use the term packet delay variation. The preferred form for this
98 metric is the PDV form of delay variation defined in `RFC5481
99 <https://www.rfc-editor.org/rfc/rfc5481.txt>`__. The most relevant
100 measurement of PDV considers the delay variation of a single user flow,
101 as this will be relevant to the size of end-system buffers to compensate
102 for delay variation. The measurement system's ability to store the
103 delays of individual packets in the flow of interest is a key factor
104 that determines the specific measurement method. At the outset, it is
105 ideal to view the complete PDV distribution. Systems that can capture
106 and store packets and their delays have the freedom to calculate the
107 reference minimum delay and to determine various quantiles of the PDV
108 distribution accurately (in post-measurement processing routines).
109 Systems without storage must apply algorithms to calculate delay and
110 statistical measurements on the fly. For example, a system may store
111 temporary estimates of the mimimum delay and the set of (100) packets
112 with the longest delays during measurement (to calculate a high quantile,
113 and update these sets with new values periodically.
114 In some cases, a limited number of delay histogram bins will be
115 available, and the bin limits will need to be set using results from
116 repeated experiments. See section 8 of `RFC5481
117 <https://www.rfc-editor.org/rfc/rfc5481.txt>`__.
118 - **Packet loss** (within a configured waiting time at the receiver): All
119 packets sent to the DUT should be accounted for.
120 - **Burst behaviour**: measures the ability of the DUT to buffer packets.
121 - **Packet re-ordering**: measures the ability of the device under test to
122 maintain sending order throughout transfer to the destination.
123 - **Packet correctness**: packets or Frames must be well-formed, in that
124 they include all required fields, conform to length requirements, pass
125 integrity checks, etc.
126 - **Availability and capacity** of the DUT i.e. when the DUT is fully “up”
129 - Includes power consumption of the CPU (in various power states) and
131 - Includes CPU utilization.
132 - Includes the number of NIC interfaces supported.
133 - Includes headroom of VM workload processing cores (i.e. available
140 In order to determine the packet transfer characteristics of a virtual
141 switch, the tests will be broken down into the following categories:
143 2.2.1 Test Categories
144 ----------------------
145 - **Throughput Tests** to measure the maximum forwarding rate (in
146 frames per second or fps) and bit rate (in Mbps) for a constant load
147 (as defined by `RFC1242 <https://www.rfc-editor.org/rfc/rfc1242.txt>`__)
148 without traffic loss.
149 - **Packet and Frame Delay Tests** to measure average, min and max
150 packet and frame delay for constant loads.
151 - **Stream Performance Tests** (TCP, UDP) to measure bulk data transfer
152 performance, i.e. how fast systems can send and receive data through
154 - **Request/Response Performance** Tests (TCP, UDP) the measure the
155 transaction rate through the virtual switch.
156 - **Packet Delay Tests** to understand latency distribution for
157 different packet sizes and over an extended test run to uncover
159 - **Scalability Tests** to understand how the virtual switch performs
160 as the number of flows, active ports, complexity of the forwarding
161 logic's configuration... it has to deal with increases.
162 - **Control Path and Datapath Coupling** Tests, to understand how
163 closely coupled the datapath and the control path are as well as the
164 effect of this coupling on the performance of the DUT.
165 - **CPU and Memory Consumption Tests** to understand the virtual
166 switch’s footprint on the system, this includes:
171 * Time To Establish Flows Tests.
173 - **Noisy Neighbour Tests**, to understand the effects of resource
174 sharing on the performance of a virtual switch.
176 **Note:** some of the tests above can be conducted simultaneously where
177 the combined results would be insightful, for example Packet/Frame Delay
180 2.2.2 Deployment Scenarios
181 --------------------------
182 The following represents possible deployments which can help to
183 determine the performance of both the virtual switch and the datapath
186 Physical port → vSwitch → physical port
187 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
188 .. code-block:: console
191 +--------------------------------------------------+ |
192 | +--------------------+ | |
195 | +--------------+ +--------------+ | |
196 | | phy port | vSwitch | phy port | | |
197 +---+--------------+------------+--------------+---+ _|
201 +--------------------------------------------------+
203 | traffic generator |
205 +--------------------------------------------------+
208 Physical port → vSwitch → VNF → vSwitch → physical port
209 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
210 .. code-block:: console
213 +---------------------------------------------------+ |
215 | +-------------------------------------------+ | |
216 | | Application | | |
217 | +-------------------------------------------+ | |
221 | +---------------+ +---------------+ | |
222 | | logical port 0| | logical port 1| | |
223 +---+---------------+-----------+---------------+---+ _|
227 +---+---------------+----------+---------------+---+ |
228 | | logical port 0| | logical port 1| | |
229 | +---------------+ +---------------+ | |
233 | +--------------+ +--------------+ | |
234 | | phy port | vSwitch | phy port | | |
235 +---+--------------+------------+--------------+---+ _|
239 +--------------------------------------------------+
241 | traffic generator |
243 +--------------------------------------------------+
246 Physical port → vSwitch → VNF → vSwitch → VNF → vSwitch → physical port
247 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
249 .. code-block:: console
252 +----------------------+ +----------------------+ |
253 | Guest 1 | | Guest 2 | |
254 | +---------------+ | | +---------------+ | |
255 | | Application | | | | Application | | |
256 | +---------------+ | | +---------------+ | |
258 | | v | | | v | | Guests
259 | +---------------+ | | +---------------+ | |
260 | | logical ports | | | | logical ports | | |
261 | | 0 1 | | | | 0 1 | | |
262 +---+---------------+--+ +---+---------------+--+ _|
266 +---+---------------+---------+---------------+--+ |
267 | | 0 1 | | 3 4 | | |
268 | | logical ports | | logical ports | | |
269 | +---------------+ +---------------+ | |
271 | | L-----------------+ v | |
272 | +--------------+ +--------------+ | |
273 | | phy ports | vSwitch | phy ports | | |
274 +---+--------------+----------+--------------+---+ _|
278 +--------------------------------------------------+
280 | traffic generator |
282 +--------------------------------------------------+
285 Physical port → vSwitch → VNF
286 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
288 .. code-block:: console
291 +---------------------------------------------------+ |
293 | +-------------------------------------------+ | |
294 | | Application | | |
295 | +-------------------------------------------+ | |
299 | +---------------+ | |
300 | | logical port 0| | |
301 +---+---------------+-------------------------------+ _|
305 +---+---------------+------------------------------+ |
306 | | logical port 0| | |
307 | +---------------+ | |
311 | +--------------+ | |
312 | | phy port | vSwitch | |
313 +---+--------------+------------ -------------- ---+ _|
317 +--------------------------------------------------+
319 | traffic generator |
321 +--------------------------------------------------+
323 VNF → vSwitch → physical port
324 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
326 .. code-block:: console
329 +---------------------------------------------------+ |
331 | +-------------------------------------------+ | |
332 | | Application | | |
333 | +-------------------------------------------+ | |
337 | +---------------+ | |
338 | | logical port | | |
339 +-------------------------------+---------------+---+ _|
343 +------------------------------+---------------+---+ |
344 | | logical port | | |
345 | +---------------+ | |
349 | +--------------+ | |
350 | vSwitch | phy port | | |
351 +-------------------------------+--------------+---+ _|
355 +--------------------------------------------------+
357 | traffic generator |
359 +--------------------------------------------------+
361 VNF → vSwitch → VNF → vSwitch
362 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
364 .. code-block:: console
367 +-------------------------+ +-------------------------+ |
368 | Guest 1 | | Guest 2 | |
369 | +-----------------+ | | +-----------------+ | |
370 | | Application | | | | Application | | |
371 | +-----------------+ | | +-----------------+ | |
375 | +---------------+ | | +---------------+ | |
376 | | logical port 0| | | | logical port 0| | |
377 +-----+---------------+---+ +---+---------------+-----+ _|
381 +----+---------------+------------+---------------+-----+ |
382 | | port 0 | | port 1 | | |
383 | +---------------+ +---------------+ | |
386 | +--------------------+ | |
389 +-------------------------------------------------------+ _|
391 HOST 1(Physical port → virtual switch → VNF → virtual switch →
392 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
393 Physical port) → HOST 2(Physical port → virtual switch → VNF →
394 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
395 virtual switch → Physical port)
396 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
398 .. code-block:: console
401 +----------------------+ +----------------------+ |
402 | Guest 1 | | Guest 2 | |
403 | +---------------+ | | +---------------+ | |
404 | | Application | | | | Application | | |
405 | +---------------+ | | +---------------+ | |
407 | | v | | | v | | Guests
408 | +---------------+ | | +---------------+ | |
409 | | logical ports | | | | logical ports | | |
410 | | 0 1 | | | | 0 1 | | |
411 +---+---------------+--+ +---+---------------+--+ _|
415 +---+---------------+--+ +---+---------------+--+ |
416 | | 0 1 | | | | 3 4 | | |
417 | | logical ports | | | | logical ports | | |
418 | +---------------+ | | +---------------+ | |
419 | ^ | | | ^ | | | Hosts
421 | +--------------+ | | +--------------+ | |
422 | | phy ports | | | | phy ports | | |
423 +---+--------------+---+ +---+--------------+---+ _|
425 | +-----------------+ |
427 +--------------------------------------------------+
429 | traffic generator |
431 +--------------------------------------------------+
435 **Note:** For tests where the traffic generator and/or measurement
436 receiver are implemented on VM and connected to the virtual switch
437 through vNIC, the issues of shared resources and interactions between
438 the measurement devices and the device under test must be considered.
440 **Note:** Some RFC 2889 tests require a full-mesh sending and receiving
441 pattern involving more than two ports. This possibility is illustrated in the
442 Physical port → vSwitch → VNF → vSwitch → VNF → vSwitch → physical port
443 diagram above (with 2 sending and 2 receiving ports, though all ports
444 could be used bi-directionally).
446 **Note:** When Deployment Scenarios are used in RFC 2889 address learning
447 or cache capacity testing, an additional port from the vSwitch must be
448 connected to the test device. This port is used to listen for flooded
451 2.2.3 General Methodology:
452 --------------------------
453 To establish the baseline performance of the virtual switch, tests would
454 initially be run with a simple workload in the VNF (the recommended
455 simple workload VNF would be `DPDK <http://www.dpdk.org/>`__'s testpmd
456 application forwarding packets in a VM or vloop\_vnf a simple kernel
457 module that forwards traffic between two network interfaces inside the
458 virtualized environment while bypassing the networking stack).
459 Subsequently, the tests would also be executed with a real Telco
460 workload running in the VNF, which would exercise the virtual switch in
461 the context of higher level Telco NFV use cases, and prove that its
462 underlying characteristics and behaviour can be measured and validated.
463 Suitable real Telco workload VNFs are yet to be identified.
465 2.2.3.1 Default Test Parameters
466 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
468 The following list identifies the default parameters for suite of
471 - Reference application: Simple forwarding or Open Source VNF.
472 - Frame size (bytes): 64, 128, 256, 512, 1024, 1280, 1518, 2K, 4k OR
473 Packet size based on use-case (e.g. RTP 64B, 256B) OR Mix of packet sizes as
474 maintained by the Functest project <https://wiki.opnfv.org/traffic_profile_management>.
475 - Reordering check: Tests should confirm that packets within a flow are
477 - Duplex: Unidirectional / Bidirectional. Default: Full duplex with
478 traffic transmitting in both directions, as network traffic generally
479 does not flow in a single direction. By default the data rate of
480 transmitted traffic should be the same in both directions, please
481 note that asymmetric traffic (e.g. downlink-heavy) tests will be
482 mentioned explicitly for the relevant test cases.
483 - Number of Flows: Default for non scalability tests is a single flow.
484 For scalability tests the goal is to test with maximum supported
485 flows but where possible will test up to 10 Million flows. Start with
486 a single flow and scale up. By default flows should be added
487 sequentially, tests that add flows simultaneously will explicitly
488 call out their flow addition behaviour. Packets are generated across
489 the flows uniformly with no burstiness.
490 - Traffic Types: UDP, SCTP, RTP, GTP and UDP traffic.
491 - Deployment scenarios are:
492 - Physical → virtual switch → physical.
493 - Physical → virtual switch → VNF → virtual switch → physical.
494 - Physical → virtual switch → VNF → virtual switch → VNF → virtual
496 - Physical → virtual switch → VNF.
497 - VNF → virtual switch → Physical.
498 - VNF → virtual switch → VNF.
500 Tests MUST have these parameters unless otherwise stated. **Test cases
501 with non default parameters will be stated explicitly**.
503 **Note**: For throughput tests unless stated otherwise, test
504 configurations should ensure that traffic traverses the installed flows
505 through the virtual switch, i.e. flows are installed and have an appropriate
506 time out that doesn't expire before packet transmission starts.
508 2.2.3.2 Flow Classification
509 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
511 Virtual switches classify packets into flows by processing and matching
512 particular header fields in the packet/frame and/or the input port where
513 the packets/frames arrived. The vSwitch then carries out an action on
514 the group of packets that match the classification parameters. Thus a
515 flow is considered to be a sequence of packets that have a shared set of
516 header field values or have arrived on the same port and have the same
517 action applied to them. Performance results can vary based on the
518 parameters the vSwitch uses to match for a flow. The recommended flow
519 classification parameters for L3 vSwitch performance tests are: the
520 input port, the source IP address, the destination IP address and the
521 Ethernet protocol type field. It is essential to increase the flow
522 time-out time on a vSwitch before conducting any performance tests that
523 do not measure the flow set-up time. Normally the first packet of a
524 particular flow will install the flow in the vSwitch which adds an
525 additional latency, subsequent packets of the same flow are not subject
526 to this latency if the flow is already installed on the vSwitch.
528 2.2.3.3 Test Priority
529 ~~~~~~~~~~~~~~~~~~~~~
531 Tests will be assigned a priority in order to determine which tests
532 should be implemented immediately and which tests implementations
535 Priority can be of following types: - Urgent: Must be implemented
536 immediately. - High: Must be implemented in the next release. - Medium:
537 May be implemented after the release. - Low: May or may not be
543 The SUT should be configured to its "default" state. The
544 SUT's configuration or set-up must not change between tests in any way
545 other than what is required to do the test. All supported protocols must
546 be configured and enabled for each test set up.
548 2.2.3.4.1 Port Configuration
549 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
551 The DUT should be configured with n ports where
552 n is a multiple of 2. Half of the ports on the DUT should be used as
553 ingress ports and the other half of the ports on the DUT should be used
554 as egress ports. Where a DUT has more than 2 ports, the ingress data
555 streams should be set-up so that they transmit packets to the egress
556 ports in sequence so that there is an even distribution of traffic
557 across ports. For example, if a DUT has 4 ports 0(ingress), 1(ingress),
558 2(egress) and 3(egress), the traffic stream directed at port 0 should
559 output a packet to port 2 followed by a packet to port 3. The traffic
560 stream directed at port 1 should also output a packet to port 2 followed
561 by a packet to port 3.
563 2.2.3.4.2 Frame Formats
564 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
566 Frame formats Layer 2 (data link layer) protocols
567 ++++++++++++++++++++++++++++++++++++++++++++++++++
570 .. code-block:: console
572 +---------------------+--------------------+-----------+
573 | Ethernet Header | Payload | Check Sum |
574 +---------------------+--------------------+-----------+
575 |_____________________|____________________|___________|
576 14 Bytes 46 - 1500 Bytes 4 Bytes
578 Layer 3 (network layer) protocols
579 ++++++++++++++++++++++++++++++++++
583 .. code-block:: console
585 +---------------------+--------------------+--------------------+-----------+
586 | Ethernet Header | IP Header | Payload | Check Sum |
587 +---------------------+--------------------+--------------------+-----------+
588 |_____________________|____________________|____________________|___________|
589 14 Bytes 20 bytes 26 - 1480 Bytes 4 Bytes
593 .. code-block:: console
595 +---------------------+--------------------+--------------------+-----------+
596 | Ethernet Header | IP Header | Payload | Check Sum |
597 +---------------------+--------------------+--------------------+-----------+
598 |_____________________|____________________|____________________|___________|
599 14 Bytes 40 bytes 26 - 1460 Bytes 4 Bytes
601 Layer 4 (transport layer) protocols
602 ++++++++++++++++++++++++++++++++++++
607 .. code-block:: console
609 +---------------------+--------------------+-----------------+--------------------+-----------+
610 | Ethernet Header | IP Header | Layer 4 Header | Payload | Check Sum |
611 +---------------------+--------------------+-----------------+--------------------+-----------+
612 |_____________________|____________________|_________________|____________________|___________|
613 14 Bytes 40 bytes 20 Bytes 6 - 1460 Bytes 4 Bytes
615 Layer 5 (application layer) protocols
616 +++++++++++++++++++++++++++++++++++++
620 .. code-block:: console
622 +---------------------+--------------------+-----------------+--------------------+-----------+
623 | Ethernet Header | IP Header | Layer 4 Header | Payload | Check Sum |
624 +---------------------+--------------------+-----------------+--------------------+-----------+
625 |_____________________|____________________|_________________|____________________|___________|
626 14 Bytes 20 bytes 20 Bytes Min 6 Bytes 4 Bytes
629 2.2.3.4.3 Packet Throughput
630 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
631 There is a difference between an Ethernet frame,
632 an IP packet, and a UDP datagram. In the seven-layer OSI model of
633 computer networking, packet refers to a data unit at layer 3 (network
634 layer). The correct term for a data unit at layer 2 (data link layer) is
635 a frame, and at layer 4 (transport layer) is a segment or datagram.
637 Important concepts related to 10GbE performance are frame rate and
638 throughput. The MAC bit rate of 10GbE, defined in the IEEE standard 802
639 .3ae, is 10 billion bits per second. Frame rate is based on the bit rate
640 and frame format definitions. Throughput, defined in IETF RFC 1242, is
641 the highest rate at which the system under test can forward the offered
644 The frame rate for 10GbE is determined by a formula that divides the 10
645 billion bits per second by the preamble + frame length + inter-frame
648 The maximum frame rate is calculated using the minimum values of the
649 following parameters, as described in the IEEE 802 .3ae standard:
651 - Preamble: 8 bytes \* 8 = 64 bits
652 - Frame Length: 64 bytes (minimum) \* 8 = 512 bits
653 - Inter-frame Gap: 12 bytes (minimum) \* 8 = 96 bits
655 Therefore, Maximum Frame Rate (64B Frames)
656 = MAC Transmit Bit Rate / (Preamble + Frame Length + Inter-frame Gap)
657 = 10,000,000,000 / (64 + 512 + 96)
658 = 10,000,000,000 / 672
659 = 14,880,952.38 frame per second (fps)
661 2.2.3.4.4 System isolation and validation
662 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
664 A key consideration when conducting any sort of benchmark is trying to
665 ensure the consistency and repeatability of test results between runs.
666 When benchmarking the performance of a virtual switch there are many
667 factors that can affect the consistency of results. This section
668 describes these factors and the measures that can be taken to limit
669 their effects. In addition, this section will outline some system tests
670 to validate the platform and the VNF before conducting any vSwitch
675 When conducting a benchmarking test on any SUT, it is essential to limit
676 (and if reasonable, eliminate) any noise that may interfere with the
677 accuracy of the metrics collected by the test. This noise may be
678 introduced by other hardware or software (OS, other applications), and
679 can result in significantly varying performance metrics being collected
680 between consecutive runs of the same test. In the case of characterizing
681 the performance of a virtual switch, there are a number of configuration
682 parameters that can help increase the repeatability and stability of
683 test results, including:
685 - OS/GRUB configuration:
687 - maxcpus = n where n >= 0; limits the kernel to using 'n'
688 processors. Only use exactly what you need.
689 - isolcpus: Isolate CPUs from the general scheduler. Isolate all
690 CPUs bar one which will be used by the OS.
691 - use taskset to affinitize the forwarding application and the VNFs
692 onto isolated cores. VNFs and the vSwitch should be allocated
693 their own cores, i.e. must not share the same cores. vCPUs for the
694 VNF should be affinitized to individual cores also.
695 - Limit the amount of background applications that are running and
696 set OS to boot to runlevel 3. Make sure to kill any unnecessary
697 system processes/daemons.
698 - Only enable hardware that you need to use for your test – to
699 ensure there are no other interrupts on the system.
700 - Configure NIC interrupts to only use the cores that are not
701 allocated to any other process (VNF/vSwitch).
703 - NUMA configuration: Any unused sockets in a multi-socket system
705 - CPU pinning: The vSwitch and the VNF should each be affinitized to
706 separate logical cores using a combination of maxcpus, isolcpus and
708 - BIOS configuration: BIOS should be configured for performance where
709 an explicit option exists, sleep states should be disabled, any
710 virtualization optimization technologies should be enabled, and
711 hyperthreading should also be enabled.
715 System validation is broken down into two sub-categories: Platform
716 validation and VNF validation. The validation test itself involves
717 verifying the forwarding capability and stability for the sub-system
718 under test. The rationale behind system validation is two fold. Firstly
719 to give a tester confidence in the stability of the platform or VNF that
720 is being tested; and secondly to provide base performance comparison
721 points to understand the overhead introduced by the virtual switch.
723 * Benchmark platform forwarding capability: This is an OPTIONAL test
724 used to verify the platform and measure the base performance (maximum
725 forwarding rate in fps and latency) that can be achieved by the
726 platform without a vSwitch or a VNF. The following diagram outlines
727 the set-up for benchmarking Platform forwarding capability:
729 .. code-block:: console
732 +--------------------------------------------------+ |
733 | +------------------------------------------+ | |
735 | | l2fw or DPDK L2FWD app | | Host
737 | +------------------------------------------+ | |
739 +---+------------------------------------------+---+ __|
743 +--------------------------------------------------+
745 | traffic generator |
747 +--------------------------------------------------+
749 * Benchmark VNF forwarding capability: This test is used to verify
750 the VNF and measure the base performance (maximum forwarding rate in
751 fps and latency) that can be achieved by the VNF without a vSwitch.
752 The performance metrics collected by this test will serve as a key
753 comparison point for NIC passthrough technologies and vSwitches. VNF
754 in this context refers to the hypervisor and the VM. The following
755 diagram outlines the set-up for benchmarking VNF forwarding
758 .. code-block:: console
761 +--------------------------------------------------+ |
762 | +------------------------------------------+ | |
766 | +------------------------------------------+ | |
767 | | Passthrough/SR-IOV | | Host
768 | +------------------------------------------+ | |
770 +---+------------------------------------------+---+ __|
774 +--------------------------------------------------+
776 | traffic generator |
778 +--------------------------------------------------+
781 Methodology to benchmark Platform/VNF forwarding capability
782 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
784 The recommended methodology for the platform/VNF validation and
785 benchmark is: - Run `RFC2889 <https://www.rfc-editor.org/rfc/rfc2289.txt>`__
786 Maximum Forwarding Rate test, this test will produce maximum
787 forwarding rate and latency results that will serve as the
788 expected values. These expected values can be used in
789 subsequent steps or compared with in subsequent validation tests. -
790 Transmit bidirectional traffic at line rate/max forwarding rate
791 (whichever is higher) for at least 72 hours, measure throughput (fps)
792 and latency. - Note: Traffic should be bidirectional. - Establish a
793 baseline forwarding rate for what the platform can achieve. - Additional
794 validation: After the test has completed for 72 hours run bidirectional
795 traffic at the maximum forwarding rate once more to see if the system is
796 still functional and measure throughput (fps) and latency. Compare the
797 measure the new obtained values with the expected values.
799 **NOTE 1**: How the Platform is configured for its forwarding capability
800 test (BIOS settings, GRUB configuration, runlevel...) is how the
801 platform should be configured for every test after this
803 **NOTE 2**: How the VNF is configured for its forwarding capability test
804 (# of vCPUs, vNICs, Memory, affinitization…) is how it should be
805 configured for every test that uses a VNF after this.
807 2.2.4 RFCs for testing virtual switch performance
808 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
810 The starting point for defining the suite of tests for benchmarking the
811 performance of a virtual switch is to take existing RFCs and standards
812 that were designed to test their physical counterparts and adapting them
813 for testing virtual switches. The rationale behind this is to establish
814 a fair comparison between the performance of virtual and physical
815 switches. This section outlines the RFCs that are used by this
818 RFC 1242 Benchmarking Terminology for Network Interconnection
819 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
820 Devices RFC 1242 defines the terminology that is used in describing
821 performance benchmarking tests and their results. Definitions and
822 discussions covered include: Back-to-back, bridge, bridge/router,
823 constant load, data link frame size, frame loss rate, inter frame gap,
824 latency, and many more.
826 RFC 2544 Benchmarking Methodology for Network Interconnect Devices
827 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
828 RFC 2544 outlines a benchmarking methodology for network Interconnect
829 Devices. The methodology results in performance metrics such as latency,
830 frame loss percentage, and maximum data throughput.
832 In this document network “throughput” (measured in millions of frames
833 per second) is based on RFC 2544, unless otherwise noted. Frame size
834 refers to Ethernet frames ranging from smallest frames of 64 bytes to
835 largest frames of 4K bytes.
839 1. Throughput test defines the maximum number of frames per second
840 that can be transmitted without any error.
842 2. Latency test measures the time required for a frame to travel from
843 the originating device through the network to the destination device.
844 Please note that RFC2544 Latency measurement will be superseded with
845 a measurement of average latency over all successfully transferred
848 3. Frame loss test measures the network’s
849 response in overload conditions - a critical indicator of the
850 network’s ability to support real-time applications in which a
851 large amount of frame loss will rapidly degrade service quality.
853 4. Burst test assesses the buffering capability of a virtual switch. It
854 measures the maximum number of frames received at full line rate
855 before a frame is lost. In carrier Ethernet networks, this
856 measurement validates the excess information rate (EIR) as defined in
859 5. System recovery to characterize speed of recovery from an overload
862 6. Reset to characterize speed of recovery from device or software
863 reset. This type of test has been updated by `RFC6201 <https://www.rfc-editor.org/rfc/rfc6201.txt>`__ as such,
864 the methodology defined by this specification will be that of RFC 6201.
866 Although not included in the defined RFC 2544 standard, another crucial
867 measurement in Ethernet networking is packet delay variation. The
868 definition set out by this specification comes from
869 `RFC5481 <https://www.rfc-editor.org/rfc/rfc5481.txt>`__.
871 RFC 2285 Benchmarking Terminology for LAN Switching Devices
872 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
873 RFC 2285 defines the terminology that is used to describe the
874 terminology for benchmarking a LAN switching device. It extends RFC
875 1242 and defines: DUTs, SUTs, Traffic orientation and distribution,
876 bursts, loads, forwarding rates, etc.
878 RFC 2889 Benchmarking Methodology for LAN Switching
879 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
880 RFC 2889 outlines a benchmarking methodology for LAN switching, it
881 extends RFC 2544. The outlined methodology gathers performance
882 metrics for forwarding, congestion control, latency, address handling
883 and finally filtering.
885 RFC 3918 Methodology for IP Multicast Benchmarking
886 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
887 RFC 3918 outlines a methodology for IP Multicast benchmarking.
889 RFC 4737 Packet Reordering Metrics
890 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
891 RFC 4737 describes metrics for identifying and counting re-ordered
892 packets within a stream, and metrics to measure the extent each
893 packet has been re-ordered.
895 RFC 5481 Packet Delay Variation Applicability Statement
896 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
897 RFC 5481 defined two common, but different forms of delay variation
898 metrics, and compares the metrics over a range of networking
899 circumstances and tasks. The most suitable form for vSwitch
900 benchmarking is the "PDV" form.
902 RFC 6201 Device Reset Characterization
903 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
904 RFC 6201 extends the methodology for characterizing the speed of
905 recovery of the DUT from device or software reset described in RFC
908 2.2.5 Details of the Test Report
909 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
911 There are a number of parameters related to the system, DUT and tests
912 that can affect the repeatability of a test results and should be
913 recorded. In order to minimise the variation in the results of a test,
914 it is recommended that the test report includes the following information:
916 - Hardware details including:
920 - Memory information (see below)
921 - Number of enabled cores.
922 - Number of cores used for the test.
923 - Number of physical NICs, as well as their details (manufacturer,
924 versions, type and the PCI slot they are plugged into).
925 - NIC interrupt configuration.
926 - BIOS version, release date and any configurations that were
929 - Software details including:
931 - OS version (for host and VNF)
932 - Kernel version (for host and VNF)
933 - GRUB boot parameters (for host and VNF).
934 - Hypervisor details (Type and version).
935 - Selected vSwitch, version number or commit id used.
936 - vSwitch launch command line if it has been parameterised.
937 - Memory allocation to the vSwitch – which NUMA node it is using,
938 and how many memory channels.
939 - Where the vswitch is built from source: compiler details including
940 versions and the flags that were used to compile the vSwitch.
941 - DPDK or any other SW dependency version number or commit id used.
942 - Memory allocation to a VM - if it's from Hugpages/elsewhere.
943 - VM storage type: snapshot/independent persistent/independent
946 - Number of Virtual NICs (vNICs), versions, type and driver.
947 - Number of virtual CPUs and their core affinity on the host.
948 - Number vNIC interrupt configuration.
949 - Thread affinitization for the applications (including the vSwitch
951 - Details of Resource isolation, such as CPUs designated for
952 Host/Kernel (isolcpu) and CPUs designated for specific processes
971 - Traffic Information:
973 - Traffic type - UDP, TCP, IMIX / Other.
976 - Deployment Scenario.
978 **Note**: Tests that require additional parameters to be recorded will
979 explicitly specify this.
981 2.3. Test identification
982 ------------------------
983 2.3.1 Throughput tests
984 ~~~~~~~~~~~~~~~~~~~~~~
985 The following tests aim to determine the maximum forwarding rate that
986 can be achieved with a virtual switch. The list is not exhaustive but
987 should indicate the type of tests that should be required. It is
988 expected that more will be added.
990 Test ID: LTD.Throughput.RFC2544.PacketLossRatio
991 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
992 **Title**: RFC 2544 X% packet loss ratio Throughput and Latency Test
994 **Prerequisite Test**: N/A
1000 This test determines the DUT's maximum forwarding rate with X% traffic
1001 loss for a constant load (fixed length frames at a fixed interval time).
1002 The default loss percentages to be tested are: - X = 0% - X = 10^-7%
1004 Note: Other values can be tested if required by the user.
1006 The selected frame sizes are those previously defined under `Default
1007 Test Parameters <#DefaultParams>`__. The test can also be used to
1008 determine the average latency of the traffic.
1010 Under the `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__
1011 test methodology, the test duration will
1012 include a number of trials; each trial should run for a minimum period
1013 of 60 seconds. A binary search methodology must be applied for each
1014 trial to obtain the final result.
1016 **Expected Result**: At the end of each trial, the presence or absence
1017 of loss determines the modification of offered load for the next trial,
1018 converging on a maximum rate, or
1019 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__ Throughput with X% loss.
1020 The Throughput load is re-used in related
1021 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__ tests and other
1024 **Metrics Collected**:
1026 The following are the metrics collected for this test:
1028 - The maximum forwarding rate in Frames Per Second (FPS) and Mbps of
1029 the DUT for each frame size with X% packet loss.
1030 - The average latency of the traffic flow when passing through the DUT
1031 (if testing for latency, note that this average is different from the
1032 test specified in Section 26.3 of
1033 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__).
1034 - CPU and memory utilization may also be collected as part of this
1035 test, to determine the vSwitch's performance footprint on the system.
1037 Test ID: LTD.Throughput.RFC2544.PacketLossRatioFrameModification
1038 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1039 **Title**: RFC 2544 X% packet loss Throughput and Latency Test with
1042 **Prerequisite Test**: N/A
1048 This test determines the DUT's maximum forwarding rate with X% traffic
1049 loss for a constant load (fixed length frames at a fixed interval time).
1050 The default loss percentages to be tested are: - X = 0% - X = 10^-7%
1052 Note: Other values can be tested if required by the user.
1054 The selected frame sizes are those previously defined under `Default
1055 Test Parameters <#DefaultParams>`__. The test can also be used to
1056 determine the average latency of the traffic.
1058 Under the `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__
1059 test methodology, the test duration will
1060 include a number of trials; each trial should run for a minimum period
1061 of 60 seconds. A binary search methodology must be applied for each
1062 trial to obtain the final result.
1064 During this test, the DUT must perform the following operations on the
1067 - Perform packet parsing on the DUT's ingress port.
1068 - Perform any relevant address look-ups on the DUT's ingress ports.
1069 - Modify the packet header before forwarding the packet to the DUT's
1070 egress port. Packet modifications include:
1072 - Modifying the Ethernet source or destination MAC address.
1073 - Modifying/adding a VLAN tag. (**Recommended**).
1074 - Modifying/adding a MPLS tag.
1075 - Modifying the source or destination ip address.
1076 - Modifying the TOS/DSCP field.
1077 - Modifying the source or destination ports for UDP/TCP/SCTP.
1078 - Modifying the TTL.
1080 **Expected Result**: The Packet parsing/modifications require some
1081 additional degree of processing resource, therefore the
1082 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__
1083 Throughput is expected to be somewhat lower than the Throughput level
1084 measured without additional steps. The reduction is expected to be
1085 greatest on tests with the smallest packet sizes (greatest header
1088 **Metrics Collected**:
1090 The following are the metrics collected for this test:
1092 - The maximum forwarding rate in Frames Per Second (FPS) and Mbps of
1093 the DUT for each frame size with X% packet loss and packet
1094 modification operations being performed by the DUT.
1095 - The average latency of the traffic flow when passing through the DUT
1096 (if testing for latency, note that this average is different from the
1097 test specified in Section 26.3 of
1098 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__).
1099 - The `RFC5481 <https://www.rfc-editor.org/rfc/rfc5481.txt>`__
1100 PDV form of delay variation on the traffic flow,
1101 using the 99th percentile.
1102 - CPU and memory utilization may also be collected as part of this
1103 test, to determine the vSwitch's performance footprint on the system.
1105 Test ID: LTD.Throughput.RFC2544.Profile
1106 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1107 **Title**: RFC 2544 Throughput and Latency Profile
1109 **Prerequisite Test**: N/A
1115 This test reveals how throughput and latency degrades as the offered
1116 rate varies in the region of the DUT's maximum forwarding rate as
1117 determined by LTD.Throughput.RFC2544.PacketLossRatio (0% Packet Loss).
1118 For example it can be used to determine if the degradation of throughput
1119 and latency as the offered rate increases is slow and graceful or sudden
1122 The selected frame sizes are those previously defined under `Default
1123 Test Parameters <#DefaultParams>`__.
1125 The offered traffic rate is described as a percentage delta with respect
1126 to the DUT's RFC 2544 Throughput as determined by
1127 LTD.Throughput.RFC2544.PacketLoss Ratio (0% Packet Loss case). A delta
1128 of 0% is equivalent to an offered traffic rate equal to the RFC 2544
1129 Throughput; A delta of +50% indicates an offered rate half-way
1130 between the Throughput and line-rate, whereas a delta of
1131 -50% indicates an offered rate of half the maximum rate. Therefore the
1132 range of the delta figure is natuarlly bounded at -100% (zero offered
1133 traffic) and +100% (traffic offered at line rate).
1135 The following deltas to the maximum forwarding rate should be applied:
1137 - -50%, -10%, 0%, +10% & +50%
1139 **Expected Result**: For each packet size a profile should be produced
1140 of how throughput and latency vary with offered rate.
1142 **Metrics Collected**:
1144 The following are the metrics collected for this test:
1146 - The forwarding rate in Frames Per Second (FPS) and Mbps of the DUT
1147 for each delta to the maximum forwarding rate and for each frame
1149 - The average latency for each delta to the maximum forwarding rate and
1150 for each frame size.
1151 - CPU and memory utilization may also be collected as part of this
1152 test, to determine the vSwitch's performance footprint on the system.
1153 - Any failures experienced (for example if the vSwitch crashes, stops
1154 processing packets, restarts or becomes unresponsive to commands)
1155 when the offered load is above Maximum Throughput MUST be recorded
1156 and reported with the results.
1158 Test ID: LTD.Throughput.RFC2544.SystemRecoveryTime
1159 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1160 **Title**: RFC 2544 System Recovery Time Test
1162 **Prerequisite Test** LTD.Throughput.RFC2544.PacketLossRatio
1168 The aim of this test is to determine the length of time it takes the DUT
1169 to recover from an overload condition for a constant load (fixed length
1170 frames at a fixed interval time). The selected frame sizes are those
1171 previously defined under `Default Test Parameters <#DefaultParams>`__,
1172 traffic should be sent to the DUT under normal conditions. During the
1173 duration of the test and while the traffic flows are passing though the
1174 DUT, at least one situation leading to an overload condition for the DUT
1175 should occur. The time from the end of the overload condition to when
1176 the DUT returns to normal operations should be measured to determine
1177 recovery time. Prior to overloading the DUT, one should record the
1178 average latency for 10,000 packets forwarded through the DUT.
1180 The overload condition SHOULD be to transmit traffic at a very high
1181 frame rate to the DUT (150% of the maximum 0% packet loss rate as
1182 determined by LTD.Throughput.RFC2544.PacketLossRatio or line-rate
1183 whichever is lower), for at least 60 seconds, then reduce the frame rate
1184 to 75% of the maximum 0% packet loss rate. A number of time-stamps
1185 should be recorded: - Record the time-stamp at which the frame rate was
1186 reduced and record a second time-stamp at the time of the last frame
1187 lost. The recovery time is the difference between the two timestamps. -
1188 Record the average latency for 10,000 frames after the last frame loss
1189 and continue to record average latency measurements for every 10,000
1190 frames, when latency returns to within 10% of pre-overload levels record
1193 **Expected Result**:
1195 **Metrics collected**
1197 The following are the metrics collected for this test:
1199 - The length of time it takes the DUT to recover from an overload
1201 - The length of time it takes the DUT to recover the average latency to
1202 pre-overload conditions.
1204 **Deployment scenario**:
1206 - Physical → virtual switch → physical.
1208 Test ID: LTD.Throughput.RFC2544.BackToBackFrames
1209 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1210 **Title**: RFC2544 Back To Back Frames Test
1212 **Prerequisite Test**: N
1218 The aim of this test is to characterize the ability of the DUT to
1219 process back-to-back frames. For each frame size previously defined
1220 under `Default Test Parameters <#DefaultParams>`__, a burst of traffic
1221 is sent to the DUT with the minimum inter-frame gap between each frame.
1222 If the number of received frames equals the number of frames that were
1223 transmitted, the burst size should be increased and traffic is sent to
1224 the DUT again. The value measured is the back-to-back value, that is the
1225 maximum burst size the DUT can handle without any frame loss.
1227 **Expected Result**:
1229 Tests of back-to-back frames with physical devices have produced
1230 unstable results in some cases. All tests should be repeated in multiple
1231 test sessions and results stability should be examined.
1233 **Metrics collected**
1235 The following are the metrics collected for this test:
1237 - The back-to-back value, which is the the number of frames in the
1238 longest burst that the DUT will handle without the loss of any
1240 - CPU and memory utilization may also be collected as part of this
1241 test, to determine the vSwitch's performance footprint on the system.
1243 **Deployment scenario**:
1245 - Physical → virtual switch → physical.
1247 Test ID: LTD.Throughput.RFC2889.Soak
1248 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1249 **Title**: RFC 2889 X% packet loss Throughput Soak Test
1251 **Prerequisite Test** LTD.Throughput.RFC2544.PacketLossRatio
1257 The aim of this test is to understand the Throughput stability over an
1258 extended test duration in order to uncover any outliers. To allow for an
1259 extended test duration, the test should ideally run for 24 hours or, if
1260 this is not possible, for at least 6 hours. For this test, each frame
1261 size must be sent at the highest Throughput with X% packet loss, as
1262 determined in the prerequisite test. The default loss percentages to be
1263 tested are: - X = 0% - X = 10^-7%
1265 Note: Other values can be tested if required by the user.
1267 **Expected Result**:
1269 **Metrics Collected**:
1271 The following are the metrics collected for this test:
1273 - Throughput stability of the DUT.
1275 - This means reporting the number of packets lost per time interval
1276 and reporting any time intervals with packet loss. The
1277 `RFC2889 <https://www.rfc-editor.org/rfc/rfc2289.txt>`__
1278 Forwarding Rate shall be measured in each interval.
1279 An interval of 60s is suggested.
1281 - CPU and memory utilization may also be collected as part of this
1282 test, to determine the vSwitch's performance footprint on the system.
1283 - The `RFC5481 <https://www.rfc-editor.org/rfc/rfc5481.txt>`__
1284 PDV form of delay variation on the traffic flow,
1285 using the 99th percentile.
1287 Test ID: LTD.Throughput.RFC2889.SoakFrameModification
1288 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1289 **Title**: RFC 2889 Throughput Soak Test with Frame Modification
1291 **Prerequisite Test**: LTD.Throughput.RFC2544.PacketLossRatioFrameModification (0% Packet Loss)
1297 The aim of this test is to understand the throughput stability over an
1298 extended test duration in order to uncover any outliers. To allow for an
1299 extended test duration, the test should ideally run for 24 hours or, if
1300 this is not possible, for at least 6 hour. For this test, each frame
1301 size must be sent at the highest Throughput with 0% packet loss, as
1302 determined in the prerequisite test.
1304 During this test, the DUT must perform the following operations on the
1307 - Perform packet parsing on the DUT's ingress port.
1308 - Perform any relevant address look-ups on the DUT's ingress ports.
1309 - Modify the packet header before forwarding the packet to the DUT's
1310 egress port. Packet modifications include:
1312 - Modifying the Ethernet source or destination MAC address.
1313 - Modifying/adding a VLAN tag (**Recommended**).
1314 - Modifying/adding a MPLS tag.
1315 - Modifying the source or destination ip address.
1316 - Modifying the TOS/DSCP field.
1317 - Modifying the source or destination ports for UDP/TCP/SCTP.
1318 - Modifying the TTL.
1320 **Expected Result**:
1322 **Metrics Collected**:
1324 The following are the metrics collected for this test:
1326 - Throughput stability of the DUT.
1328 - This means reporting the number of packets lost per time interval
1329 and reporting any time intervals with packet loss. The
1330 `RFC2889 <https://www.rfc-editor.org/rfc/rfc2289.txt>`__
1331 Forwarding Rate shall be measured in each interval.
1332 An interval of 60s is suggested.
1334 - CPU and memory utilization may also be collected as part of this
1335 test, to determine the vSwitch's performance footprint on the system.
1336 - The `RFC5481 <https://www.rfc-editor.org/rfc/rfc5481.txt>`__ PDV form of delay variation on the traffic flow,
1337 using the 99th percentile.
1339 Test ID: LTD.Throughput.RFC6201.ResetTime
1340 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1341 **Title**: RFC 6201 Reset Time Test
1343 **Prerequisite Test**: N/A
1349 The aim of this test is to determine the length of time it takes the DUT
1350 to recover from a reset.
1352 Two reset methods are defined - planned and unplanned. A planned reset
1353 requires stopping and restarting the virtual switch by the usual
1354 'graceful' method defined by it's documentation. An unplanned reset
1355 requires simulating a fatal internal fault in the virtual switch - for
1356 example by using kill -SIGKILL on a Linux environment.
1358 Both reset methods SHOULD be exercised.
1360 For each frame size previously defined under `Default Test
1361 Parameters <#DefaultParams>`__, traffic should be sent to the DUT under
1362 normal conditions. During the duration of the test and while the traffic
1363 flows are passing through the DUT, the DUT should be reset and the Reset
1364 time measured. The Reset time is the total time that a device is
1365 determined to be out of operation and includes the time to perform the
1366 reset and the time to recover from it (cf. `RFC6201 <https://www.rfc-editor.org/rfc/rfc6201.txt>`__).
1368 `RFC6201 <https://www.rfc-editor.org/rfc/rfc6201.txt>`__ defines two methods to measure the Reset time:
1369 - Frame-Loss Method: which requires the monitoring of the number of
1370 lost frames and calculates the Reset time based on the number of
1371 frames lost and the offered rate according to the following
1374 .. code-block:: console
1376 Frames_lost (packets)
1377 Reset_time = -------------------------------------
1378 Offered_rate (packets per second)
1380 - Timestamp Method: which measures the time from which the last frame
1381 is forwarded from the DUT to the time the first frame is forwarded
1382 after the reset. This involves time-stamping all transmitted frames
1383 and recording the timestamp of the last frame that was received prior
1384 to the reset and also measuring the timestamp of the first frame that
1385 is received after the reset. The Reset time is the difference between
1386 these two timestamps.
1388 According to `RFC6201 <https://www.rfc-editor.org/rfc/rfc6201.txt>`__ the choice of method depends on the test
1389 tool's capability; the Frame-Loss method SHOULD be used if the test tool
1390 supports: - Counting the number of lost frames per stream. -
1391 Transmitting test frame despite the physical link status.
1393 whereas the Timestamp method SHOULD be used if the test tool supports: -
1394 Timestamping each frame. - Monitoring received frame's timestamp. -
1395 Transmitting frames only if the physical link status is up.
1397 **Expected Result**:
1399 **Metrics collected**
1401 The following are the metrics collected for this test: - Average Reset
1402 Time over the number of trials performed.
1404 Results of this test should include the following information: - The
1405 reset method used. - Throughput in Fps and Mbps. - Average Frame Loss
1406 over the number of trials performed. - Average Reset Time in
1407 milliseconds over the number of trials performed. - Number of trials
1408 performed. - Protocol: IPv4, IPv6, MPLS, etc. - Frame Size in Octets -
1409 Port Media: Ethernet, Gigabit Ethernet (GbE), etc. - Port Speed: 10
1410 Gbps, 40 Gbps etc. - Interface Encapsulation: Ethernet, Ethernet VLAN,
1413 **Deployment scenario**:
1415 - Physical → virtual switch → physical.
1417 Test ID: LTD.Throughput.RFC2889.MaxForwardingRate
1418 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1419 **Title**: RFC2889 Forwarding Rate Test
1421 **Prerequisite Test**: LTD.Throughput.RFC2544.PacketLossRatio
1427 This test measures the DUT's Max Forwarding Rate when the Offered Load
1428 is varied between the throughput and the Maximum Offered Load for fixed
1429 length frames at a fixed time interval. The selected frame sizes are
1430 those previously defined under `Default Test
1431 Parameters <#DefaultParams>`__. The throughput is the maximum offered
1432 load with 0% frame loss (measured by the prerequisite test), and the
1433 Maximum Offered Load (as defined by
1434 `RFC2285 <https://www.rfc-editor.org/rfc/rfc2285.txt>`__) is *"the highest
1435 number of frames per second that an external source can transmit to a
1436 DUT/SUT for forwarding to a specified output interface or interfaces"*.
1438 Traffic should be sent to the DUT at a particular rate (TX rate)
1439 starting with TX rate equal to the throughput rate. The rate of
1440 successfully received frames at the destination counted (in FPS). If the
1441 RX rate is equal to the TX rate, the TX rate should be increased by a
1442 fixed step size and the RX rate measured again until the Max Forwarding
1445 The trial duration for each iteration should last for the period of time
1446 needed for the system to reach steady state for the frame size being
1447 tested. Under `RFC2889 <https://www.rfc-editor.org/rfc/rfc2289.txt>`__
1448 (Sec. 5.6.3.1) test methodology, the test
1449 duration should run for a minimum period of 30 seconds, regardless
1450 whether the system reaches steady state before the minimum duration
1453 **Expected Result**: According to
1454 `RFC2889 <https://www.rfc-editor.org/rfc/rfc2289.txt>`__ The Max Forwarding Rate
1455 is the highest forwarding rate of a DUT taken from an iterative set of
1456 forwarding rate measurements. The iterative set of forwarding rate
1457 measurements are made by setting the intended load transmitted from an
1458 external source and measuring the offered load (i.e what the DUT is
1459 capable of forwarding). If the Throughput == the Maximum Offered Load,
1460 it follows that Max Forwarding Rate is equal to the Maximum Offered
1463 **Metrics Collected**:
1465 The following are the metrics collected for this test:
1467 - The Max Forwarding Rate for the DUT for each packet size.
1468 - CPU and memory utilization may also be collected as part of this
1469 test, to determine the vSwitch's performance footprint on the system.
1471 Test ID: LTD.Throughput.RFC2889.ForwardPressure
1472 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1473 **Title**: RFC2889 Forward Pressure Test
1475 **Prerequisite Test**: LTD.Throughput.RFC2889.MaxForwardingRate
1481 The aim of this test is to determine if the DUT transmits frames with an
1482 inter-frame gap that is less than 12 bytes. This test overloads the DUT
1483 and measures the output for forward pressure. Traffic should be
1484 transmitted to the DUT with an inter-frame gap of 11 bytes, this will
1485 overload the DUT by 1 byte per frame. The forwarding rate of the DUT
1488 **Expected Result**: The forwarding rate should not exceed the maximum
1489 forwarding rate of the DUT collected by
1490 LTD.Throughput.RFC2889.MaxForwardingRate.
1492 **Metrics collected**
1494 The following are the metrics collected for this test:
1496 - Forwarding rate of the DUT in FPS or Mbps.
1497 - CPU and memory utilization may also be collected as part of this
1498 test, to determine the vSwitch's performance footprint on the system.
1500 **Deployment scenario**:
1502 - Physical → virtual switch → physical.
1505 Test ID: LTD.Throughput.RFC2889.ErrorFramesFiltering
1506 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1507 **Title**: RFC2889 Error Frames Filtering Test
1509 **Prerequisite Test**: N/A
1515 The aim of this test is to determine whether the DUT will propagate any
1516 erroneous frames it receives or whether it is capable of filtering out
1517 the erroneous frames. Traffic should be sent with erroneous frames
1518 included within the flow at random intervals. Illegal frames that must
1519 be tested include: - Oversize Frames. - Undersize Frames. - CRC Errored
1520 Frames. - Dribble Bit Errored Frames - Alignment Errored Frames
1522 The traffic flow exiting the DUT should be recorded and checked to
1523 determine if the erroneous frames where passed through the DUT.
1525 **Expected Result**: Broken frames are not passed!
1527 **Metrics collected**
1529 No Metrics are collected in this test, instead it determines:
1531 - Whether the DUT will propagate erroneous frames.
1532 - Or whether the DUT will correctly filter out any erroneous frames
1533 from traffic flow with out removing correct frames.
1535 **Deployment scenario**:
1537 - Physical → virtual switch → physical.
1539 Test ID: LTD.Throughput.RFC2889.BroadcastFrameForwarding
1540 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1541 **Title**: RFC2889 Broadcast Frame Forwarding Test
1543 **Prerequisite Test**: N
1549 The aim of this test is to determine the maximum forwarding rate of the
1550 DUT when forwarding broadcast traffic. For each frame previously defined
1551 under `Default Test Parameters <#DefaultParams>`__, the traffic should
1552 be set up as broadcast traffic. The traffic throughput of the DUT should
1555 The test should be conducted with at least 4 physical ports on the DUT.
1556 The number of ports used MUST be recorded.
1558 As broadcast involves forwarding a single incoming packet to several
1559 destinations, the latency of a single packet is defined as the average
1560 of the latencies for each of the broadcast destinations.
1562 The incoming packet is transmitted on each of the other physical ports,
1563 it is not transmitted on the port on which it was received. The test MAY
1564 be conducted using different broadcasting ports to uncover any
1565 performance differences.
1567 **Expected Result**:
1569 **Metrics collected**:
1571 The following are the metrics collected for this test:
1573 - The forwarding rate of the DUT when forwarding broadcast traffic.
1574 - The minimum, average & maximum packets latencies observed.
1576 **Deployment scenario**:
1578 - Physical → virtual switch 3x physical. In the Broadcast rate testing,
1579 four test ports are required. One of the ports is connected to the test
1580 device, so it can send broadcast frames and listen for miss-routed frames.
1582 2.3.2 Packet Latency tests
1583 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
1584 These tests will measure the store and forward latency as well as the packet
1585 delay variation for various packet types through the virtual switch. The
1586 following list is not exhaustive but should indicate the type of tests
1587 that should be required. It is expected that more will be added.
1589 Test ID: LTD.PacketLatency.InitialPacketProcessingLatency
1590 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1591 **Title**: Initial Packet Processing Latency
1593 **Prerequisite Test**: N/A
1599 In some virtual switch architectures, the first packets of a flow will
1600 take the system longer to process than subsequent packets in the flow.
1601 This test determines the latency for these packets. The test will
1602 measure the latency of the packets as they are processed by the
1603 flow-setup-path of the DUT. There are two methods for this test, a
1604 recommended method and a nalternative method that can be used if it is
1605 possible to disable the fastpath of the virtual switch.
1607 Recommended method: This test will send 64,000 packets to the DUT, each
1608 belonging to a different flow. Average packet latency will be determined
1609 over the 64,000 packets.
1611 Alternative method: This test will send a single packet to the DUT after
1612 a fixed interval of time. The time interval will be equivalent to the
1613 amount of time it takes for a flow to time out in the virtual switch
1614 plus 10%. Average packet latency will be determined over 1,000,000
1617 This test is intended only for non-learning virtual switches; For learning
1618 virtual switches use RFC2889.
1620 For this test, only unidirectional traffic is required.
1622 **Expected Result**: The average latency for the initial packet of all
1623 flows should be greater than the latency of subsequent traffic.
1625 **Metrics Collected**:
1627 The following are the metrics collected for this test:
1629 - Average latency of the initial packets of all flows that are
1630 processed by the DUT.
1632 **Deployment scenario**:
1634 - Physical → Virtual Switch → Physical.
1636 Test ID: LTD.PacketDelayVariation.RFC3393.Soak
1637 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1638 **Title**: Packet Delay Variation Soak Test
1640 **Prerequisite Tests**: LTD.Throughput.RFC2544.PacketLossRatio (0% Packet Loss)
1646 The aim of this test is to understand the distribution of packet delay
1647 variation for different frame sizes over an extended test duration and
1648 to determine if there are any outliers. To allow for an extended test
1649 duration, the test should ideally run for 24 hours or, if this is not
1650 possible, for at least 6 hour. For this test, each frame size must be
1651 sent at the highest possible throughput with 0% packet loss, as
1652 determined in the prerequisite test.
1654 **Expected Result**:
1656 **Metrics Collected**:
1658 The following are the metrics collected for this test:
1660 - The packet delay variation value for traffic passing through the DUT.
1661 - The `RFC5481 <https://www.rfc-editor.org/rfc/rfc5481.txt>`__
1662 PDV form of delay variation on the traffic flow,
1663 using the 99th percentile, for each 60s interval during the test.
1664 - CPU and memory utilization may also be collected as part of this
1665 test, to determine the vSwitch's performance footprint on the system.
1667 2.3.3 Scalability tests
1668 ~~~~~~~~~~~~~~~~~~~~~~~~
1669 The general aim of these tests is to understand the impact of large flow
1670 table size and flow lookups on throughput. The following list is not
1671 exhaustive but should indicate the type of tests that should be required.
1672 It is expected that more will be added.
1674 Test ID: LTD.Scalability.RFC2544.0PacketLoss
1675 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1676 **Title**: RFC 2544 0% loss Scalability throughput test
1678 **Prerequisite Test**: LTD.Throughput.RFC2544.PacketLossRatio, IF the
1679 delta Throughput between the single-flow RFC2544 test and this test with
1680 a variable number of flows is desired.
1686 The aim of this test is to measure how throughput changes as the number
1687 of flows in the DUT increases. The test will measure the throughput
1688 through the fastpath, as such the flows need to be installed on the DUT
1689 before passing traffic.
1691 For each frame size previously defined under `Default Test
1692 Parameters <#DefaultParams>`__ and for each of the following number of
1702 - Max supported number of flows.
1704 This test will be conducted under two conditions following the
1705 establishment of all flows as required by RFC 2544, regarding the flow
1706 expiration time-out:
1708 1) The time-out never expires during each trial.
1710 2) The time-out expires for all flows periodically. This would require a
1711 short time-out compared with flow re-appearance for a small number of
1712 flows, and may not be possible for all flow conditions.
1714 The maximum 0% packet loss Throughput should be determined in a manner
1715 identical to LTD.Throughput.RFC2544.PacketLossRatio.
1717 **Expected Result**:
1719 **Metrics Collected**:
1721 The following are the metrics collected for this test:
1723 - The maximum number of frames per second that can be forwarded at the
1724 specified number of flows and the specified frame size, with zero
1727 Test ID: LTD.MemoryBandwidth.RFC2544.0PacketLoss.Scalability
1728 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1729 **Title**: RFC 2544 0% loss Memory Bandwidth Scalability test
1731 **Prerequisite Tests**: LTD.Throughput.RFC2544.PacketLossRatio, IF the
1732 delta Throughput between an undisturbed RFC2544 test and this test with
1733 the Throughput affected by cache and memory bandwidth contention is desired.
1739 The aim of this test is to understand how the DUT's performance is
1740 affected by cache sharing and memory bandwidth between processes.
1742 During the test all cores not used by the vSwitch should be running a
1743 memory intensive application. This application should read and write
1744 random data to random addresses in unused physical memory. The random
1745 nature of the data and addresses is intended to consume cache, exercise
1746 main memory access (as opposed to cache) and exercise all memory buses
1747 equally. Furthermore:
1749 - the ratio of reads to writes should be recorded. A ratio of 1:1
1751 - the reads and writes MUST be of cache-line size and be cache-line aligned.
1752 - in NUMA architectures memory access SHOULD be local to the core's node.
1753 Whether only local memory or a mix of local and remote memory is used
1755 - the memory bandwidth (reads plus writes) used per-core MUST be recorded;
1756 the test MUST be run with a per-core memory bandwidth equal to half the
1757 maximum system memory bandwidth divided by the number of cores. The test
1758 MAY be run with other values for the per-core memory bandwidth.
1759 - the test MAY also be run with the memory intensive application running
1762 Under these conditions the DUT's 0% packet loss throughput is determined
1763 as per LTD.Throughput.RFC2544.PacketLossRatio.
1765 **Expected Result**:
1767 **Metrics Collected**:
1769 The following are the metrics collected for this test:
1771 - The DUT's 0% packet loss throughput in the presence of cache sharing and memory bandwidth between processes.
1773 2.3.4 Activation tests
1774 ~~~~~~~~~~~~~~~~~~~~~~~~
1775 The general aim of these tests is to understand the capacity of the
1776 and speed with which the vswitch can accomodate new flows.
1778 Test ID: LTD.Activation.RFC2889.AddressCachingCapacity
1779 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1780 **Title**: RFC2889 Address Caching Capacity Test
1782 **Prerequisite Test**: N/A
1788 Please note this test is only applicable to virtual switches that are capable of
1789 MAC learning. The aim of this test is to determine the address caching
1790 capacity of the DUT for a constant load (fixed length frames at a fixed
1791 interval time). The selected frame sizes are those previously defined
1792 under `Default Test Parameters <#DefaultParams>`__.
1794 In order to run this test the aging time, that is the maximum time the
1795 DUT will keep a learned address in its flow table, and a set of initial
1796 addresses, whose value should be >= 1 and <= the max number supported by
1797 the implementation must be known. Please note that if the aging time is
1798 configurable it must be longer than the time necessary to produce frames
1799 from the external source at the specified rate. If the aging time is
1800 fixed the frame rate must be brought down to a value that the external
1801 source can produce in a time that is less than the aging time.
1803 Learning Frames should be sent from an external source to the DUT to
1804 install a number of flows. The Learning Frames must have a fixed
1805 destination address and must vary the source address of the frames. The
1806 DUT should install flows in its flow table based on the varying source
1807 addresses. Frames should then be transmitted from an external source at
1808 a suitable frame rate to see if the DUT has properly learned all of the
1809 addresses. If there is no frame loss and no flooding, the number of
1810 addresses sent to the DUT should be increased and the test is repeated
1811 until the max number of cached addresses supported by the DUT
1814 **Expected Result**:
1816 **Metrics collected**:
1818 The following are the metrics collected for this test:
1820 - Number of cached addresses supported by the DUT.
1821 - CPU and memory utilization may also be collected as part of this
1822 test, to determine the vSwitch's performance footprint on the system.
1824 **Deployment scenario**:
1826 - Physical → virtual switch → 2 x physical (one receiving, one listening).
1828 Test ID: LTD.Activation.RFC2889.AddressLearningRate
1829 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1830 **Title**: RFC2889 Address Learning Rate Test
1832 **Prerequisite Test**: LTD.Memory.RFC2889.AddressCachingCapacity
1838 Please note this test is only applicable to virtual switches that are capable of
1839 MAC learning. The aim of this test is to determine the rate of address
1840 learning of the DUT for a constant load (fixed length frames at a fixed
1841 interval time). The selected frame sizes are those previously defined
1842 under `Default Test Parameters <#DefaultParams>`__, traffic should be
1843 sent with each IPv4/IPv6 address incremented by one. The rate at which
1844 the DUT learns a new address should be measured. The maximum caching
1845 capacity from LTD.Memory.RFC2889.AddressCachingCapacity should be taken
1846 into consideration as the maximum number of addresses for which the
1847 learning rate can be obtained.
1849 **Expected Result**: It may be worthwhile to report the behaviour when
1850 operating beyond address capacity - some DUTs may be more friendly to
1851 new addresses than others.
1853 **Metrics collected**:
1855 The following are the metrics collected for this test:
1857 - The address learning rate of the DUT.
1859 **Deployment scenario**:
1861 - Physical → virtual switch → 2 x physical (one receiving, one listening).
1864 2.3.5 Coupling between control path and datapath Tests
1865 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1866 The following tests aim to determine how tightly coupled the datapath
1867 and the control path are within a virtual switch. The following list
1868 is not exhaustive but should indicate the type of tests that should be
1869 required. It is expected that more will be added.
1871 Test ID: LTD.CPDPCouplingFlowAddition
1872 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1873 **Title**: Control Path and Datapath Coupling
1875 **Prerequisite Test**:
1881 The aim of this test is to understand how exercising the DUT's control
1882 path affects datapath performance.
1884 Initially a certain number of flow table entries are installed in the
1885 vSwitch. Then over the duration of an RFC2544 throughput test
1886 flow-entries are added and removed at the rates specified below. No
1887 traffic is 'hitting' these flow-entries, they are simply added and
1890 The test MUST be repeated with the following initial number of
1891 flow-entries installed: - < 10 - 1000 - 100,000 - 10,000,000 (or the
1892 maximum supported number of flow-entries)
1894 The test MUST be repeated with the following rates of flow-entry
1895 addition and deletion per second: - 0 - 1 (i.e. 1 addition plus 1
1896 deletion) - 100 - 10,000
1898 **Expected Result**:
1900 **Metrics Collected**:
1902 The following are the metrics collected for this test:
1904 - The maximum forwarding rate in Frames Per Second (FPS) and Mbps of
1906 - The average latency of the traffic flow when passing through the DUT
1907 (if testing for latency, note that this average is different from the
1908 test specified in Section 26.3 of
1909 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__).
1910 - CPU and memory utilization may also be collected as part of this
1911 test, to determine the vSwitch's performance footprint on the system.
1913 **Deployment scenario**:
1915 - Physical → virtual switch → physical.
1917 2.3.6 CPU and memory consumption
1918 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1919 The following tests will profile a virtual switch's CPU and memory
1920 utilization under various loads and circumstances. The following
1921 list is not exhaustive but should indicate the type of tests that
1922 should be required. It is expected that more will be added.
1924 Test ID: LTD.CPU.RFC2544.0PacketLoss
1925 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1926 **Title**: RFC 2544 0% Loss Compute Test
1928 **Prerequisite Test**:
1934 The aim of this test is to understand the overall performance of the
1935 system when a CPU intensive application is run on the same DUT as the
1936 Virtual Switch. For each frame size, an
1937 LTD.Throughput.RFC2544.PacketLossRatio (0% Packet Loss) test should be
1938 performed. Throughout the entire test a CPU intensive application should
1939 be run on all cores on the system not in use by the Virtual Switch. For
1940 NUMA system only cores on the same NUMA node are loaded.
1942 It is recommended that stress-ng be used for loading the non-Virtual
1943 Switch cores but any stress tool MAY be used.
1945 **Expected Result**:
1947 **Metrics Collected**:
1949 The following are the metrics collected for this test:
1951 - CPU utilization of the cores running the Virtual Switch.
1952 - The number of identity of the cores allocated to the Virtual Switch.
1953 - The configuration of the stress tool (for example the command line
1954 parameters used to start it.)
1956 2.3.7 Summary List of Tests
1957 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
1960 - Test ID: LTD.Throughput.RFC2544.PacketLossRatio
1961 - Test ID: LTD.Throughput.RFC2544.PacketLossRatioFrameModification
1962 - Test ID: LTD.Throughput.RFC2544.Profile
1963 - Test ID: LTD.Throughput.RFC2544.SystemRecoveryTime
1964 - Test ID: LTD.Throughput.RFC2544.BackToBackFrames
1965 - Test ID: LTD.Throughput.RFC2889.Soak
1966 - Test ID: LTD.Throughput.RFC2889.SoakFrameModification
1967 - Test ID: LTD.Throughput.RFC6201.ResetTime
1968 - Test ID: LTD.Throughput.RFC2889.MaxForwardingRate
1969 - Test ID: LTD.Throughput.RFC2889.ForwardPressure
1970 - Test ID: LTD.Throughput.RFC2889.ErrorFramesFiltering
1971 - Test ID: LTD.Throughput.RFC2889.BroadcastFrameForwarding
1973 2. Packet Latency tests
1975 - Test ID: LTD.PacketLatency.InitialPacketProcessingLatency
1976 - Test ID: LTD.PacketDelayVariation.RFC3393.Soak
1978 3. Scalability tests
1980 - Test ID: LTD.Scalability.RFC2544.0PacketLoss
1981 - Test ID: LTD.MemoryBandwidth.RFC2544.0PacketLoss.Scalability
1985 - Test ID: LTD.Activation.RFC2889.AddressCachingCapacity
1986 - Test ID: LTD.Activation.RFC2889.AddressLearningRate
1988 5. Coupling between control path and datapath Tests
1990 - Test ID: LTD.CPDPCouplingFlowAddition
1992 6. CPU and memory consumption
1994 - Test ID: LTD.CPU.RFC2544.0PacketLoss
Code Review / vswitchperf.git / blob