8 The objective of the OPNFV project titled
9 **“Characterize vSwitch Performance for Telco NFV Use Cases”**, is to
10 evaluate a virtual switch to identify its suitability for a Telco
11 Network Function Virtualization (NFV) environment. The intention of this
12 Level Test Design (LTD) document is to specify the set of tests to carry
13 out in order to objectively measure the current characteristics of a
14 virtual switch in the Network Function Virtualization Infrastructure
15 (NFVI) as well as the test pass criteria. The detailed test cases will
16 be defined in `Section 2 <#DetailsOfTheLevelTestDesign>`__, preceded by
17 the `Document identifier <#DocId>`__ and the `Scope <#Scope>`__.
19 This document is currently in draft form.
24 =========================
26 The document id will be used to uniquely
27 identify versions of the LTD. The format for the document id will be:
28 OPNFV\_vswitchperf\_LTD\_ver\_NUM\_MONTH\_YEAR\_STATUS, where by the
29 status is one of: draft, reviewed, corrected or final. The document id
30 for this version of the LTD is:
31 OPNFV\_vswitchperf\_LTD\_ver\_1.6\_Jan\_15\_DRAFT.
38 The main purpose of this project is to specify a suite of
39 performance tests in order to objectively measure the current packet
40 transfer characteristics of a virtual switch in the NFVI. The intent of
41 the project is to facilitate testing of any virtual switch. Thus, a
42 generic suite of tests shall be developed, with no hard dependencies to
43 a single implementation. In addition, the test case suite shall be
44 architecture independent.
46 The test cases developed in this project shall not form part of a
47 separate test framework, all of these tests may be inserted into the
48 Continuous Integration Test Framework and/or the Platform Functionality
49 Test Framework - if a vSwitch becomes a standard component of an OPNFV
57 * `RFC 1242 Benchmarking Terminology for Network Interconnection
58 Devices <http://www.ietf.org/rfc/rfc1242.txt>`__
59 * `RFC 2544 Benchmarking Methodology for Network Interconnect
60 Devices <http://www.ietf.org/rfc/rfc2544.txt>`__
61 * `RFC 2285 Benchmarking Terminology for LAN Switching
62 Devices <http://www.ietf.org/rfc/rfc2285.txt>`__
63 * `RFC 2889 Benchmarking Methodology for LAN Switching
64 Devices <http://www.ietf.org/rfc/rfc2889.txt>`__
65 * `RFC 3918 Methodology for IP Multicast
66 Benchmarking <http://www.ietf.org/rfc/rfc3918.txt>`__
67 * `RFC 4737 Packet Reordering
68 Metrics <http://www.ietf.org/rfc/rfc4737.txt>`__
69 * `RFC 5481 Packet Delay Variation Applicability
70 Statement <http://www.ietf.org/rfc/rfc5481.txt>`__
71 * `RFC 6201 Device Reset
72 Characterization <http://tools.ietf.org/html/rfc6201>`__
76 ===================================
77 Details of the Level Test Design
78 ===================================
80 This section describes the features to be tested (
81 :ref:_FeaturesToBeTested), the test approach (:ref:_Approach);
82 it also identifies the sets of test cases or scenarios (
83 :ref:_TestIdentification) along with the pass/fail criteria and
84 the test deliverables.
88 .. _FeaturesToBeTested:
91 ==========================
93 Characterizing virtual switches (i.e. Device Under Test (DUT) in this document)
94 includes measuring the following performance metrics:
96 - **Throughput** as defined by `RFC1242
97 <https://www.rfc-editor.org/rfc/rfc1242.txt>`__: The maximum rate at which
98 **none** of the offered frames are dropped by the DUT. The maximum frame
99 rate and bit rate that can be transmitted by the DUT without any error
100 should be recorded. Note there is an equivalent bit rate and a specific
101 layer at which the payloads contribute to the bits. Errors and
102 improperly formed frames or packets are dropped.
103 - **Packet delay** introduced by the DUT and its cumulative effect on
104 E2E networks. Frame delay can be measured equivalently.
105 - **Packet delay variation**: measured from the perspective of the
106 VNF/application. Packet delay variation is sometimes called "jitter".
107 However, we will avoid the term "jitter" as the term holds different
108 meaning to different groups of people. In this document we will
109 simply use the term packet delay variation. The preferred form for this
110 metric is the PDV form of delay variation defined in `RFC5481
111 <https://www.rfc-editor.org/rfc/rfc5481.txt>`__. The most relevant
112 measurement of PDV considers the delay variation of a single user flow,
113 as this will be relevant to the size of end-system buffers to compensate
114 for delay variation. The measurement system's ability to store the
115 delays of individual packets in the flow of interest is a key factor
116 that determines the specific measurement method. At the outset, it is
117 ideal to view the complete PDV distribution. Systems that can capture
118 and store packets and their delays have the freedom to calculate the
119 reference minimum delay and to determine various quantiles of the PDV
120 distribution accurately (in post-measurement processing routines).
121 Systems without storage must apply algorithms to calculate delay and
122 statistical measurements on the fly. For example, a system may store
123 temporary estimates of the mimimum delay and the set of (100) packets
124 with the longest delays during measurement (to calculate a high quantile,
125 and update these sets with new values periodically.
126 In some cases, a limited number of delay histogram bins will be
127 available, and the bin limits will need to be set using results from
128 repeated experiments. See section 8 of `RFC5481
129 <https://www.rfc-editor.org/rfc/rfc5481.txt>`__.
130 - **Packet loss** (within a configured waiting time at the receiver): All
131 packets sent to the DUT should be accounted for.
132 - **Burst behaviour**: measures the ability of the DUT to buffer packets.
133 - **Packet re-ordering**: measures the ability of the device under test to
134 maintain sending order throughout transfer to the destination.
135 - **Packet correctness**: packets or Frames must be well-formed, in that
136 they include all required fields, conform to length requirements, pass
137 integrity checks, etc.
138 - **Availability and capacity** of the DUT i.e. when the DUT is fully “up”
141 - Includes power consumption of the CPU (in various power states) and
143 - Includes CPU utilization.
144 - Includes the number of NIC interfaces supported.
145 - Includes headroom of VM workload processing cores (i.e. available
155 In order to determine the packet transfer characteristics of a virtual
156 switch, the tests will be broken down into the following categories:
161 ----------------------
162 - **Throughput Tests** to measure the maximum forwarding rate (in
163 frames per second or fps) and bit rate (in Mbps) for a constant load
164 (as defined by `RFC1242 <https://www.rfc-editor.org/rfc/rfc1242.txt>`__)
165 without traffic loss.
166 - **Packet and Frame Delay Tests** to measure average, min and max
167 packet and frame delay for constant loads.
168 - **Stream Performance Tests** (TCP, UDP) to measure bulk data transfer
169 performance, i.e. how fast systems can send and receive data through
171 - **Request/Response Performance** Tests (TCP, UDP) the measure the
172 transaction rate through the virtual switch.
173 - **Packet Delay Tests** to understand latency distribution for
174 different packet sizes and over an extended test run to uncover
176 - **Scalability Tests** to understand how the virtual switch performs
177 as the number of flows, active ports, complexity of the forwarding
178 logic's configuration... it has to deal with increases.
179 - **Control Path and Datapath Coupling** Tests, to understand how
180 closely coupled the datapath and the control path are as well as the
181 effect of this coupling on the performance of the DUT.
182 - **CPU and Memory Consumption Tests** to understand the virtual
183 switch’s footprint on the system, this includes:
188 * Time To Establish Flows Tests.
190 - **Noisy Neighbour Tests**, to understand the effects of resource
191 sharing on the performance of a virtual switch.
193 **Note:** some of the tests above can be conducted simultaneously where
194 the combined results would be insightful, for example Packet/Frame Delay
200 --------------------------
201 The following represents possible deployments which can help to
202 determine the performance of both the virtual switch and the datapath
207 Physical port → vSwitch → physical port
208 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
209 .. code-block:: console
212 +--------------------------------------------------+ |
213 | +--------------------+ | |
216 | +--------------+ +--------------+ | |
217 | | phy port | vSwitch | phy port | | |
218 +---+--------------+------------+--------------+---+ _|
222 +--------------------------------------------------+
224 | traffic generator |
226 +--------------------------------------------------+
230 Physical port → vSwitch → VNF → vSwitch → physical port
231 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
232 .. code-block:: console
235 +---------------------------------------------------+ |
237 | +-------------------------------------------+ | |
238 | | Application | | |
239 | +-------------------------------------------+ | |
243 | +---------------+ +---------------+ | |
244 | | logical port 0| | logical port 1| | |
245 +---+---------------+-----------+---------------+---+ _|
249 +---+---------------+----------+---------------+---+ |
250 | | logical port 0| | logical port 1| | |
251 | +---------------+ +---------------+ | |
255 | +--------------+ +--------------+ | |
256 | | phy port | vSwitch | phy port | | |
257 +---+--------------+------------+--------------+---+ _|
261 +--------------------------------------------------+
263 | traffic generator |
265 +--------------------------------------------------+
269 Physical port → vSwitch → VNF → vSwitch → VNF → vSwitch → physical port
270 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
272 .. code-block:: console
275 +----------------------+ +----------------------+ |
276 | Guest 1 | | Guest 2 | |
277 | +---------------+ | | +---------------+ | |
278 | | Application | | | | Application | | |
279 | +---------------+ | | +---------------+ | |
281 | | v | | | v | | Guests
282 | +---------------+ | | +---------------+ | |
283 | | logical ports | | | | logical ports | | |
284 | | 0 1 | | | | 0 1 | | |
285 +---+---------------+--+ +---+---------------+--+ _|
289 +---+---------------+---------+---------------+--+ |
290 | | 0 1 | | 3 4 | | |
291 | | logical ports | | logical ports | | |
292 | +---------------+ +---------------+ | |
294 | | L-----------------+ v | |
295 | +--------------+ +--------------+ | |
296 | | phy ports | vSwitch | phy ports | | |
297 +---+--------------+----------+--------------+---+ _|
301 +--------------------------------------------------+
303 | traffic generator |
305 +--------------------------------------------------+
309 Physical port → VNF → vSwitch → VNF → physical port
310 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
312 .. code-block:: console
315 +----------------------+ +----------------------+ |
316 | Guest 1 | | Guest 2 | |
317 |+-------------------+ | | +-------------------+| |
318 || Application | | | | Application || |
319 |+-------------------+ | | +-------------------+| |
320 | ^ | | | ^ | | | Guests
322 |+-------------------+ | | +-------------------+| |
323 || logical ports | | | | logical ports || |
324 || 0 1 | | | | 0 1 || |
325 ++--------------------++ ++--------------------++ _|
327 (PCI passthrough) | | (PCI passthrough)
329 +--------++------------+-+------------++---------+ |
330 | | || 0 | | 1 || | | |
331 | | ||logical port| |logical port|| | | |
332 | | |+------------+ +------------+| | | |
334 | | | L-----------------+ | | | |
336 | | | vSwitch | | | |
337 | | +-----------------------------+ | | |
340 | +--------------+ +--------------+ | |
341 | | phy port/VF | | phy port/VF | | |
342 +-+--------------+--------------+--------------+-+ _|
346 +--------------------------------------------------+
348 | traffic generator |
350 +--------------------------------------------------+
354 Physical port → vSwitch → VNF
355 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
357 .. code-block:: console
360 +---------------------------------------------------+ |
362 | +-------------------------------------------+ | |
363 | | Application | | |
364 | +-------------------------------------------+ | |
368 | +---------------+ | |
369 | | logical port 0| | |
370 +---+---------------+-------------------------------+ _|
374 +---+---------------+------------------------------+ |
375 | | logical port 0| | |
376 | +---------------+ | |
380 | +--------------+ | |
381 | | phy port | vSwitch | |
382 +---+--------------+------------ -------------- ---+ _|
386 +--------------------------------------------------+
388 | traffic generator |
390 +--------------------------------------------------+
394 VNF → vSwitch → physical port
395 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
397 .. code-block:: console
400 +---------------------------------------------------+ |
402 | +-------------------------------------------+ | |
403 | | Application | | |
404 | +-------------------------------------------+ | |
408 | +---------------+ | |
409 | | logical port | | |
410 +-------------------------------+---------------+---+ _|
414 +------------------------------+---------------+---+ |
415 | | logical port | | |
416 | +---------------+ | |
420 | +--------------+ | |
421 | vSwitch | phy port | | |
422 +-------------------------------+--------------+---+ _|
426 +--------------------------------------------------+
428 | traffic generator |
430 +--------------------------------------------------+
434 VNF → vSwitch → VNF → vSwitch
435 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
437 .. code-block:: console
440 +-------------------------+ +-------------------------+ |
441 | Guest 1 | | Guest 2 | |
442 | +-----------------+ | | +-----------------+ | |
443 | | Application | | | | Application | | |
444 | +-----------------+ | | +-----------------+ | |
448 | +---------------+ | | +---------------+ | |
449 | | logical port 0| | | | logical port 0| | |
450 +-----+---------------+---+ +---+---------------+-----+ _|
454 +----+---------------+------------+---------------+-----+ |
455 | | port 0 | | port 1 | | |
456 | +---------------+ +---------------+ | |
459 | +--------------------+ | |
462 +-------------------------------------------------------+ _|
466 HOST 1(Physical port → virtual switch → VNF → virtual switch → Physical port)
467 → HOST 2(Physical port → virtual switch → VNF → virtual switch → Physical port)
469 HOST 1 (PVP) → HOST 2 (PVP)
470 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
472 .. code-block:: console
475 +----------------------+ +----------------------+ |
476 | Guest 1 | | Guest 2 | |
477 | +---------------+ | | +---------------+ | |
478 | | Application | | | | Application | | |
479 | +---------------+ | | +---------------+ | |
481 | | v | | | v | | Guests
482 | +---------------+ | | +---------------+ | |
483 | | logical ports | | | | logical ports | | |
484 | | 0 1 | | | | 0 1 | | |
485 +---+---------------+--+ +---+---------------+--+ _|
489 +---+---------------+--+ +---+---------------+--+ |
490 | | 0 1 | | | | 3 4 | | |
491 | | logical ports | | | | logical ports | | |
492 | +---------------+ | | +---------------+ | |
493 | ^ | | | ^ | | | Hosts
495 | +--------------+ | | +--------------+ | |
496 | | phy ports | | | | phy ports | | |
497 +---+--------------+---+ +---+--------------+---+ _|
499 | +-----------------+ |
501 +--------------------------------------------------+
503 | traffic generator |
505 +--------------------------------------------------+
509 **Note:** For tests where the traffic generator and/or measurement
510 receiver are implemented on VM and connected to the virtual switch
511 through vNIC, the issues of shared resources and interactions between
512 the measurement devices and the device under test must be considered.
514 **Note:** Some RFC 2889 tests require a full-mesh sending and receiving
515 pattern involving more than two ports. This possibility is illustrated in the
516 Physical port → vSwitch → VNF → vSwitch → VNF → vSwitch → physical port
517 diagram above (with 2 sending and 2 receiving ports, though all ports
518 could be used bi-directionally).
520 **Note:** When Deployment Scenarios are used in RFC 2889 address learning
521 or cache capacity testing, an additional port from the vSwitch must be
522 connected to the test device. This port is used to listen for flooded
528 --------------------------
529 To establish the baseline performance of the virtual switch, tests would
530 initially be run with a simple workload in the VNF (the recommended
531 simple workload VNF would be `DPDK <http://www.dpdk.org/>`__'s testpmd
532 application forwarding packets in a VM or vloop\_vnf a simple kernel
533 module that forwards traffic between two network interfaces inside the
534 virtualized environment while bypassing the networking stack).
535 Subsequently, the tests would also be executed with a real Telco
536 workload running in the VNF, which would exercise the virtual switch in
537 the context of higher level Telco NFV use cases, and prove that its
538 underlying characteristics and behaviour can be measured and validated.
539 Suitable real Telco workload VNFs are yet to be identified.
543 Default Test Parameters
544 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
546 The following list identifies the default parameters for suite of
549 - Reference application: Simple forwarding or Open Source VNF.
550 - Frame size (bytes): 64, 128, 256, 512, 1024, 1280, 1518, 2K, 4k OR
551 Packet size based on use-case (e.g. RTP 64B, 256B) OR Mix of packet sizes as
552 maintained by the Functest project <https://wiki.opnfv.org/traffic_profile_management>.
553 - Reordering check: Tests should confirm that packets within a flow are
555 - Duplex: Unidirectional / Bidirectional. Default: Full duplex with
556 traffic transmitting in both directions, as network traffic generally
557 does not flow in a single direction. By default the data rate of
558 transmitted traffic should be the same in both directions, please
559 note that asymmetric traffic (e.g. downlink-heavy) tests will be
560 mentioned explicitly for the relevant test cases.
561 - Number of Flows: Default for non scalability tests is a single flow.
562 For scalability tests the goal is to test with maximum supported
563 flows but where possible will test up to 10 Million flows. Start with
564 a single flow and scale up. By default flows should be added
565 sequentially, tests that add flows simultaneously will explicitly
566 call out their flow addition behaviour. Packets are generated across
567 the flows uniformly with no burstiness. For multi-core tests should
568 consider the number of packet flows based on vSwitch/VNF multi-thread
569 implementation and behavior.
571 - Traffic Types: UDP, SCTP, RTP, GTP and UDP traffic.
572 - Deployment scenarios are:
573 - Physical → virtual switch → physical.
574 - Physical → virtual switch → VNF → virtual switch → physical.
575 - Physical → virtual switch → VNF → virtual switch → VNF → virtual
577 - Physical → VNF → virtual switch → VNF → physical.
578 - Physical → virtual switch → VNF.
579 - VNF → virtual switch → Physical.
580 - VNF → virtual switch → VNF.
582 Tests MUST have these parameters unless otherwise stated. **Test cases
583 with non default parameters will be stated explicitly**.
585 **Note**: For throughput tests unless stated otherwise, test
586 configurations should ensure that traffic traverses the installed flows
587 through the virtual switch, i.e. flows are installed and have an appropriate
588 time out that doesn't expire before packet transmission starts.
593 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
595 Virtual switches classify packets into flows by processing and matching
596 particular header fields in the packet/frame and/or the input port where
597 the packets/frames arrived. The vSwitch then carries out an action on
598 the group of packets that match the classification parameters. Thus a
599 flow is considered to be a sequence of packets that have a shared set of
600 header field values or have arrived on the same port and have the same
601 action applied to them. Performance results can vary based on the
602 parameters the vSwitch uses to match for a flow. The recommended flow
603 classification parameters for L3 vSwitch performance tests are: the
604 input port, the source IP address, the destination IP address and the
605 Ethernet protocol type field. It is essential to increase the flow
606 time-out time on a vSwitch before conducting any performance tests that
607 do not measure the flow set-up time. Normally the first packet of a
608 particular flow will install the flow in the vSwitch which adds an
609 additional latency, subsequent packets of the same flow are not subject
610 to this latency if the flow is already installed on the vSwitch.
615 ~~~~~~~~~~~~~~~~~~~~~
617 Tests will be assigned a priority in order to determine which tests
618 should be implemented immediately and which tests implementations
621 Priority can be of following types: - Urgent: Must be implemented
622 immediately. - High: Must be implemented in the next release. - Medium:
623 May be implemented after the release. - Low: May or may not be
631 The SUT should be configured to its "default" state. The
632 SUT's configuration or set-up must not change between tests in any way
633 other than what is required to do the test. All supported protocols must
634 be configured and enabled for each test set up.
639 ~~~~~~~~~~~~~~~~~~~~~~~~~~
641 The DUT should be configured with n ports where
642 n is a multiple of 2. Half of the ports on the DUT should be used as
643 ingress ports and the other half of the ports on the DUT should be used
644 as egress ports. Where a DUT has more than 2 ports, the ingress data
645 streams should be set-up so that they transmit packets to the egress
646 ports in sequence so that there is an even distribution of traffic
647 across ports. For example, if a DUT has 4 ports 0(ingress), 1(ingress),
648 2(egress) and 3(egress), the traffic stream directed at port 0 should
649 output a packet to port 2 followed by a packet to port 3. The traffic
650 stream directed at port 1 should also output a packet to port 2 followed
651 by a packet to port 3.
656 ~~~~~~~~~~~~~~~~~~~~~
658 **Frame formats Layer 2 (data link layer) protocols**
662 .. code-block:: console
664 +---------------------------+-----------+
665 | Ethernet Header | Payload | Check Sum |
666 +-----------------+---------+-----------+
667 |_________________|_________|___________|
668 14 Bytes 46 - 1500 4 Bytes
672 **Layer 3 (network layer) protocols**
676 .. code-block:: console
678 +-----------------+-----------+---------+-----------+
679 | Ethernet Header | IP Header | Payload | Checksum |
680 +-----------------+-----------+---------+-----------+
681 |_________________|___________|_________|___________|
682 14 Bytes 20 bytes 26 - 1480 4 Bytes
687 .. code-block:: console
689 +-----------------+-----------+---------+-----------+
690 | Ethernet Header | IP Header | Payload | Checksum |
691 +-----------------+-----------+---------+-----------+
692 |_________________|___________|_________|___________|
693 14 Bytes 40 bytes 26 - 1460 4 Bytes
696 **Layer 4 (transport layer) protocols**
702 .. code-block:: console
704 +-----------------+-----------+-----------------+---------+-----------+
705 | Ethernet Header | IP Header | Layer 4 Header | Payload | Checksum |
706 +-----------------+-----------+-----------------+---------+-----------+
707 |_________________|___________|_________________|_________|___________|
708 14 Bytes 40 bytes 20 Bytes 6 - 1460 4 Bytes
712 **Layer 5 (application layer) protocols**
717 .. code-block:: console
719 +-----------------+-----------+-----------------+---------+-----------+
720 | Ethernet Header | IP Header | Layer 4 Header | Payload | Checksum |
721 +-----------------+-----------+-----------------+---------+-----------+
722 |_________________|___________|_________________|_________|___________|
723 14 Bytes 20 bytes 20 Bytes >= 6 Bytes 4 Bytes
728 ~~~~~~~~~~~~~~~~~~~~~~~~~
729 There is a difference between an Ethernet frame,
730 an IP packet, and a UDP datagram. In the seven-layer OSI model of
731 computer networking, packet refers to a data unit at layer 3 (network
732 layer). The correct term for a data unit at layer 2 (data link layer) is
733 a frame, and at layer 4 (transport layer) is a segment or datagram.
735 Important concepts related to 10GbE performance are frame rate and
736 throughput. The MAC bit rate of 10GbE, defined in the IEEE standard 802
737 .3ae, is 10 billion bits per second. Frame rate is based on the bit rate
738 and frame format definitions. Throughput, defined in IETF RFC 1242, is
739 the highest rate at which the system under test can forward the offered
742 The frame rate for 10GbE is determined by a formula that divides the 10
743 billion bits per second by the preamble + frame length + inter-frame
746 The maximum frame rate is calculated using the minimum values of the
747 following parameters, as described in the IEEE 802 .3ae standard:
749 - Preamble: 8 bytes \* 8 = 64 bits
750 - Frame Length: 64 bytes (minimum) \* 8 = 512 bits
751 - Inter-frame Gap: 12 bytes (minimum) \* 8 = 96 bits
753 Therefore, Maximum Frame Rate (64B Frames)
754 = MAC Transmit Bit Rate / (Preamble + Frame Length + Inter-frame Gap)
755 = 10,000,000,000 / (64 + 512 + 96)
756 = 10,000,000,000 / 672
757 = 14,880,952.38 frame per second (fps)
761 System isolation and validation
762 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
764 A key consideration when conducting any sort of benchmark is trying to
765 ensure the consistency and repeatability of test results between runs.
766 When benchmarking the performance of a virtual switch there are many
767 factors that can affect the consistency of results. This section
768 describes these factors and the measures that can be taken to limit
769 their effects. In addition, this section will outline some system tests
770 to validate the platform and the VNF before conducting any vSwitch
773 **System Isolation:**
775 When conducting a benchmarking test on any SUT, it is essential to limit
776 (and if reasonable, eliminate) any noise that may interfere with the
777 accuracy of the metrics collected by the test. This noise may be
778 introduced by other hardware or software (OS, other applications), and
779 can result in significantly varying performance metrics being collected
780 between consecutive runs of the same test. In the case of characterizing
781 the performance of a virtual switch, there are a number of configuration
782 parameters that can help increase the repeatability and stability of
783 test results, including:
785 - OS/GRUB configuration:
787 - maxcpus = n where n >= 0; limits the kernel to using 'n'
788 processors. Only use exactly what you need.
789 - isolcpus: Isolate CPUs from the general scheduler. Isolate all
790 CPUs bar one which will be used by the OS.
791 - use taskset to affinitize the forwarding application and the VNFs
792 onto isolated cores. VNFs and the vSwitch should be allocated
793 their own cores, i.e. must not share the same cores. vCPUs for the
794 VNF should be affinitized to individual cores also.
795 - Limit the amount of background applications that are running and
796 set OS to boot to runlevel 3. Make sure to kill any unnecessary
797 system processes/daemons.
798 - Only enable hardware that you need to use for your test – to
799 ensure there are no other interrupts on the system.
800 - Configure NIC interrupts to only use the cores that are not
801 allocated to any other process (VNF/vSwitch).
803 - NUMA configuration: Any unused sockets in a multi-socket system
805 - CPU pinning: The vSwitch and the VNF should each be affinitized to
806 separate logical cores using a combination of maxcpus, isolcpus and
808 - BIOS configuration: BIOS should be configured for performance where
809 an explicit option exists, sleep states should be disabled, any
810 virtualization optimization technologies should be enabled, and
811 hyperthreading should also be enabled.
813 **System Validation:**
815 System validation is broken down into two sub-categories: Platform
816 validation and VNF validation. The validation test itself involves
817 verifying the forwarding capability and stability for the sub-system
818 under test. The rationale behind system validation is two fold. Firstly
819 to give a tester confidence in the stability of the platform or VNF that
820 is being tested; and secondly to provide base performance comparison
821 points to understand the overhead introduced by the virtual switch.
823 * Benchmark platform forwarding capability: This is an OPTIONAL test
824 used to verify the platform and measure the base performance (maximum
825 forwarding rate in fps and latency) that can be achieved by the
826 platform without a vSwitch or a VNF. The following diagram outlines
827 the set-up for benchmarking Platform forwarding capability:
829 .. code-block:: console
832 +--------------------------------------------------+ |
833 | +------------------------------------------+ | |
835 | | l2fw or DPDK L2FWD app | | Host
837 | +------------------------------------------+ | |
839 +---+------------------------------------------+---+ __|
843 +--------------------------------------------------+
845 | traffic generator |
847 +--------------------------------------------------+
849 * Benchmark VNF forwarding capability: This test is used to verify
850 the VNF and measure the base performance (maximum forwarding rate in
851 fps and latency) that can be achieved by the VNF without a vSwitch.
852 The performance metrics collected by this test will serve as a key
853 comparison point for NIC passthrough technologies and vSwitches. VNF
854 in this context refers to the hypervisor and the VM. The following
855 diagram outlines the set-up for benchmarking VNF forwarding
858 .. code-block:: console
861 +--------------------------------------------------+ |
862 | +------------------------------------------+ | |
866 | +------------------------------------------+ | |
867 | | Passthrough/SR-IOV | | Host
868 | +------------------------------------------+ | |
870 +---+------------------------------------------+---+ __|
874 +--------------------------------------------------+
876 | traffic generator |
878 +--------------------------------------------------+
881 **Methodology to benchmark Platform/VNF forwarding capability**
884 The recommended methodology for the platform/VNF validation and
885 benchmark is: - Run `RFC2889 <https://www.rfc-editor.org/rfc/rfc2289.txt>`__
886 Maximum Forwarding Rate test, this test will produce maximum
887 forwarding rate and latency results that will serve as the
888 expected values. These expected values can be used in
889 subsequent steps or compared with in subsequent validation tests. -
890 Transmit bidirectional traffic at line rate/max forwarding rate
891 (whichever is higher) for at least 72 hours, measure throughput (fps)
892 and latency. - Note: Traffic should be bidirectional. - Establish a
893 baseline forwarding rate for what the platform can achieve. - Additional
894 validation: After the test has completed for 72 hours run bidirectional
895 traffic at the maximum forwarding rate once more to see if the system is
896 still functional and measure throughput (fps) and latency. Compare the
897 measure the new obtained values with the expected values.
899 **NOTE 1**: How the Platform is configured for its forwarding capability
900 test (BIOS settings, GRUB configuration, runlevel...) is how the
901 platform should be configured for every test after this
903 **NOTE 2**: How the VNF is configured for its forwarding capability test
904 (# of vCPUs, vNICs, Memory, affinitization…) is how it should be
905 configured for every test that uses a VNF after this.
909 RFCs for testing virtual switch performance
910 --------------------------------------------------
912 The starting point for defining the suite of tests for benchmarking the
913 performance of a virtual switch is to take existing RFCs and standards
914 that were designed to test their physical counterparts and adapting them
915 for testing virtual switches. The rationale behind this is to establish
916 a fair comparison between the performance of virtual and physical
917 switches. This section outlines the RFCs that are used by this
922 RFC 1242 Benchmarking Terminology for Network Interconnection
923 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
924 Devices RFC 1242 defines the terminology that is used in describing
925 performance benchmarking tests and their results. Definitions and
926 discussions covered include: Back-to-back, bridge, bridge/router,
927 constant load, data link frame size, frame loss rate, inter frame gap,
928 latency, and many more.
932 RFC 2544 Benchmarking Methodology for Network Interconnect Devices
933 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
934 RFC 2544 outlines a benchmarking methodology for network Interconnect
935 Devices. The methodology results in performance metrics such as latency,
936 frame loss percentage, and maximum data throughput.
938 In this document network “throughput” (measured in millions of frames
939 per second) is based on RFC 2544, unless otherwise noted. Frame size
940 refers to Ethernet frames ranging from smallest frames of 64 bytes to
941 largest frames of 4K bytes.
945 1. Throughput test defines the maximum number of frames per second
946 that can be transmitted without any error.
948 2. Latency test measures the time required for a frame to travel from
949 the originating device through the network to the destination device.
950 Please note that RFC2544 Latency measurement will be superseded with
951 a measurement of average latency over all successfully transferred
954 3. Frame loss test measures the network’s
955 response in overload conditions - a critical indicator of the
956 network’s ability to support real-time applications in which a
957 large amount of frame loss will rapidly degrade service quality.
959 4. Burst test assesses the buffering capability of a virtual switch. It
960 measures the maximum number of frames received at full line rate
961 before a frame is lost. In carrier Ethernet networks, this
962 measurement validates the excess information rate (EIR) as defined in
965 5. System recovery to characterize speed of recovery from an overload
968 6. Reset to characterize speed of recovery from device or software
969 reset. This type of test has been updated by `RFC6201
970 <https://www.rfc-editor.org/rfc/rfc6201.txt>`__ as such,
971 the methodology defined by this specification will be that of RFC 6201.
973 Although not included in the defined RFC 2544 standard, another crucial
974 measurement in Ethernet networking is packet delay variation. The
975 definition set out by this specification comes from
976 `RFC5481 <https://www.rfc-editor.org/rfc/rfc5481.txt>`__.
980 RFC 2285 Benchmarking Terminology for LAN Switching Devices
981 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
982 RFC 2285 defines the terminology that is used to describe the
983 terminology for benchmarking a LAN switching device. It extends RFC
984 1242 and defines: DUTs, SUTs, Traffic orientation and distribution,
985 bursts, loads, forwarding rates, etc.
989 RFC 2889 Benchmarking Methodology for LAN Switching
990 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
991 RFC 2889 outlines a benchmarking methodology for LAN switching, it
992 extends RFC 2544. The outlined methodology gathers performance
993 metrics for forwarding, congestion control, latency, address handling
994 and finally filtering.
998 RFC 3918 Methodology for IP Multicast Benchmarking
999 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1000 RFC 3918 outlines a methodology for IP Multicast benchmarking.
1004 RFC 4737 Packet Reordering Metrics
1005 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1006 RFC 4737 describes metrics for identifying and counting re-ordered
1007 packets within a stream, and metrics to measure the extent each
1008 packet has been re-ordered.
1012 RFC 5481 Packet Delay Variation Applicability Statement
1013 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1014 RFC 5481 defined two common, but different forms of delay variation
1015 metrics, and compares the metrics over a range of networking
1016 circumstances and tasks. The most suitable form for vSwitch
1017 benchmarking is the "PDV" form.
1021 RFC 6201 Device Reset Characterization
1022 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1023 RFC 6201 extends the methodology for characterizing the speed of
1024 recovery of the DUT from device or software reset described in RFC
1029 Details of the Test Report
1030 ---------------------------------
1032 There are a number of parameters related to the system, DUT and tests
1033 that can affect the repeatability of a test results and should be
1034 recorded. In order to minimise the variation in the results of a test,
1035 it is recommended that the test report includes the following information:
1037 - Hardware details including:
1040 - Processor details.
1041 - Memory information (see below)
1042 - Number of enabled cores.
1043 - Number of cores used for the test.
1044 - Number of physical NICs, as well as their details (manufacturer,
1045 versions, type and the PCI slot they are plugged into).
1046 - NIC interrupt configuration.
1047 - BIOS version, release date and any configurations that were
1050 - Software details including:
1052 - OS version (for host and VNF)
1053 - Kernel version (for host and VNF)
1054 - GRUB boot parameters (for host and VNF).
1055 - Hypervisor details (Type and version).
1056 - Selected vSwitch, version number or commit id used.
1057 - vSwitch launch command line if it has been parameterised.
1058 - Memory allocation to the vSwitch – which NUMA node it is using,
1059 and how many memory channels.
1060 - Where the vswitch is built from source: compiler details including
1061 versions and the flags that were used to compile the vSwitch.
1062 - DPDK or any other SW dependency version number or commit id used.
1063 - Memory allocation to a VM - if it's from Hugpages/elsewhere.
1064 - VM storage type: snapshot/independent persistent/independent
1067 - Number of Virtual NICs (vNICs), versions, type and driver.
1068 - Number of virtual CPUs and their core affinity on the host.
1069 - Number vNIC interrupt configuration.
1070 - Thread affinitization for the applications (including the vSwitch
1071 itself) on the host.
1072 - Details of Resource isolation, such as CPUs designated for
1073 Host/Kernel (isolcpu) and CPUs designated for specific processes
1092 - Traffic Information:
1094 - Traffic type - UDP, TCP, IMIX / Other.
1097 - Deployment Scenario.
1099 **Note**: Tests that require additional parameters to be recorded will
1100 explicitly specify this.
1102 .. _TestIdentification:
1107 =========================
1112 ----------------------
1113 The following tests aim to determine the maximum forwarding rate that
1114 can be achieved with a virtual switch. The list is not exhaustive but
1115 should indicate the type of tests that should be required. It is
1116 expected that more will be added.
1120 Test ID: LTD.Throughput.RFC2544.PacketLossRatio
1121 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1122 **Title**: RFC 2544 X% packet loss ratio Throughput and Latency Test
1124 **Prerequisite Test**: N/A
1130 This test determines the DUT's maximum forwarding rate with X% traffic
1131 loss for a constant load (fixed length frames at a fixed interval time).
1132 The default loss percentages to be tested are: - X = 0% - X = 10^-7%
1134 Note: Other values can be tested if required by the user.
1136 The selected frame sizes are those previously defined under `Default
1137 Test Parameters <#DefaultParams>`__. The test can also be used to
1138 determine the average latency of the traffic.
1140 Under the `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__
1141 test methodology, the test duration will
1142 include a number of trials; each trial should run for a minimum period
1143 of 60 seconds. A binary search methodology must be applied for each
1144 trial to obtain the final result.
1146 **Expected Result**: At the end of each trial, the presence or absence
1147 of loss determines the modification of offered load for the next trial,
1148 converging on a maximum rate, or
1149 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__ Throughput with X%
1151 The Throughput load is re-used in related
1152 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__ tests and other
1155 **Metrics Collected**:
1157 The following are the metrics collected for this test:
1159 - The maximum forwarding rate in Frames Per Second (FPS) and Mbps of
1160 the DUT for each frame size with X% packet loss.
1161 - The average latency of the traffic flow when passing through the DUT
1162 (if testing for latency, note that this average is different from the
1163 test specified in Section 26.3 of
1164 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__).
1165 - CPU and memory utilization may also be collected as part of this
1166 test, to determine the vSwitch's performance footprint on the system.
1170 Test ID: LTD.Throughput.RFC2544.PacketLossRatioFrameModification
1171 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1172 **Title**: RFC 2544 X% packet loss Throughput and Latency Test with
1175 **Prerequisite Test**: N/A
1181 This test determines the DUT's maximum forwarding rate with X% traffic
1182 loss for a constant load (fixed length frames at a fixed interval time).
1183 The default loss percentages to be tested are: - X = 0% - X = 10^-7%
1185 Note: Other values can be tested if required by the user.
1187 The selected frame sizes are those previously defined under `Default
1188 Test Parameters <#DefaultParams>`__. The test can also be used to
1189 determine the average latency of the traffic.
1191 Under the `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__
1192 test methodology, the test duration will
1193 include a number of trials; each trial should run for a minimum period
1194 of 60 seconds. A binary search methodology must be applied for each
1195 trial to obtain the final result.
1197 During this test, the DUT must perform the following operations on the
1200 - Perform packet parsing on the DUT's ingress port.
1201 - Perform any relevant address look-ups on the DUT's ingress ports.
1202 - Modify the packet header before forwarding the packet to the DUT's
1203 egress port. Packet modifications include:
1205 - Modifying the Ethernet source or destination MAC address.
1206 - Modifying/adding a VLAN tag. (**Recommended**).
1207 - Modifying/adding a MPLS tag.
1208 - Modifying the source or destination ip address.
1209 - Modifying the TOS/DSCP field.
1210 - Modifying the source or destination ports for UDP/TCP/SCTP.
1211 - Modifying the TTL.
1213 **Expected Result**: The Packet parsing/modifications require some
1214 additional degree of processing resource, therefore the
1215 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__
1216 Throughput is expected to be somewhat lower than the Throughput level
1217 measured without additional steps. The reduction is expected to be
1218 greatest on tests with the smallest packet sizes (greatest header
1221 **Metrics Collected**:
1223 The following are the metrics collected for this test:
1225 - The maximum forwarding rate in Frames Per Second (FPS) and Mbps of
1226 the DUT for each frame size with X% packet loss and packet
1227 modification operations being performed by the DUT.
1228 - The average latency of the traffic flow when passing through the DUT
1229 (if testing for latency, note that this average is different from the
1230 test specified in Section 26.3 of
1231 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__).
1232 - The `RFC5481 <https://www.rfc-editor.org/rfc/rfc5481.txt>`__
1233 PDV form of delay variation on the traffic flow,
1234 using the 99th percentile.
1235 - CPU and memory utilization may also be collected as part of this
1236 test, to determine the vSwitch's performance footprint on the system.
1240 Test ID: LTD.Throughput.RFC2544.Profile
1241 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1242 **Title**: RFC 2544 Throughput and Latency Profile
1244 **Prerequisite Test**: N/A
1250 This test reveals how throughput and latency degrades as the offered
1251 rate varies in the region of the DUT's maximum forwarding rate as
1252 determined by LTD.Throughput.RFC2544.PacketLossRatio (0% Packet Loss).
1253 For example it can be used to determine if the degradation of throughput
1254 and latency as the offered rate increases is slow and graceful or sudden
1257 The selected frame sizes are those previously defined under `Default
1258 Test Parameters <#DefaultParams>`__.
1260 The offered traffic rate is described as a percentage delta with respect
1261 to the DUT's RFC 2544 Throughput as determined by
1262 LTD.Throughput.RFC2544.PacketLoss Ratio (0% Packet Loss case). A delta
1263 of 0% is equivalent to an offered traffic rate equal to the RFC 2544
1264 Throughput; A delta of +50% indicates an offered rate half-way
1265 between the Throughput and line-rate, whereas a delta of
1266 -50% indicates an offered rate of half the maximum rate. Therefore the
1267 range of the delta figure is natuarlly bounded at -100% (zero offered
1268 traffic) and +100% (traffic offered at line rate).
1270 The following deltas to the maximum forwarding rate should be applied:
1272 - -50%, -10%, 0%, +10% & +50%
1274 **Expected Result**: For each packet size a profile should be produced
1275 of how throughput and latency vary with offered rate.
1277 **Metrics Collected**:
1279 The following are the metrics collected for this test:
1281 - The forwarding rate in Frames Per Second (FPS) and Mbps of the DUT
1282 for each delta to the maximum forwarding rate and for each frame
1284 - The average latency for each delta to the maximum forwarding rate and
1285 for each frame size.
1286 - CPU and memory utilization may also be collected as part of this
1287 test, to determine the vSwitch's performance footprint on the system.
1288 - Any failures experienced (for example if the vSwitch crashes, stops
1289 processing packets, restarts or becomes unresponsive to commands)
1290 when the offered load is above Maximum Throughput MUST be recorded
1291 and reported with the results.
1295 Test ID: LTD.Throughput.RFC2544.SystemRecoveryTime
1296 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1297 **Title**: RFC 2544 System Recovery Time Test
1299 **Prerequisite Test** LTD.Throughput.RFC2544.PacketLossRatio
1305 The aim of this test is to determine the length of time it takes the DUT
1306 to recover from an overload condition for a constant load (fixed length
1307 frames at a fixed interval time). The selected frame sizes are those
1308 previously defined under `Default Test Parameters <#DefaultParams>`__,
1309 traffic should be sent to the DUT under normal conditions. During the
1310 duration of the test and while the traffic flows are passing though the
1311 DUT, at least one situation leading to an overload condition for the DUT
1312 should occur. The time from the end of the overload condition to when
1313 the DUT returns to normal operations should be measured to determine
1314 recovery time. Prior to overloading the DUT, one should record the
1315 average latency for 10,000 packets forwarded through the DUT.
1317 The overload condition SHOULD be to transmit traffic at a very high
1318 frame rate to the DUT (150% of the maximum 0% packet loss rate as
1319 determined by LTD.Throughput.RFC2544.PacketLossRatio or line-rate
1320 whichever is lower), for at least 60 seconds, then reduce the frame rate
1321 to 75% of the maximum 0% packet loss rate. A number of time-stamps
1322 should be recorded: - Record the time-stamp at which the frame rate was
1323 reduced and record a second time-stamp at the time of the last frame
1324 lost. The recovery time is the difference between the two timestamps. -
1325 Record the average latency for 10,000 frames after the last frame loss
1326 and continue to record average latency measurements for every 10,000
1327 frames, when latency returns to within 10% of pre-overload levels record
1330 **Expected Result**:
1332 **Metrics collected**
1334 The following are the metrics collected for this test:
1336 - The length of time it takes the DUT to recover from an overload
1338 - The length of time it takes the DUT to recover the average latency to
1339 pre-overload conditions.
1341 **Deployment scenario**:
1343 - Physical → virtual switch → physical.
1347 Test ID: LTD.Throughput.RFC2544.BackToBackFrames
1348 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1349 **Title**: RFC2544 Back To Back Frames Test
1351 **Prerequisite Test**: N
1357 The aim of this test is to characterize the ability of the DUT to
1358 process back-to-back frames. For each frame size previously defined
1359 under `Default Test Parameters <#DefaultParams>`__, a burst of traffic
1360 is sent to the DUT with the minimum inter-frame gap between each frame.
1361 If the number of received frames equals the number of frames that were
1362 transmitted, the burst size should be increased and traffic is sent to
1363 the DUT again. The value measured is the back-to-back value, that is the
1364 maximum burst size the DUT can handle without any frame loss. Please note
1365 a trial must run for a minimum of 2 seconds and should be repeated 50
1366 times (at a minimum).
1368 **Expected Result**:
1370 Tests of back-to-back frames with physical devices have produced
1371 unstable results in some cases. All tests should be repeated in multiple
1372 test sessions and results stability should be examined.
1374 **Metrics collected**
1376 The following are the metrics collected for this test:
1378 - The average back-to-back value across the trials, which is
1379 the number of frames in the longest burst that the DUT will
1380 handle without the loss of any frames.
1381 - CPU and memory utilization may also be collected as part of this
1382 test, to determine the vSwitch's performance footprint on the system.
1384 **Deployment scenario**:
1386 - Physical → virtual switch → physical.
1390 Test ID: LTD.Throughput.RFC2889.MaxForwardingRateSoak
1391 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1392 **Title**: RFC 2889 X% packet loss Max Forwarding Rate Soak Test
1394 **Prerequisite Test** LTD.Throughput.RFC2544.PacketLossRatio
1400 The aim of this test is to understand the Max Forwarding Rate stability
1401 over an extended test duration in order to uncover any outliers. To allow
1402 for an extended test duration, the test should ideally run for 24 hours
1403 or, if this is not possible, for at least 6 hours. For this test, each frame
1404 size must be sent at the highest Throughput rate with X% packet loss, as
1405 determined in the prerequisite test. The default loss percentages to be
1406 tested are: - X = 0% - X = 10^-7%
1408 Note: Other values can be tested if required by the user.
1410 **Expected Result**:
1412 **Metrics Collected**:
1414 The following are the metrics collected for this test:
1416 - Max Forwarding Rate stability of the DUT.
1418 - This means reporting the number of packets lost per time interval
1419 and reporting any time intervals with packet loss. The
1420 `RFC2889 <https://www.rfc-editor.org/rfc/rfc2289.txt>`__
1421 Forwarding Rate shall be measured in each interval.
1422 An interval of 60s is suggested.
1424 - CPU and memory utilization may also be collected as part of this
1425 test, to determine the vSwitch's performance footprint on the system.
1426 - The `RFC5481 <https://www.rfc-editor.org/rfc/rfc5481.txt>`__
1427 PDV form of delay variation on the traffic flow,
1428 using the 99th percentile.
1432 Test ID: LTD.Throughput.RFC2889.MaxForwardingRateSoakFrameModification
1433 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1434 **Title**: RFC 2889 Max Forwarding Rate Soak Test with Frame Modification
1436 **Prerequisite Test**:
1437 LTD.Throughput.RFC2544.PacketLossRatioFrameModification (0% Packet Loss)
1443 The aim of this test is to understand the Max Forwarding Rate stability over an
1444 extended test duration in order to uncover any outliers. To allow for an
1445 extended test duration, the test should ideally run for 24 hours or, if
1446 this is not possible, for at least 6 hour. For this test, each frame
1447 size must be sent at the highest Throughput rate with 0% packet loss, as
1448 determined in the prerequisite test.
1450 During this test, the DUT must perform the following operations on the
1453 - Perform packet parsing on the DUT's ingress port.
1454 - Perform any relevant address look-ups on the DUT's ingress ports.
1455 - Modify the packet header before forwarding the packet to the DUT's
1456 egress port. Packet modifications include:
1458 - Modifying the Ethernet source or destination MAC address.
1459 - Modifying/adding a VLAN tag (**Recommended**).
1460 - Modifying/adding a MPLS tag.
1461 - Modifying the source or destination ip address.
1462 - Modifying the TOS/DSCP field.
1463 - Modifying the source or destination ports for UDP/TCP/SCTP.
1464 - Modifying the TTL.
1466 **Expected Result**:
1468 **Metrics Collected**:
1470 The following are the metrics collected for this test:
1472 - Max Forwarding Rate stability of the DUT.
1474 - This means reporting the number of packets lost per time interval
1475 and reporting any time intervals with packet loss. The
1476 `RFC2889 <https://www.rfc-editor.org/rfc/rfc2289.txt>`__
1477 Forwarding Rate shall be measured in each interval.
1478 An interval of 60s is suggested.
1480 - CPU and memory utilization may also be collected as part of this
1481 test, to determine the vSwitch's performance footprint on the system.
1482 - The `RFC5481 <https://www.rfc-editor.org/rfc/rfc5481.txt>`__
1483 PDV form of delay variation on the traffic flow, using the 99th
1488 Test ID: LTD.Throughput.RFC6201.ResetTime
1489 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1490 **Title**: RFC 6201 Reset Time Test
1492 **Prerequisite Test**: N/A
1498 The aim of this test is to determine the length of time it takes the DUT
1499 to recover from a reset.
1501 Two reset methods are defined - planned and unplanned. A planned reset
1502 requires stopping and restarting the virtual switch by the usual
1503 'graceful' method defined by it's documentation. An unplanned reset
1504 requires simulating a fatal internal fault in the virtual switch - for
1505 example by using kill -SIGKILL on a Linux environment.
1507 Both reset methods SHOULD be exercised.
1509 For each frame size previously defined under `Default Test
1510 Parameters <#DefaultParams>`__, traffic should be sent to the DUT under
1511 normal conditions. During the duration of the test and while the traffic
1512 flows are passing through the DUT, the DUT should be reset and the Reset
1513 time measured. The Reset time is the total time that a device is
1514 determined to be out of operation and includes the time to perform the
1515 reset and the time to recover from it (cf. `RFC6201
1516 <https://www.rfc-editor.org/rfc/rfc6201.txt>`__).
1518 `RFC6201 <https://www.rfc-editor.org/rfc/rfc6201.txt>`__ defines two methods
1519 to measure the Reset time:
1521 - Frame-Loss Method: which requires the monitoring of the number of
1522 lost frames and calculates the Reset time based on the number of
1523 frames lost and the offered rate according to the following
1526 .. code-block:: console
1528 Frames_lost (packets)
1529 Reset_time = -------------------------------------
1530 Offered_rate (packets per second)
1532 - Timestamp Method: which measures the time from which the last frame
1533 is forwarded from the DUT to the time the first frame is forwarded
1534 after the reset. This involves time-stamping all transmitted frames
1535 and recording the timestamp of the last frame that was received prior
1536 to the reset and also measuring the timestamp of the first frame that
1537 is received after the reset. The Reset time is the difference between
1538 these two timestamps.
1540 According to `RFC6201 <https://www.rfc-editor.org/rfc/rfc6201.txt>`__ the
1541 choice of method depends on the test tool's capability; the Frame-Loss
1542 method SHOULD be used if the test tool supports:
1544 * Counting the number of lost frames per stream.
1545 * Transmitting test frame despite the physical link status.
1547 whereas the Timestamp method SHOULD be used if the test tool supports:
1548 * Timestamping each frame.
1549 * Monitoring received frame's timestamp.
1550 * Transmitting frames only if the physical link status is up.
1552 **Expected Result**:
1554 **Metrics collected**
1556 The following are the metrics collected for this test:
1558 * Average Reset Time over the number of trials performed.
1560 Results of this test should include the following information:
1562 * The reset method used.
1563 * Throughput in Fps and Mbps.
1564 * Average Frame Loss over the number of trials performed.
1565 * Average Reset Time in milliseconds over the number of trials performed.
1566 * Number of trials performed.
1567 * Protocol: IPv4, IPv6, MPLS, etc.
1568 * Frame Size in Octets
1569 * Port Media: Ethernet, Gigabit Ethernet (GbE), etc.
1570 * Port Speed: 10 Gbps, 40 Gbps etc.
1571 * Interface Encapsulation: Ethernet, Ethernet VLAN, etc.
1573 **Deployment scenario**:
1575 * Physical → virtual switch → physical.
1579 Test ID: LTD.Throughput.RFC2889.MaxForwardingRate
1580 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1581 **Title**: RFC2889 Forwarding Rate Test
1583 **Prerequisite Test**: LTD.Throughput.RFC2544.PacketLossRatio
1589 This test measures the DUT's Max Forwarding Rate when the Offered Load
1590 is varied between the throughput and the Maximum Offered Load for fixed
1591 length frames at a fixed time interval. The selected frame sizes are
1592 those previously defined under `Default Test
1593 Parameters <#DefaultParams>`__. The throughput is the maximum offered
1594 load with 0% frame loss (measured by the prerequisite test), and the
1595 Maximum Offered Load (as defined by
1596 `RFC2285 <https://www.rfc-editor.org/rfc/rfc2285.txt>`__) is *"the highest
1597 number of frames per second that an external source can transmit to a
1598 DUT/SUT for forwarding to a specified output interface or interfaces"*.
1600 Traffic should be sent to the DUT at a particular rate (TX rate)
1601 starting with TX rate equal to the throughput rate. The rate of
1602 successfully received frames at the destination counted (in FPS). If the
1603 RX rate is equal to the TX rate, the TX rate should be increased by a
1604 fixed step size and the RX rate measured again until the Max Forwarding
1607 The trial duration for each iteration should last for the period of time
1608 needed for the system to reach steady state for the frame size being
1609 tested. Under `RFC2889 <https://www.rfc-editor.org/rfc/rfc2289.txt>`__
1610 (Sec. 5.6.3.1) test methodology, the test
1611 duration should run for a minimum period of 30 seconds, regardless
1612 whether the system reaches steady state before the minimum duration
1615 **Expected Result**: According to
1616 `RFC2889 <https://www.rfc-editor.org/rfc/rfc2289.txt>`__ The Max Forwarding
1617 Rate is the highest forwarding rate of a DUT taken from an iterative set of
1618 forwarding rate measurements. The iterative set of forwarding rate measurements
1619 are made by setting the intended load transmitted from an external source and
1620 measuring the offered load (i.e what the DUT is capable of forwarding). If the
1621 Throughput == the Maximum Offered Load, it follows that Max Forwarding Rate is
1622 equal to the Maximum Offered Load.
1624 **Metrics Collected**:
1626 The following are the metrics collected for this test:
1628 - The Max Forwarding Rate for the DUT for each packet size.
1629 - CPU and memory utilization may also be collected as part of this
1630 test, to determine the vSwitch's performance footprint on the system.
1632 **Deployment scenario**:
1634 - Physical → virtual switch → physical. Note: Full mesh tests with
1635 multiple ingress and egress ports are a key aspect of RFC 2889
1636 benchmarks, and scenarios with both 2 and 4 ports should be tested.
1637 In any case, the number of ports used must be reported.
1641 Test ID: LTD.Throughput.RFC2889.ForwardPressure
1642 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1643 **Title**: RFC2889 Forward Pressure Test
1645 **Prerequisite Test**: LTD.Throughput.RFC2889.MaxForwardingRate
1651 The aim of this test is to determine if the DUT transmits frames with an
1652 inter-frame gap that is less than 12 bytes. This test overloads the DUT
1653 and measures the output for forward pressure. Traffic should be
1654 transmitted to the DUT with an inter-frame gap of 11 bytes, this will
1655 overload the DUT by 1 byte per frame. The forwarding rate of the DUT
1658 **Expected Result**: The forwarding rate should not exceed the maximum
1659 forwarding rate of the DUT collected by
1660 LTD.Throughput.RFC2889.MaxForwardingRate.
1662 **Metrics collected**
1664 The following are the metrics collected for this test:
1666 - Forwarding rate of the DUT in FPS or Mbps.
1667 - CPU and memory utilization may also be collected as part of this
1668 test, to determine the vSwitch's performance footprint on the system.
1670 **Deployment scenario**:
1672 - Physical → virtual switch → physical.
1676 Test ID: LTD.Throughput.RFC2889.ErrorFramesFiltering
1677 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1678 **Title**: RFC2889 Error Frames Filtering Test
1680 **Prerequisite Test**: N/A
1686 The aim of this test is to determine whether the DUT will propagate any
1687 erroneous frames it receives or whether it is capable of filtering out
1688 the erroneous frames. Traffic should be sent with erroneous frames
1689 included within the flow at random intervals. Illegal frames that must
1690 be tested include: - Oversize Frames. - Undersize Frames. - CRC Errored
1691 Frames. - Dribble Bit Errored Frames - Alignment Errored Frames
1693 The traffic flow exiting the DUT should be recorded and checked to
1694 determine if the erroneous frames where passed through the DUT.
1696 **Expected Result**: Broken frames are not passed!
1698 **Metrics collected**
1700 No Metrics are collected in this test, instead it determines:
1702 - Whether the DUT will propagate erroneous frames.
1703 - Or whether the DUT will correctly filter out any erroneous frames
1704 from traffic flow with out removing correct frames.
1706 **Deployment scenario**:
1708 - Physical → virtual switch → physical.
1712 Test ID: LTD.Throughput.RFC2889.BroadcastFrameForwarding
1713 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1714 **Title**: RFC2889 Broadcast Frame Forwarding Test
1716 **Prerequisite Test**: N
1722 The aim of this test is to determine the maximum forwarding rate of the
1723 DUT when forwarding broadcast traffic. For each frame previously defined
1724 under `Default Test Parameters <#DefaultParams>`__, the traffic should
1725 be set up as broadcast traffic. The traffic throughput of the DUT should
1728 The test should be conducted with at least 4 physical ports on the DUT.
1729 The number of ports used MUST be recorded.
1731 As broadcast involves forwarding a single incoming packet to several
1732 destinations, the latency of a single packet is defined as the average
1733 of the latencies for each of the broadcast destinations.
1735 The incoming packet is transmitted on each of the other physical ports,
1736 it is not transmitted on the port on which it was received. The test MAY
1737 be conducted using different broadcasting ports to uncover any
1738 performance differences.
1740 **Expected Result**:
1742 **Metrics collected**:
1744 The following are the metrics collected for this test:
1746 - The forwarding rate of the DUT when forwarding broadcast traffic.
1747 - The minimum, average & maximum packets latencies observed.
1749 **Deployment scenario**:
1751 - Physical → virtual switch 3x physical. In the Broadcast rate testing,
1752 four test ports are required. One of the ports is connected to the test
1753 device, so it can send broadcast frames and listen for miss-routed frames.
1757 Packet Latency tests
1758 ---------------------------
1759 These tests will measure the store and forward latency as well as the packet
1760 delay variation for various packet types through the virtual switch. The
1761 following list is not exhaustive but should indicate the type of tests
1762 that should be required. It is expected that more will be added.
1766 Test ID: LTD.PacketLatency.InitialPacketProcessingLatency
1767 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1768 **Title**: Initial Packet Processing Latency
1770 **Prerequisite Test**: N/A
1776 In some virtual switch architectures, the first packets of a flow will
1777 take the system longer to process than subsequent packets in the flow.
1778 This test determines the latency for these packets. The test will
1779 measure the latency of the packets as they are processed by the
1780 flow-setup-path of the DUT. There are two methods for this test, a
1781 recommended method and a nalternative method that can be used if it is
1782 possible to disable the fastpath of the virtual switch.
1784 Recommended method: This test will send 64,000 packets to the DUT, each
1785 belonging to a different flow. Average packet latency will be determined
1786 over the 64,000 packets.
1788 Alternative method: This test will send a single packet to the DUT after
1789 a fixed interval of time. The time interval will be equivalent to the
1790 amount of time it takes for a flow to time out in the virtual switch
1791 plus 10%. Average packet latency will be determined over 1,000,000
1794 This test is intended only for non-learning virtual switches; For learning
1795 virtual switches use RFC2889.
1797 For this test, only unidirectional traffic is required.
1799 **Expected Result**: The average latency for the initial packet of all
1800 flows should be greater than the latency of subsequent traffic.
1802 **Metrics Collected**:
1804 The following are the metrics collected for this test:
1806 - Average latency of the initial packets of all flows that are
1807 processed by the DUT.
1809 **Deployment scenario**:
1811 - Physical → Virtual Switch → Physical.
1815 Test ID: LTD.PacketDelayVariation.RFC3393.Soak
1816 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1817 **Title**: Packet Delay Variation Soak Test
1819 **Prerequisite Tests**: LTD.Throughput.RFC2544.PacketLossRatio (0% Packet Loss)
1825 The aim of this test is to understand the distribution of packet delay
1826 variation for different frame sizes over an extended test duration and
1827 to determine if there are any outliers. To allow for an extended test
1828 duration, the test should ideally run for 24 hours or, if this is not
1829 possible, for at least 6 hour. For this test, each frame size must be
1830 sent at the highest possible throughput with 0% packet loss, as
1831 determined in the prerequisite test.
1833 **Expected Result**:
1835 **Metrics Collected**:
1837 The following are the metrics collected for this test:
1839 - The packet delay variation value for traffic passing through the DUT.
1840 - The `RFC5481 <https://www.rfc-editor.org/rfc/rfc5481.txt>`__
1841 PDV form of delay variation on the traffic flow,
1842 using the 99th percentile, for each 60s interval during the test.
1843 - CPU and memory utilization may also be collected as part of this
1844 test, to determine the vSwitch's performance footprint on the system.
1849 ------------------------
1850 The general aim of these tests is to understand the impact of large flow
1851 table size and flow lookups on throughput. The following list is not
1852 exhaustive but should indicate the type of tests that should be required.
1853 It is expected that more will be added.
1857 Test ID: LTD.Scalability.RFC2544.0PacketLoss
1858 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1859 **Title**: RFC 2544 0% loss Scalability throughput test
1861 **Prerequisite Test**: LTD.Throughput.RFC2544.PacketLossRatio, IF the
1862 delta Throughput between the single-flow RFC2544 test and this test with
1863 a variable number of flows is desired.
1869 The aim of this test is to measure how throughput changes as the number
1870 of flows in the DUT increases. The test will measure the throughput
1871 through the fastpath, as such the flows need to be installed on the DUT
1872 before passing traffic.
1874 For each frame size previously defined under `Default Test
1875 Parameters <#DefaultParams>`__ and for each of the following number of
1885 - Max supported number of flows.
1887 This test will be conducted under two conditions following the
1888 establishment of all flows as required by RFC 2544, regarding the flow
1889 expiration time-out:
1891 1) The time-out never expires during each trial.
1893 2) The time-out expires for all flows periodically. This would require a
1894 short time-out compared with flow re-appearance for a small number of
1895 flows, and may not be possible for all flow conditions.
1897 The maximum 0% packet loss Throughput should be determined in a manner
1898 identical to LTD.Throughput.RFC2544.PacketLossRatio.
1900 **Expected Result**:
1902 **Metrics Collected**:
1904 The following are the metrics collected for this test:
1906 - The maximum number of frames per second that can be forwarded at the
1907 specified number of flows and the specified frame size, with zero
1912 Test ID: LTD.MemoryBandwidth.RFC2544.0PacketLoss.Scalability
1913 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1914 **Title**: RFC 2544 0% loss Memory Bandwidth Scalability test
1916 **Prerequisite Tests**: LTD.Throughput.RFC2544.PacketLossRatio, IF the
1917 delta Throughput between an undisturbed RFC2544 test and this test with
1918 the Throughput affected by cache and memory bandwidth contention is desired.
1924 The aim of this test is to understand how the DUT's performance is
1925 affected by cache sharing and memory bandwidth between processes.
1927 During the test all cores not used by the vSwitch should be running a
1928 memory intensive application. This application should read and write
1929 random data to random addresses in unused physical memory. The random
1930 nature of the data and addresses is intended to consume cache, exercise
1931 main memory access (as opposed to cache) and exercise all memory buses
1932 equally. Furthermore:
1934 - the ratio of reads to writes should be recorded. A ratio of 1:1
1936 - the reads and writes MUST be of cache-line size and be cache-line aligned.
1937 - in NUMA architectures memory access SHOULD be local to the core's node.
1938 Whether only local memory or a mix of local and remote memory is used
1940 - the memory bandwidth (reads plus writes) used per-core MUST be recorded;
1941 the test MUST be run with a per-core memory bandwidth equal to half the
1942 maximum system memory bandwidth divided by the number of cores. The test
1943 MAY be run with other values for the per-core memory bandwidth.
1944 - the test MAY also be run with the memory intensive application running
1947 Under these conditions the DUT's 0% packet loss throughput is determined
1948 as per LTD.Throughput.RFC2544.PacketLossRatio.
1950 **Expected Result**:
1952 **Metrics Collected**:
1954 The following are the metrics collected for this test:
1956 - The DUT's 0% packet loss throughput in the presence of cache sharing and
1957 memory bandwidth between processes.
1962 -----------------------
1963 The general aim of these tests is to understand the capacity of the
1964 and speed with which the vswitch can accommodate new flows.
1968 Test ID: LTD.Activation.RFC2889.AddressCachingCapacity
1969 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1970 **Title**: RFC2889 Address Caching Capacity Test
1972 **Prerequisite Test**: N/A
1978 Please note this test is only applicable to virtual switches that are capable of
1979 MAC learning. The aim of this test is to determine the address caching
1980 capacity of the DUT for a constant load (fixed length frames at a fixed
1981 interval time). The selected frame sizes are those previously defined
1982 under `Default Test Parameters <#DefaultParams>`__.
1984 In order to run this test the aging time, that is the maximum time the
1985 DUT will keep a learned address in its flow table, and a set of initial
1986 addresses, whose value should be >= 1 and <= the max number supported by
1987 the implementation must be known. Please note that if the aging time is
1988 configurable it must be longer than the time necessary to produce frames
1989 from the external source at the specified rate. If the aging time is
1990 fixed the frame rate must be brought down to a value that the external
1991 source can produce in a time that is less than the aging time.
1993 Learning Frames should be sent from an external source to the DUT to
1994 install a number of flows. The Learning Frames must have a fixed
1995 destination address and must vary the source address of the frames. The
1996 DUT should install flows in its flow table based on the varying source
1997 addresses. Frames should then be transmitted from an external source at
1998 a suitable frame rate to see if the DUT has properly learned all of the
1999 addresses. If there is no frame loss and no flooding, the number of
2000 addresses sent to the DUT should be increased and the test is repeated
2001 until the max number of cached addresses supported by the DUT
2004 **Expected Result**:
2006 **Metrics collected**:
2008 The following are the metrics collected for this test:
2010 - Number of cached addresses supported by the DUT.
2011 - CPU and memory utilization may also be collected as part of this
2012 test, to determine the vSwitch's performance footprint on the system.
2014 **Deployment scenario**:
2016 - Physical → virtual switch → 2 x physical (one receiving, one listening).
2020 Test ID: LTD.Activation.RFC2889.AddressLearningRate
2021 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2022 **Title**: RFC2889 Address Learning Rate Test
2024 **Prerequisite Test**: LTD.Memory.RFC2889.AddressCachingCapacity
2030 Please note this test is only applicable to virtual switches that are capable of
2031 MAC learning. The aim of this test is to determine the rate of address
2032 learning of the DUT for a constant load (fixed length frames at a fixed
2033 interval time). The selected frame sizes are those previously defined
2034 under `Default Test Parameters <#DefaultParams>`__, traffic should be
2035 sent with each IPv4/IPv6 address incremented by one. The rate at which
2036 the DUT learns a new address should be measured. The maximum caching
2037 capacity from LTD.Memory.RFC2889.AddressCachingCapacity should be taken
2038 into consideration as the maximum number of addresses for which the
2039 learning rate can be obtained.
2041 **Expected Result**: It may be worthwhile to report the behaviour when
2042 operating beyond address capacity - some DUTs may be more friendly to
2043 new addresses than others.
2045 **Metrics collected**:
2047 The following are the metrics collected for this test:
2049 - The address learning rate of the DUT.
2051 **Deployment scenario**:
2053 - Physical → virtual switch → 2 x physical (one receiving, one listening).
2057 Coupling between control path and datapath Tests
2058 -------------------------------------------------------
2059 The following tests aim to determine how tightly coupled the datapath
2060 and the control path are within a virtual switch. The following list
2061 is not exhaustive but should indicate the type of tests that should be
2062 required. It is expected that more will be added.
2066 Test ID: LTD.CPDPCouplingFlowAddition
2067 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2068 **Title**: Control Path and Datapath Coupling
2070 **Prerequisite Test**:
2076 The aim of this test is to understand how exercising the DUT's control
2077 path affects datapath performance.
2079 Initially a certain number of flow table entries are installed in the
2080 vSwitch. Then over the duration of an RFC2544 throughput test
2081 flow-entries are added and removed at the rates specified below. No
2082 traffic is 'hitting' these flow-entries, they are simply added and
2085 The test MUST be repeated with the following initial number of
2086 flow-entries installed: - < 10 - 1000 - 100,000 - 10,000,000 (or the
2087 maximum supported number of flow-entries)
2089 The test MUST be repeated with the following rates of flow-entry
2090 addition and deletion per second: - 0 - 1 (i.e. 1 addition plus 1
2091 deletion) - 100 - 10,000
2093 **Expected Result**:
2095 **Metrics Collected**:
2097 The following are the metrics collected for this test:
2099 - The maximum forwarding rate in Frames Per Second (FPS) and Mbps of
2101 - The average latency of the traffic flow when passing through the DUT
2102 (if testing for latency, note that this average is different from the
2103 test specified in Section 26.3 of
2104 `RFC2544 <https://www.rfc-editor.org/rfc/rfc2544.txt>`__).
2105 - CPU and memory utilization may also be collected as part of this
2106 test, to determine the vSwitch's performance footprint on the system.
2108 **Deployment scenario**:
2110 - Physical → virtual switch → physical.
2114 CPU and memory consumption
2115 ---------------------------------
2116 The following tests will profile a virtual switch's CPU and memory
2117 utilization under various loads and circumstances. The following
2118 list is not exhaustive but should indicate the type of tests that
2119 should be required. It is expected that more will be added.
2123 Test ID: LTD.CPU.RFC2544.0PacketLoss
2124 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2125 **Title**: RFC 2544 0% Loss Compute Test
2127 **Prerequisite Test**:
2133 The aim of this test is to understand the overall performance of the
2134 system when a CPU intensive application is run on the same DUT as the
2135 Virtual Switch. For each frame size, an
2136 LTD.Throughput.RFC2544.PacketLossRatio (0% Packet Loss) test should be
2137 performed. Throughout the entire test a CPU intensive application should
2138 be run on all cores on the system not in use by the Virtual Switch. For
2139 NUMA system only cores on the same NUMA node are loaded.
2141 It is recommended that stress-ng be used for loading the non-Virtual
2142 Switch cores but any stress tool MAY be used.
2144 **Expected Result**:
2146 **Metrics Collected**:
2148 The following are the metrics collected for this test:
2150 - CPU utilization of the cores running the Virtual Switch.
2151 - The number of identity of the cores allocated to the Virtual Switch.
2152 - The configuration of the stress tool (for example the command line
2153 parameters used to start it.)
2157 Summary List of Tests
2158 ----------------------------
2161 - Test ID: LTD.Throughput.RFC2544.PacketLossRatio
2162 - Test ID: LTD.Throughput.RFC2544.PacketLossRatioFrameModification
2163 - Test ID: LTD.Throughput.RFC2544.Profile
2164 - Test ID: LTD.Throughput.RFC2544.SystemRecoveryTime
2165 - Test ID: LTD.Throughput.RFC2544.BackToBackFrames
2166 - Test ID: LTD.Throughput.RFC2889.Soak
2167 - Test ID: LTD.Throughput.RFC2889.SoakFrameModification
2168 - Test ID: LTD.Throughput.RFC6201.ResetTime
2169 - Test ID: LTD.Throughput.RFC2889.MaxForwardingRate
2170 - Test ID: LTD.Throughput.RFC2889.ForwardPressure
2171 - Test ID: LTD.Throughput.RFC2889.ErrorFramesFiltering
2172 - Test ID: LTD.Throughput.RFC2889.BroadcastFrameForwarding
2174 2. Packet Latency tests
2176 - Test ID: LTD.PacketLatency.InitialPacketProcessingLatency
2177 - Test ID: LTD.PacketDelayVariation.RFC3393.Soak
2179 3. Scalability tests
2181 - Test ID: LTD.Scalability.RFC2544.0PacketLoss
2182 - Test ID: LTD.MemoryBandwidth.RFC2544.0PacketLoss.Scalability
2186 - Test ID: LTD.Activation.RFC2889.AddressCachingCapacity
2187 - Test ID: LTD.Activation.RFC2889.AddressLearningRate
2189 5. Coupling between control path and datapath Tests
2191 - Test ID: LTD.CPDPCouplingFlowAddition
2193 6. CPU and memory consumption
2195 - Test ID: LTD.CPU.RFC2544.0PacketLoss
Code Review / vswitchperf.git / blob