1 <?xml version="1.0" encoding="US-ASCII"?>
2 <!DOCTYPE rfc SYSTEM "rfc2629.dtd">
4 <?rfc tocompact="yes"?>
6 <?rfc tocindent="yes"?>
12 <?rfc subcompact="no"?>
13 <rfc category="info" docName="draft-vsperf-bmwg-vswitch-opnfv-01"
16 <title abbrev="Benchmarking vSwitches">Benchmarking Virtual Switches in
19 <author fullname="Maryam Tahhan" initials="M." surname="Tahhan">
20 <organization>Intel</organization>
39 <email>maryam.tahhan@intel.com</email>
45 <author fullname="Billy O'Mahony" initials="B." surname="O'Mahony">
46 <organization>Intel</organization>
65 <email>billy.o.mahony@intel.com</email>
71 <author fullname="Al Morton" initials="A." surname="Morton">
72 <organization>AT&T Labs</organization>
76 <street>200 Laurel Avenue South</street>
78 <city>Middletown,</city>
84 <country>USA</country>
87 <phone>+1 732 420 1571</phone>
89 <facsimile>+1 732 368 1192</facsimile>
91 <email>acmorton@att.com</email>
93 <uri>http://home.comcast.net/~acmacm/</uri>
97 <date day="14" month="October" year="2015"/>
100 <t>This memo describes the progress of the Open Platform for NFV (OPNFV)
101 project on virtual switch performance "VSWITCHPERF". This project
102 intends to build on the current and completed work of the Benchmarking
103 Methodology Working Group in IETF, by referencing existing literature.
104 The Benchmarking Methodology Working Group has traditionally conducted
105 laboratory characterization of dedicated physical implementations of
106 internetworking functions. Therefore, this memo begins to describe the
107 additional considerations when virtual switches are implemented in
108 general-purpose hardware. The expanded tests and benchmarks are also
109 influenced by the OPNFV mission to support virtualization of the "telco"
113 <note title="Requirements Language">
114 <t>The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
115 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
116 document are to be interpreted as described in <xref
117 target="RFC2119">RFC 2119</xref>.</t>
124 <section title="Introduction">
125 <t>Benchmarking Methodology Working Group (BMWG) has traditionally
126 conducted laboratory characterization of dedicated physical
127 implementations of internetworking functions. The Black-box Benchmarks
128 of Throughput, Latency, Forwarding Rates and others have served our
129 industry for many years. Now, Network Function Virtualization (NFV) has
130 the goal to transform how internetwork functions are implemented, and
131 therefore has garnered much attention.</t>
133 <t>This memo describes the progress of the Open Platform for NFV (OPNFV)
134 project on virtual switch performance characterization, "VSWITCHPERF".
135 This project intends to build on the current and completed work of the
136 Benchmarking Methodology Working Group in IETF, by referencing existing
137 literature. For example, currently the most often referenced RFC is
138 <xref target="RFC2544"/> (which depends on <xref target="RFC1242"/>) and
139 foundation of the benchmarking work in OPNFV is common and strong.</t>
142 https://wiki.opnfv.org/characterize_vswitch_performance_for_telco_nfv_use_cases
143 for more background, and the OPNFV website for general information:
144 https://www.opnfv.org/</t>
146 <t>The authors note that OPNFV distinguishes itself from other open
147 source compute and networking projects through its emphasis on existing
148 "telco" services as opposed to cloud-computing. There are many ways in
149 which telco requirements have different emphasis on performance
150 dimensions when compared to cloud computing: support for and transfer of
151 isochronous media streams is one example.</t>
153 <t>Note also that the move to NFV Infrastructure has resulted in many
154 new benchmarking initiatives across the industry, and the authors are
155 currently doing their best to maintain alignment with many other
156 projects, and this Internet Draft is evidence of the efforts.</t>
159 <section title="Scope">
160 <t>The primary purpose and scope of the memo is to inform BMWG of
161 work-in-progress that builds on the body of extensive literature and
162 experience. Additionally, once the initial information conveyed here is
163 received, this memo may be expanded to include more detail and
164 commentary from both BMWG and OPNFV communities, under BMWG's chartered
165 work to characterize the NFV Infrastructure (a virtual switch is an
166 important aspect of that infrastructure).</t>
169 <section title="Benchmarking Considerations">
170 <t>This section highlights some specific considerations (from <xref
171 target="I-D.ietf-bmwg-virtual-net"/>)related to Benchmarks for virtual
172 switches. The OPNFV project is sharing its present view on these areas,
173 as they develop their specifications in the Level Test Design (LTD)
176 <section title="Comparison with Physical Network Functions">
177 <t>To compare the performance of virtual designs and implementations
178 with their physical counterparts, identical benchmarks are needed.
179 BMWG has developed specifications for many network functions this memo
180 re-uses existing benchmarks through references, and expands them
181 during development of new methods. A key configuration aspect is the
182 number of parallel cores required to achieve comparable performance
183 with a given physical device, or whether some limit of scale was
184 reached before the cores could achieve the comparable level.</t>
186 <t>It's unlikely that the virtual switch will be the only application
187 running on the SUT, so CPU utilization, Cache utilization, and Memory
188 footprint should also be recorded for the virtual implementations of
189 internetworking functions.</t>
192 <section title="Continued Emphasis on Black-Box Benchmarks">
193 <t>External observations remain essential as the basis for Benchmarks.
194 Internal observations with fixed specification and interpretation will
195 be provided in parallel to assist the development of operations
196 procedures when the technology is deployed.</t>
199 <section title="New Configuration Parameters">
200 <t>A key consideration when conducting any sort of benchmark is trying
201 to ensure the consistency and repeatability of test results. When
202 benchmarking the performance of a vSwitch there are many factors that
203 can affect the consistency of results, one key factor is matching the
204 various hardware and software details of the SUT. This section lists
205 some of the many new parameters which this project believes are
206 critical to report in order to achieve repeatability.</t>
208 <t>Hardware details including:</t>
210 <t><list style="symbols">
211 <t>Platform details</t>
213 <t>Processor details</t>
215 <t>Memory information (type and size)</t>
217 <t>Number of enabled cores</t>
219 <t>Number of cores used for the test</t>
221 <t>Number of physical NICs, as well as their details
222 (manufacturer, versions, type and the PCI slot they are plugged
225 <t>NIC interrupt configuration</t>
227 <t>BIOS version, release date and any configurations that were
230 <t>CPU microcode level</t>
232 <t>Memory DIMM configurations (quad rank performance may not be
233 the same as dual rank) in size, freq and slot locations</t>
235 <t>PCI configuration parameters (payload size, early ack
238 <t>Power management at all levels (ACPI sleep states, processor
240 </list>Software details including:</t>
242 <t><list style="symbols">
243 <t>OS parameters and behavior (text vs graphical no one typing at
244 the console on one system)</t>
246 <t>OS version (for host and VNF)</t>
248 <t>Kernel version (for host and VNF)</t>
250 <t>GRUB boot parameters (for host and VNF)</t>
252 <t>Hypervisor details (Type and version)</t>
254 <t>Selected vSwitch, version number or commit id used</t>
256 <t>vSwitch launch command line if it has been parameterised</t>
258 <t>Memory allocation to the vSwitch</t>
260 <t>which NUMA node it is using, and how many memory channels</t>
262 <t>DPDK or any other SW dependency version number or commit id
265 <t>Memory allocation to a VM - if it's from Hugpages/elsewhere</t>
267 <t>VM storage type: snapshot/independent persistent/independent
272 <t>Number of Virtual NICs (vNICs), versions, type and driver</t>
274 <t>Number of virtual CPUs and their core affinity on the host</t>
276 <t>Number vNIC interrupt configuration</t>
278 <t>Thread affinitization for the applications (including the
279 vSwitch itself) on the host</t>
281 <t>Details of Resource isolation, such as CPUs designated for
282 Host/Kernel (isolcpu) and CPUs designated for specific processes
283 (taskset). - Test duration. - Number of flows.</t>
286 <t>Test Traffic Information:<list style="symbols">
287 <t>Traffic type - UDP, TCP, IMIX / Other</t>
291 <t>Deployment Scenario</t>
297 <section title="Flow classification">
298 <t>Virtual switches group packets into flows by processing and
299 matching particular packet or frame header information, or by matching
300 packets based on the input ports. Thus a flow can be thought of a
301 sequence of packets that have the same set of header field values or
302 have arrived on the same port. Performance results can vary based on
303 the parameters the vSwitch uses to match for a flow. The recommended
304 flow classification parameters for any vSwitch performance tests are:
305 the input port, the source IP address, the destination IP address and
306 the Ethernet protocol type field. It is essential to increase the flow
307 timeout time on a vSwitch before conducting any performance tests that
308 do not measure the flow setup time. Normally the first packet of a
309 particular stream will install the flow in the virtual switch which
310 adds an additional latency, subsequent packets of the same flow are
311 not subject to this latency if the flow is already installed on the
315 <section title="Benchmarks using Baselines with Resource Isolation">
316 <t>This outline describes measurement of baseline with isolated
317 resources at a high level, which is the intended approach at this
320 <t><list style="numbers">
321 <t>Baselines: <list style="symbols">
322 <t>Optional: Benchmark platform forwarding capability without
323 a vswitch or VNF for at least 72 hours (serves as a means of
324 platform validation and a means to obtain the base performance
325 for the platform in terms of its maximum forwarding rate and
327 <preamble>Benchmark platform forwarding
328 capability</preamble>
330 <artwork align="right"><![CDATA[ __
331 +--------------------------------------------------+ |
332 | +------------------------------------------+ | |
334 | | Simple Forwarding App | | Host
336 | +------------------------------------------+ | |
338 +---+------------------------------------------+---+ __|
342 +--------------------------------------------------+
344 | traffic generator |
346 +--------------------------------------------------+]]></artwork>
351 <t>Benchmark VNF forwarding capability with direct
352 connectivity (vSwitch bypass, e.g., SR/IOV) for at least 72
353 hours (serves as a means of VNF validation and a means to
354 obtain the base performance for the VNF in terms of its
355 maximum forwarding rate and latency). The metrics gathered
356 from this test will serve as a key comparison point for
357 vSwitch bypass technologies performance and vSwitch
358 performance. <figure align="right">
359 <preamble>Benchmark VNF forwarding capability</preamble>
361 <artwork><![CDATA[ __
362 +--------------------------------------------------+ |
363 | +------------------------------------------+ | |
367 | +------------------------------------------+ | |
368 | | Passthrough/SR-IOV | | Host
369 | +------------------------------------------+ | |
371 +---+------------------------------------------+---+ __|
375 +--------------------------------------------------+
377 | traffic generator |
379 +--------------------------------------------------+]]></artwork>
384 <t>Benchmarking with isolated resources alone, with other
385 resources (both HW&SW) disabled Example, vSw and VM are
388 <t>Benchmarking with isolated resources alone, leaving some
391 <t>Benchmark with isolated resources and all resources
395 <t>Next Steps<list style="symbols">
396 <t>Limited sharing</t>
398 <t>Production scenarios</t>
400 <t>Stressful scenarios</t>
406 <section title="VSWITCHPERF Specification Summary">
407 <t>The overall specification in preparation is referred to as a Level
408 Test Design (LTD) document, which will contain a suite of performance
409 tests. The base performance tests in the LTD are based on the
410 pre-existing specifications developed by BMWG to test the performance of
411 physical switches. These specifications include:</t>
413 <t><list style="symbols">
414 <t><xref target="RFC2544"/> Benchmarking Methodology for Network
415 Interconnect Devices</t>
417 <t><xref target="RFC2889"/> Benchmarking Methodology for LAN
420 <t><xref target="RFC6201"/> Device Reset Characterization</t>
422 <t><xref target="RFC5481"/> Packet Delay Variation Applicability
426 <t>Some of the above/newer RFCs are being applied in benchmarking for
427 the first time, and represent a development challenge for test equipment
428 developers. Fortunately, many members of the testing system community
429 have engaged on the VSPERF project, including an open source test
432 <t>In addition to this, the LTD also re-uses the terminology defined
435 <t><list style="symbols">
436 <t><xref target="RFC2285"/> Benchmarking Terminology for LAN
437 Switching Devices</t>
439 <t><xref target="RFC5481"/> Packet Delay Variation Applicability
445 <t>Specifications to be included in future updates of the LTD
446 include:<list style="symbols">
447 <t><xref target="RFC3918"/> Methodology for IP Multicast
450 <t><xref target="RFC4737"/> Packet Reordering Metrics</t>
453 <t>As one might expect, the most fundamental internetworking
454 characteristics of Throughput and Latency remain important when the
455 switch is virtualized, and these benchmarks figure prominently in the
458 <t>When considering characteristics important to "telco" network
459 functions, we must begin to consider additional performance metrics. In
460 this case, the project specifications have referenced metrics from the
461 IETF IP Performance Metrics (IPPM) literature. This means that the <xref
462 target="RFC2544"/> test of Latency is replaced by measurement of a
463 metric derived from IPPM's <xref target="RFC2679"/>, where a set of
464 statistical summaries will be provided (mean, max, min, etc.). Further
465 metrics planned to be benchmarked include packet delay variation as
466 defined by <xref target="RFC5481"/> , reordering, burst behaviour, DUT
467 availability, DUT capacity and packet loss in long term testing at
468 Throughput level, where some low-level of background loss may be present
469 and characterized.</t>
471 <t>Tests have been (or will be) designed to collect the metrics
474 <t><list style="symbols">
475 <t>Throughput Tests to measure the maximum forwarding rate (in
476 frames per second or fps) and bit rate (in Mbps) for a constant load
477 (as defined by RFC1242) without traffic loss.</t>
479 <t>Packet and Frame Delay Distribution Tests to measure average, min
480 and max packet and frame delay for constant loads.</t>
482 <t>Packet Delay Tests to understand latency distribution for
483 different packet sizes and over an extended test run to uncover
486 <t>Scalability Tests to understand how the virtual switch performs
487 as the number of flows, active ports, complexity of the forwarding
488 logic’s configuration… it has to deal with
491 <t>Stream Performance Tests (TCP, UDP) to measure bulk data transfer
492 performance, i.e. how fast systems can send and receive data through
495 <t>Control Path and Datapath Coupling Tests, to understand how
496 closely coupled the datapath and the control path are as well as the
497 effect of this coupling on the performance of the DUT (example:
498 delay of the initial packet of a flow).</t>
500 <t>CPU and Memory Consumption Tests to understand the virtual
501 switch’s footprint on the system, usually conducted as
502 auxiliary measurements with benchmarks above. They include: CPU
503 utilization, Cache utilization and Memory footprint.</t>
506 <t>Future/planned test specs include:<list style="symbols">
507 <t>Request/Response Performance Tests (TCP, UDP) which measure the
508 transaction rate through the switch.</t>
510 <t>Noisy Neighbour Tests, to understand the effects of resource
511 sharing on the performance of a virtual switch.</t>
513 <t>Tests derived from examination of ETSI NFV Draft GS IFA003
514 requirements <xref target="IFA003"/> on characterization of
515 acceleration technologies applied to vswitches.</t>
516 </list>The flexibility of deployment of a virtual switch within a
517 network means that the BMWG IETF existing literature needs to be used to
518 characterize the performance of a switch in various deployment
519 scenarios. The deployment scenarios under consideration include:</t>
522 <preamble>Physical port to virtual switch to physical
525 <artwork><![CDATA[ __
526 +--------------------------------------------------+ |
527 | +--------------------+ | |
530 | +--------------+ +--------------+ | |
531 | | phy port | vSwitch | phy port | | |
532 +---+--------------+------------+--------------+---+ __|
536 +--------------------------------------------------+
538 | traffic generator |
540 +--------------------------------------------------+]]></artwork>
544 <preamble>Physical port to virtual switch to VNF to virtual switch
545 to physical port</preamble>
547 <artwork><![CDATA[ __
548 +---------------------------------------------------+ |
550 | +-------------------------------------------+ | |
551 | | Application | | |
552 | +-------------------------------------------+ | |
556 | +---------------+ +---------------+ | |
557 | | logical port 0| | logical port 1| | |
558 +---+---------------+-----------+---------------+---+ __|
562 +---+---------------+----------+---------------+---+ |
563 | | logical port 0| | logical port 1| | |
564 | +---------------+ +---------------+ | |
568 | +--------------+ +--------------+ | |
569 | | phy port | vSwitch | phy port | | |
570 +---+--------------+------------+--------------+---+ __|
574 +--------------------------------------------------+
576 | traffic generator |
578 +--------------------------------------------------+]]></artwork>
580 <preamble>Physical port to virtual switch to VNF to virtual switch
581 to VNF to virtual switch to physical port</preamble>
583 <artwork><![CDATA[ __
584 +----------------------+ +----------------------+ |
585 | Guest 1 | | Guest 2 | |
586 | +---------------+ | | +---------------+ | |
587 | | Application | | | | Application | | |
588 | +---------------+ | | +---------------+ | |
590 | | v | | | v | | Guests
591 | +---------------+ | | +---------------+ | |
592 | | logical ports | | | | logical ports | | |
593 | | 0 1 | | | | 0 1 | | |
594 +---+---------------+--+ +---+---------------+--+__|
598 +---+---------------+---------+---------------+--+ |
599 | | 0 1 | | 3 4 | | |
600 | | logical ports | | logical ports | | |
601 | +---------------+ +---------------+ | |
603 | | |-----------------| v | |
604 | +--------------+ +--------------+ | |
605 | | phy ports | vSwitch | phy ports | | |
606 +---+--------------+----------+--------------+---+_|
610 +--------------------------------------------------+
612 | traffic generator |
614 +--------------------------------------------------+]]></artwork>
616 <preamble>Physical port to virtual switch to VNF</preamble>
618 <artwork><![CDATA[ __
619 +---------------------------------------------------+ |
621 | +-------------------------------------------+ | |
622 | | Application | | |
623 | +-------------------------------------------+ | |
627 | +---------------+ | |
628 | | logical port 0| | |
629 +---+---------------+-------------------------------+ __|
633 +---+---------------+------------------------------+ |
634 | | logical port 0| | |
635 | +---------------+ | |
639 | +--------------+ | |
640 | | phy port | vSwitch | |
641 +---+--------------+------------ -------------- ---+ __|
645 +--------------------------------------------------+
647 | traffic generator |
649 +--------------------------------------------------+]]></artwork>
651 <preamble>VNF to virtual switch to physical port</preamble>
653 <artwork><![CDATA[ __
654 +---------------------------------------------------+ |
656 | +-------------------------------------------+ | |
657 | | Application | | |
658 | +-------------------------------------------+ | |
662 | +---------------+ | |
663 | | logical port | | |
664 +-------------------------------+---------------+---+ __|
668 +------------------------------+---------------+---+ |
669 | | logical port | | |
670 | +---------------+ | |
674 | +--------------+ | |
675 | vSwitch | phy port | | |
676 +-------------------------------+--------------+---+ __|
680 +--------------------------------------------------+
682 | traffic generator |
684 +--------------------------------------------------+]]></artwork>
686 <preamble>VNF to virtual switch to VNF</preamble>
688 <artwork><![CDATA[ __
689 +----------------------+ +----------------------+ |
690 | Guest 1 | | Guest 2 | |
691 | +---------------+ | | +---------------+ | |
692 | | Application | | | | Application | | |
693 | +---------------+ | | +---------------+ | |
696 | +---------------+ | | +---------------+ | |
697 | | logical ports | | | | logical ports | | |
698 | | 0 | | | | 0 | | |
699 +---+---------------+--+ +---+---------------+--+__|
703 +---+---------------+---------+---------------+--+ |
705 | | logical ports | | logical ports | | |
706 | +---------------+ +---------------+ | |
708 | L-----------------+ | |
711 +------------------------------------------------+_|]]></artwork>
714 <t>A set of Deployment Scenario figures is available on the VSPERF Test
715 Methodology Wiki page <xref target="TestTopo"/>. </t>
718 <section title="3x3 Matrix Coverage">
719 <t>This section organizes the many existing test specifications into the
720 "3x3" matrix (introduced in <xref target="I-D.ietf-bmwg-virtual-net"/>).
721 Because the LTD specification ID names are quite long, this section is
722 organized into lists for each occupied cell of the matrix (not all are
723 occupied, also the matrix has grown to 3x4 to accommodate scale metrics
724 when displaying the coverage of many metrics/benchmarks).</t>
726 <t>The tests listed below assess the activation of paths in the data
727 plane, rather than the control plane.</t>
729 <t>A complete list of tests with short summaries is available on the
730 VSPERF "LTD Test Spec Overview" Wiki page <xref target="LTDoverV"/>.</t>
732 <section title="Speed of Activation">
733 <t><list style="symbols">
734 <t>Activation.RFC2889.AddressLearningRate</t>
736 <t>PacketLatency.InitialPacketProcessingLatency</t>
740 <section title="Accuracy of Activation section">
741 <t><list style="symbols">
742 <t>CPDP.Coupling.Flow.Addition</t>
746 <section title="Reliability of Activation">
747 <t><list style="symbols">
748 <t>Throughput.RFC2544.SystemRecoveryTime</t>
750 <t>Throughput.RFC2544.ResetTime</t>
754 <section title="Scale of Activation">
755 <t><list style="symbols">
756 <t>Activation.RFC2889.AddressCachingCapacity</t>
760 <section title="Speed of Operation">
761 <t><list style="symbols">
762 <t>Throughput.RFC2544.PacketLossRate</t>
764 <t>CPU.RFC2544.0PacketLoss</t>
766 <t>Throughput.RFC2544.PacketLossRateFrameModification</t>
768 <t>Throughput.RFC2544.BackToBackFrames</t>
770 <t>Throughput.RFC2889.MaxForwardingRate</t>
772 <t>Throughput.RFC2889.ForwardPressure</t>
774 <t>Throughput.RFC2889.BroadcastFrameForwarding</t>
778 <section title="Accuracy of Operation">
779 <t><list style="symbols">
780 <t>Throughput.RFC2889.ErrorFramesFiltering</t>
782 <t>Throughput.RFC2544.Profile</t>
786 <section title="Reliability of Operation">
787 <t><list style="symbols">
788 <t>Throughput.RFC2889.Soak</t>
790 <t>Throughput.RFC2889.SoakFrameModification</t>
792 <t>PacketDelayVariation.RFC3393.Soak</t>
796 <section title="Scalability of Operation">
797 <t><list style="symbols">
798 <t>Scalability.RFC2544.0PacketLoss</t>
800 <t>MemoryBandwidth.RFC2544.0PacketLoss.Scalability</t>
804 <section title="Summary">
806 <artwork><![CDATA[|------------------------------------------------------------------------|
808 | | SPEED | ACCURACY | RELIABILITY | SCALE |
810 |------------------------------------------------------------------------|
812 | Activation | X | X | X | X |
814 |------------------------------------------------------------------------|
816 | Operation | X | X | X | X |
818 |------------------------------------------------------------------------|
820 | De-activation | | | | |
822 |------------------------------------------------------------------------|]]></artwork>
827 <section title="Security Considerations">
828 <t>Benchmarking activities as described in this memo are limited to
829 technology characterization of a Device Under Test/System Under Test
830 (DUT/SUT) using controlled stimuli in a laboratory environment, with
831 dedicated address space and the constraints specified in the sections
834 <t>The benchmarking network topology will be an independent test setup
835 and MUST NOT be connected to devices that may forward the test traffic
836 into a production network, or misroute traffic to the test management
839 <t>Further, benchmarking is performed on a "black-box" basis, relying
840 solely on measurements observable external to the DUT/SUT.</t>
842 <t>Special capabilities SHOULD NOT exist in the DUT/SUT specifically for
843 benchmarking purposes. Any implications for network security arising
844 from the DUT/SUT SHOULD be identical in the lab and in production
848 <section anchor="IANA" title="IANA Considerations">
849 <t>No IANA Action is requested at this time.</t>
852 <section title="Acknowledgements">
853 <t>The authors acknowledge</t>
858 <references title="Normative References">
861 <?rfc include="reference.RFC.2119"?>
863 <?rfc include="reference.RFC.2330"?>
865 <?rfc include='reference.RFC.2544'?>
867 <?rfc include="reference.RFC.2679"?>
869 <?rfc include='reference.RFC.2680'?>
871 <?rfc include='reference.RFC.3393'?>
873 <?rfc include='reference.RFC.3432'?>
875 <?rfc include='reference.RFC.2681'?>
877 <?rfc include='reference.RFC.5905'?>
879 <?rfc include='reference.RFC.4689'?>
881 <?rfc include='reference.RFC.4737'?>
883 <?rfc include='reference.RFC.5357'?>
885 <?rfc include='reference.RFC.2889'?>
887 <?rfc include='reference.RFC.3918'?>
889 <?rfc include='reference.RFC.6201'?>
891 <?rfc include='reference.RFC.2285'?>
893 <reference anchor="NFV.PER001">
895 <title>Network Function Virtualization: Performance and Portability
896 Best Practices</title>
898 <author fullname="ETSI NFV" initials="" surname="">
902 <date month="June" year="2014"/>
905 <seriesInfo name="Group Specification"
906 value="ETSI GS NFV-PER 001 V1.1.1 (2014-06)"/>
912 <references title="Informative References">
913 <?rfc include='reference.RFC.1242'?>
915 <?rfc include='reference.RFC.5481'?>
917 <?rfc include='reference.RFC.6049'?>
919 <?rfc include='reference.RFC.6248'?>
921 <?rfc include='reference.RFC.6390'?>
923 <?rfc include='reference.I-D.ietf-bmwg-virtual-net'?>
925 <reference anchor="TestTopo">
927 <title>Test Topologies
928 https://wiki.opnfv.org/vsperf/test_methodology</title>
938 <reference anchor="LTDoverV">
940 <title>LTD Test Spec Overview
941 https://wiki.opnfv.org/wiki/vswitchperf_test_spec_review </title>
951 <reference anchor="IFA003">
953 <title>https://docbox.etsi.org/ISG/NFV/Open/Drafts/IFA003_Acceleration_-_vSwitch_Spec/</title>