2.5 and 5GBASE-T Technology: Multi-vendor Interop Webinar

© 2017 Ethernet Alliancewww.ethernetalliance.org www.nbaset.orgwww.ethernetalliance.org

2.5GAND5GBASE-TTECHNOLOGY–MULTI-VENDORINTEROPHERETODAYWEBINAR

2016BASE-TPlugfestReportOut

AmrikBains,CiscoDavidChalupsky, IntelGeorgeZimmerman,CMEConsulting

March9,2017

1

© 2017 Ethernet Alliancewww.ethernetalliance.org www.nbaset.org

Disclaimers

• Opinions expressed during this presentation are the views of the presenters, and should not be considered the views or positions of the Ethernet Alliance or NBASE-T Alliance.

• Per IEEE-SA Standards Board Bylaws, Dec 2013 • “At lectures, symposia, seminars, or educational

courses, an individual presenting information on IEEE standards shall make it clear that his or her views should be considered the personal views of that individual rather than the formal position of IEEE.”

2


EthernetAlliance- MissionandPrioritiesAglobalcommunityofendusers,systemvendors,componentsuppliersandacademiaØMission

• PromoteexistingandemergingIEEE802Ethernetstandards• Accelerateindustryadoption• Demonstratemulti-vendorinteroperability

Ø2017StrategicPriorities

• SupportExistingTechnologyDeployment• SupportIEEE802StandardsDevelopment• Marketing&Education

TheVoiceofEthernet

3

NBASE-TAlliance- MissionandPriorities

TheworldwidenetworkofcompaniesthatbreathenewlifeintonetworkinfrastructureØMission§ Tofacilitate,encourageandexpandthewidespreaduseand

deploymentof2.5Gand5GBASE-TEthernet.

Ø 2017Priorities§ ExpandNBASE-Tproductecosystem§ Promote802.3bzstandardandNBASE-Tspecifications§ Sponsorinteroperabilityprograms§ Educatemarketaboutmultipleapplicationsofthetechnology

4


AboutUNH-IOL

TheUNH-IOL isaneutral,third-partylaboratorydedicatedtotestingdatanetworkingtechnologiesthroughindustrycollaboration.

21Madbury Rd.,Ste 100,Durham,NH03824http://www.iol.unh.edu

5


Today’sSpeakers

6

GeorgeZimmermanPresident, CMEConsultingTechnicalChair, EthernetAllianceChiefEditor,IEEEP802.3bzTaskForceBoD,NBASE-TAlliance

DaveChalupskyPrincipalEngineer,IntelChair,IEEEP802.3bz2.5G/5GBASE-TTaskForceBoD,EthernetAllianceBoD,NBASE-TAlliance

AmrikBainsPrincipalEngineer,Cisco SystemsTechnicalWorkingGroupChair,NBASE-TAlliance

TheBASE-TFamily!!!


Agenda

• What’saPlugfest?• Why2.5G/5GBASE-T:arecap• PlugfestEventParticipants• TestDescriptions• Results

7


What’saPlugfest?(a.k.a.InteroperabilityEvent)

• AnimportantelementintheEthernetecosystem!– Anopen,multi-vendorenvironmenthasbeenkeytothewideadoption

ofEthernet• Goal:Forthebenefitofendusers,insurethatproductsfrom

differentcompanies,designedtothesamespecification,caninteroperate.

• Bringtogetherproductsanddevelopersfrommultiplecompaniesinaneutral,confidentialenvironment– Confidentialbecauseweareoftentestingpre-releasedproducts

• Testnewtechnology,newspecification– Theremaybesurprises…

• Testpre-releaseproduct– Findthesurprisesandfixthembeforetheproductshitthemarket

8


2016BASE-TInteroperabilityEvent• Co-sponsoredbytheEthernetAllianceandtheNBASE-TAlliance• Testingvenue:Univ.ofNewHampshire’sInteroperabilityLab(UNH-IOL)• EventheldtheweekofOctober10th,2016• EventFocus:

– Newtechnology:2.5G/5GBASE-T• IEEEStd 802.3bzTM-2016ratifiedinSeptember2016

– Demonstrateinteroperabilityof2.5G/5GBASE-Tports frommultiplemanufacturers

• Mostportstestedalsosupported 10GBASE-T,allportsalsosupported 1000BASE-T– Demonstratespeednegotiationcapabilitybetweensupported rates.– DemonstratecompatibilitybetweenNBASE-TandIEEE802.3bzspecifications– Provideanopportunity toworkthekinksofoutofnewtechnologyand

products– Demonstratethatthisstuffworks!

9


ParticipantOverview15participatingcompaniesrepresentingacrosssectionoftheEthernetecosystem:NIC,wirelessAP,switches,cabling, conformancetestequipment, cabletester,protocol testequipment, PHYsilicon.

10


11

802.3Ethernetand802.11WLAN:DeploymentsToday

• Theneedsof802.11acWave2accesspointsdemandeddevelopmentofNBASE-T/802.3bz

– Mustsupport:>1Gbpsrates,installedcablingandPoE

– Rampingfast!


Why2.5Gand5GBASE-T?

• Between2003to2014~70billionmeters1ofCat5eandCat6cablingweresold….

– In2014,1.3Billionoutlets,~90%ofinstalledbase– Enormousnetworkinfrastructureinvestment

• Existingspecificationssupport1Gb/soverthiscable,butfasterdataratesarepossible

• BASE-Tallowsincrementalupgrade

– Letsgetmorevaluefromthisinvestment!– EnabletheNextEthernetEra

10metersofcableforeveryman,woman,andchildonEarth…

12

1Source:BSRIADecember2014


Wherearewenow?• Providing 2.5Gbpsand5Gbpsoverinstalledcabling• 802.11acWave2APsshipping with>1Gbpswired

backhaul• IEEE802.3bzTM receivedfinalapprovalSeptember

2016• JointEA&NBASE-TPlugfestOctober16• Cablingguideline documentsnowavailableto

support rollouts

• Rapidtransitionfrom100/1000Mto2.5G/5GBASE-T– Combinedforecastsays25%ofswitchportswillbe

2.5G/5GBASE-T– 5millionswitchportsin2017,growingtoover

100Min2021*!

*DellOro– January2017

13


FillingouttheRoadmap

NBASE-TStart

(WLAN)

MachineVision(industrial)

PoE

Homenetworking

Serviceprovider

infrastructure

1000BASE-Treplacement

EnterpriseStorage

E T H E R N E T E C O S Y S T E M

14


IEEE/NBASE-TComparison

• AutomaticDownshiftcapabilitysupportedwithNBASE-T2.3– Preventslossofconnectivityinchallengingcableenvironments

15

IEEE802.3bzfunctionality

AutomaticDownshift

NBASE-T2.3functionality


IEEEandNBASE-T:CompatibilityandfunctionalitywithNBASE-T2.3• History:NBASE-T1.0released

priortostartofIEEE802.3bz.• NBASE-T2.3coverscompatibility

withbothIEEEStd 802.3bzTM -2016andNBASE-T1.1

• RecommendbuildingnewproductswithNBASE-T2.3support(notjustIEEE802.3bz)tomaintaincompatibilitywithlegacyNBASE-T1.1ports

• UpgradeNBASE-T1.1portswithnewNBASE-T2.3firmware

16

NBASE-T1.1port

IEEE802.3bzport

NBASE-T2.0port

OKOK

NotOK


TestCaseOverviewInteroperabilitytests• Linkuptest:teststhebasiccapabilityfortwodevices undertest(DUT)toestablishlink.• FrameErrorRate(FER)tests:dataframeswerepassesbetweendevicesandmonitoredforerrorstodemonstrate1E-12biterrorratiorequiredbythespecifications.

DownshiftTesting• ByinjectingacontrolledamountofnoiseontotheCat5ecablesusingtestequipment, downshift testsdemonstratethatportscantransitiontolowerdatarateswhenthenoiseexceedsthespecifications forahigherdatarate

6Around1Testing• Althoughnotarequirementofthestandard,6-around-1 testsprovide aninformativedemonstrationofaport’sabilitytooperateinthepresenceofnoise.Inthiscasenoise isinjectedontothe“victim”cablebytightlycoupling 6“aggressor”cablesaroundthevictim,andrunningactivelinks (1000BASE-T,2.5GBASE-T)ontheaggressorcableswhileFERtestsarerunonthevictim.

ConformanceTesting:• Electricaltestequipmentwasprovidedbytwodifferentequipmentvendors.Conformancetestswererunontheactiveportsusingthisequipment totesttheirconformancetothe2.5G/5GBASE-Tspecifications. Testequipment vendorsalsobenefitfromplugfestbyhavingmoreportstotestandoptimizetheirtestsuites.

17


InteroperabilityTestCases13differentdevicesfrom10companies…78uniquepairingsForeveryinteroperabilitypair,upto12testswereperformed1. BasicLinkestablishment

• [email protected]• Link@IEEE5G• [email protected]• Link@NBASE-T5G

2. BasicFER*

• *8-minutetests,800-byteframes ->Verifies1e-12BER@[email protected]

18

WithPanduitCategory5echannels... WithBerk-TekCategory5echannels...

● [email protected]


● FER@5Gwith5meterlinksegment







BasicLink/FERPairings– 78Combinations

DUT1 DUT2 DUT3 DUT4 DUT5 DUT6 DUT7 DUT8 DUT9 DUT10 DUT11 DUT12 DUT13

DUT1

DUT2 x

DUT3 x x

DUT4 x x x

DUT5 x x x x

DUT6 x x x D x

DUT7 x x x D x x

DUT8 x x x x x x x

DUT9 x x x x x x x x

DUT10 x x x x x x x x x x

DUT11 x x x D x x x x x D x

DUT12 x x x x x x x x x x x

DUT13 x x D x x D x x

NotAvailable/Working

x Confirmed ALLdatacollected

D MissingData

19


BasicLinkResultsOverviewResultsfor591LinkUpTrials

ConfidentialtoPlugfestParticipants 20

20


LinkSuccessRate

21


SpecificationUsedforAuto-Negotiation


22


SpeedofLink


23


Category5eLinkSegmentUsed


VendorAShort(5m)VendorALong(100m)VendorBShort (5m)VendorBLong(100m)

24


FERResultsOverviewResultsover523trials

ConfidentialtoPlugfestParticipants25

25


FrameErrorRateResults

26


SpeedTested

27


LinkSegmentUsed

VendorAShort(5m)VendorALong(100m)VendorBShort (5m)VendorBLong(100m)

28


DownshiftTestSetup/Procedure

29

EthernetTrafficGen+NoiseGen(AWGN)

DUT

1mPatchcords

Ø Devices are connected to tester via 1m patch cableØ All speeds are enabled on both devicesØ Link events are initiated by the tester by restarting auto-negotiationØ After link is established at the highest speed supported by DUT (10G or 5G)Ø AWGN noise is injected into the test channel while link is activeØ The noise power is increased by 1dB; recorded noise level which triggered downshift

NoiseGeneratorallowsconsistentrepeatabletestcomparedto6-a-1test


DownshiftTest:Results• AllportsimplementingDownshiftshowedsuccess!

Ø 10G-capableports(tenDUTs)performedDownshiftfrom10G->5G->2.5G->1Gasnoiseincreasedfrom-122dBm to-108dBmwithin1dBm/Hz

Ø 5G-maxports(twoDUTs)performedDownshift from5G->2.5G->1Gasnoiseincreasedfrom-119dBm to-108dBmwithin1dBm/Hz

Ø OnlyoneDUTpresentdidnot implementdownshift feature

30

DecreasingSpeed

10G

5G

2.5G1G

IncreasingNoisedBm/Hz


Whataboutthe“6Around1”Test?

31

• UseCasesØ Most2.5G/5GBASE-Tdeploymentshave

<<100mbundling:closeproximityatswitchend,thenfanouttoendpoints. Thismakes100mofbundlingaverypessimisticmodel.

• 6-around-1cablebundleconfiguration wasusedtocalculateworst-casealiencrosstalkforCat6acablesforIEEE802.310GBASE-T…makingtheterm“6around1”wellknowninthecablingindustry for10GBASE-T.

Ø Innercableisvictimof6aggressorsØ Noisedepends onwhatsignalsarein

thecables


32

q IEEE802.3bzevaluatescablingaliencrosstalkofbydeterminingthebestSignal-to-NoiseRatio(SNR)thataPHYcouldachieveusingmeasurementsofthealiencrosstalkcharacteristicsofthecabling– thisiscalledAlienLimitedSNR(ALSNR)

q AlienLimitedSNR– 2.5G/5Gperformanceislimitedbycross-talkthananyothercableparameter,hencetermALSNR

q IEEE802.3bzspecifiesthelinksegmentbasedontheworstpossiblesetofaggressors,independentofcabletypeor6-a-1cablebundleconfiguration,thatmeetatestforaliennoiselevel(calledtheALSNRcriterion)

q IEEE802.3bzspecifythatthePHYprovideabiterrorratio(BER)betterthanorequalto10–12 attheMAC/PLSserviceinterface

DoesitMeetIEEE802.3bzSpecifications?


“6Around1”TestSetupandResults

33

• PlugfestTest

Ø ALSNRwascalculatedforeachsetofdisturbers,butduetolimitedtime,noattemptwasmadetoensureALSNRmetIEEE802.3bzLinkSegmentspecification

• Testsetup

Ø 6-a-1Cablebundleusedwasworstcase:full100mbundling ofCat5e.Exceeds802.3bzALSNRspecsforanyconfiguration.

Ø …butclosetothespecfor2.5Gvictimwith1Gaggressors.

● 5m+90m+5m linksegment

TrafficGeneratorwith7ports

DUT

• Testresults:Ø 2.5Gdeviceundertestwith1Gaggressorsgenerallyworkedwell,

matchingspecificationforthisconfiguration


FuturePlugfestRecommendations

34

• Replace6-a-1cabletestwithcontrollednoisesourceemulateALSNRasspecifiedinClause126.5.4.4Aliencross-talknoiserejection

• BenefitsØ RepeatableandconsistentresultsØ EasiertocreateALSNRfordifferentcombination ofvictim/aggressor

speedØ Mapsmorecloselytospecificationsandusecases


ResultsDiscussionInteroperabilitytests:over1,000testcasesrun!• Linkuptest:591linktrialswith100%success.• FrameErrorRatetests:523FERtestswith99%passrate.• Interoperabilitytestcaseswereanearevensplitof

– NBASE-Tvs.IEEE802.3bzautonegotiation.– 2.5GBASE-Tvs.5GBASE-Toperation– Shortvs.long(5m/100m)cablingfromtwosuppliers

DownshiftTesting• Downshifttestingwith12devicesdemonstratingthecapabilitytooperateatthebestpossiblespeedwhennoisefactorsexceedspecificationlimits.

6Around1Testing• Demonstratesthatdevicescanoperateinthepresenceofnoiseoncableplantsexceedingthespecificationlimitsforaliencrosstalk(ALSNR).

ConformanceandCableTesting:• Testequipmentavailablefrommultiplesuppliersdemonstratesthatthetest&measurementcapabilityisinplacetoenabletheworkofcomponentandsystemdesigners,equipmentandcableinstallerstodeploy2.5GBASE-Tand5GBASE-Tproductswithconfidence.

35


Summary• 2.5G/5GBASE-Tmulti-vendorinteroperabilitydemonstrated

• CompatibilitybetweenNBASE-TandIEEE802.3bzdemonstrated

• Productsarealreadyinthemarket• Manymoreindevelopment

– Weshoulddothisagain!• Widevarietyofproducttypesandtestequipmentavailable– Richecosystemdeveloping

36


Questions?

37


If you have any questions or comments, please email [email protected]

Visit the Ethernet Alliance on Facebook

Ethernet Alliance: visit www.ethernetalliance.orgJoin the Ethernet Alliance LinkedIn groupFollow @EthernetAllianc on Twitter