外文翻译--基于哈希模式的负载均衡性能研究.docx

外文翻译--基于哈希模式的负载均衡性能研究.docx

《外文翻译--基于哈希模式的负载均衡性能研究.docx》由会员分享，可在线阅读，更多相关《外文翻译--基于哈希模式的负载均衡性能研究.docx（22页珍藏版）》请在冰豆网上搜索。

.

中文5790字

毕业设计外文翻译

专 业 网络工程

班 级

学生姓名 xx

学 号 xx

指导教师

.

PerformanceofHashing-BasedSchemesforInternetLoadBalancing

ZhiruoCao,ZhengWang,EllenZegura

CollegeofComputingGeorgiaInstituteofTechnology

Atlanta,GA30332-0280BellLabs

LucentTechnologiesHolmdel,NJ07733

Abstract—LoadbalancingisakeytechniqueforimprovingInternetperformance.

Effectiveuseofloadbalancingrequiresgoodtrafficdistributionschemes.Westudytheperformanceofseveralhashingschemesfordistributingtrafficovermultiplelinkswhilepreservingtheorderofpacketswithinaflow.Althoughhashing-basedloadbalancingschemeshavebeenproposedinthepast,thisisthefirstcomprehensivestudyoftheirperformanceusingrealtraffictraces.

Weevaluatefivedirecthashingmethodsandonetable-basedhashingmethod.

Wefindthathashingusinga16-bitCRCovertheFivetuplegivesexcellentloadbalancingperformance.Further,load-adaptivetable-basedhashingusingtheexclusiveORofthesourceanddestinationIPaddressesachievescomparableperformancetothe16-bitCRC.Table-basedhashingcanalsodistributetrafficloadaccordingtounequalweights.Wealsoreportonfourotherschemeswithpoortomoderateperformance.

Keywords—Loadsharing,hashing.

I.INTRODUCTION

Loadbalancing（alsoknownasloadsharing）isakeytechniqueforimprovingtheperformanceandscalabilityoftheInternet.Forexample,manylargeenterprisenetworksareconnectedtomultipleInternetServiceProviders（ISPs）toachieveredundantconnectivityandtodistributetrafficloading.InsidetheInternet,thebackbonesareoftenengineeredtohavemultipleparalleltrunksbetweenmajorPointsofPresencetoensurehighavailability.Typically,theseparalleltrunksareconfiguredasequal-costpathsandallowloadbalancingoverthem.

TheparalleltrunksmaybecomeevenmoreubiquitouswhenthepromisingDenseWavelengthDivisionMultiplexing（DWDM）technologyisdeployedinthefutureInternetback-bone.DWDMexpandsthecapacityofcommunicationtrunksbyallowingagreaternumberofchannelstobecarriedonasingleopticalfiber.WithpotentiallytensorevenhundredsofDWDMchannelsbetweenmajorpoints,loadbalancingisessentialinbestutilizingthemultipleparallelchannels.

ParallelarchitectureshavebeenusedforpacketprocessingforcopingwithexponentialgrowthinInternettraffic,Insteadofoneprocessingengine,packetsaredispatchedtomultipleparallelenginesinsidearoutertoincreasetheoverallprocessingthroughput.Thesametechniqueisalsousedinscalingwebservers.

PopularwebserversoftenoperateafarmofmachinesandtheroutersconnectedtothemsplittheHTTPrequeststodifferentmachines.

Foralloftheseexamples,effectiveuseofloadbalancingrequiresgoodschemesforsplittingtrafficovermultiplelinks.Inaddition,sincethemajorityofthetrafficontheInternetisTCP-based[1],trafficsplittingschemesneedtoavoidpacketmisorderingwithinaTCPflow,whichcanfalselytriggercongestioncontrolmechanismsandcauseunnecessarythroughputdegradation[2],[3].

Inthispaper,weproposeandevaluateaclassofhashingbasedtrafficsplittingalgorithmswhichpreserveper-flowpacketordering.Weconsiderfivehashfunctionsthatare“direct,”meaningthatthehashfunctionproducesavalueintherangeof0...N-1,whereNisthenumberofoutgoinglinks.Wealsoconsideratable-basedgeneralizationthatinvolveshashingtoMbins,thenassigningtheMbinstotheNoutgoinglinks.Tablebasedhashingrequiresmorestatethandirecthashing,buthastheflexibilitytosupportunequalloaddistributionanddynamicadaptation.

Ourresultsareobtainedbysimulatingtheperformanceofatrafficsplitter,usingpackettracestakenfromtwotrunksofamajorInternetbackboneprovider.WefindthatdirecthashingwiththedestinationIPaddresscausessignificantimbalanceacrosstwolinks.UsingtheInternetchecksumortheexclusiveORofboththesourceIPaddressanddestinationIPaddressimprovestheperformanceconsiderably,thoughmoderateimbalancepersists.Themorecomputationallycomplex16-bitCRCofthefive-tuple（sourceaddress,destinationaddress,sourceport,destinationportand

protocolid）givesexcellentloadbalancingperformance,keepingtheloadandqueuelengthsverysimilarontwolinks.Equallygoodloadbalancingcanbeachievedusingtable-basedhashingwithadaptation,whichrequireslesscomputationthantheCRCbutnecessitatesmonitoringthelinkloadsandstoring（andadjusting）themappingfromtablebinstolinks.

Table-basedhashinghastheadditionaladvantagethatitcandistributetheloadaccordingtounequalweights.Further,anindex-basedversionofthisschemecanaltertheweightdistributionwithminimaldisruptiontoexistingflows.Ourresultsconfirmthattheindex-based