外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx

上传人:b****2 文档编号:6048981 上传时间:2023-05-05 格式:DOCX 页数:21 大小:143.21KB
下载 相关 举报
外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx_第1页
第1页 / 共21页
外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx_第2页
第2页 / 共21页
外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx_第3页
第3页 / 共21页
外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx_第4页
第4页 / 共21页
外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx_第5页
第5页 / 共21页
外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx_第6页
第6页 / 共21页
外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx_第7页
第7页 / 共21页
外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx_第8页
第8页 / 共21页
外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx_第9页
第9页 / 共21页
外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx_第10页
第10页 / 共21页
外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx_第11页
第11页 / 共21页
外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx_第12页
第12页 / 共21页
外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx_第13页
第13页 / 共21页
外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx_第14页
第14页 / 共21页
外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx_第15页
第15页 / 共21页
外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx_第16页
第16页 / 共21页
外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx_第17页
第17页 / 共21页
外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx_第18页
第18页 / 共21页
外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx_第19页
第19页 / 共21页
外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx_第20页
第20页 / 共21页
亲,该文档总共21页,到这儿已超出免费预览范围,如果喜欢就下载吧!
下载资源
资源描述

外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx

《外文翻译基于哈希模式的负载均衡性能研究Word文档格式.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Word文档格式.docx(21页珍藏版)》请在冰点文库上搜索。

外文翻译基于哈希模式的负载均衡性能研究Word文档格式.docx

 

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

CollegeofComputing

GeorgiaInstituteofTechnology

Atlanta,GA30332-0280

BellLabs

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

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

Therestofthispaperisorganizedasfollows.InSectionIIwediscussrelatedworkintrafficsplittingandloadbalancing.SectionIIIdescribesthebehaviorofanidealtrafficsplitter,explainstherequirementsforapracticalsystem,anddefinestheperformancemetricsthatwillbeusedtoassessvarioushashing-basedschemes.ThesetofschemesthatweconsideraredescribedinSectionIV.TheresultsofourstudyaredescribedinSectionV,andincludeanalysisoftherandomnessinherentinthetracedata(SectionV-A).WeconcludeandmentionareasforfutureworkinSectionVI.

II.RELATEDWORK

Loadbalancinghasbeenusedintelecommunicationnetworksintheformofinversemultiplexing[4].Inversemultiplexingallowsserviceproviderstoofferwidebandchannelsbycombiningmultiplenarrowband56kbpsand64kbpstrunks[5].Theloadbalancingininversemultiplexingistypicallybasedonroundrobindistributionofpacketsorbytes[6],[7].

Ourworkdiffersfrominversemultiplexingintwoimportantdimensions.First,inversemultiplexingisdesignedforuseoverpoint-to-pointlinks;

itstechniquesarenottypicallyapplicablefornetworklayerloadbalancing.Internetloadbalancing,however,makesuseofthenaturalredundancyinthenetworktopology.Thepathsforloadbalancing,forexample,equal-costmulti-paths,arediscovereddynamicallybyroutingprotocols,suchasOSPF[8],ratherthanthroughconfiguration.Second,inordertomaintainsynchronizationandper-flowFIFOpacketorderingininversemultiplexing,itisnecessarytoaddextrapacketheaderswithsequencenumbersortokeepstateatbothendsofthechannel.But,implementingtheseadditionalmechanismsfornetworkloadbalancingrequiresanewnetworkprotocol.Incomparison,thehashing-basedschemescanmaintainper-flowpacketorderingandcanbeimplementedwithoutrequiringanyadditionalprotocolsupport.

Hashinghasbeenwidelyusedinindexingandsearching[9].Inthenetworkingcontext,hashing-basedalgorithmsforaddresslookup[10],flowidentification[11]andpacketdemultiplexing[12]havebeenproposedinthepast.Theuseofhashingfornetworkloadbalancingisnotnew.SomecommercialrouterproductshaveimplementedsimplehashingovertheIPdestinationaddresstodistributetraffic[13].IntheOSPFOptimizedMultipathprotocol(OSPF-OMP)[14],anumberofpossibleapproachesforloadbalancingovermultiplepathsarementioned,includingper-packetroundrobin,dividingdestinationprefixesamongavailablenexthopsintheforwardingtable,anddividingtrafficaccordingtoahashfunctionappliedtothesourceanddestinationpair.However,theproposedschemesarenotevaluatedwithsimulationorrealnetworkmeasurement.InthestudyofloadbalancingwithOSPF-OMP,perfecthashingisassumed[15].

Atrafficsplittingschemeusingrandomnumbersisproposedin[16].Itappliesthename-basedmappingsapproachtoloadbalancing[17].Inthisscheme,eachnext-hopisassignedwithaweightbasedonasimplepseudo-randomnumberfunctionseededwiththeflowidentifierandthenext-hopidentifier.Whenapacketarrives,theweightsaregenerated,andthenext-hopreceivingthehighestweightisusedforforwarding.Theschemeisapproximatelytimesasexpensiveasahashing-basedscheme,whereisthenumberofoutgoinglinks.Again,noperformancestudyontheproposedschemeispresented.

Itisclearthatalthoughhashing-basedschemesfortrafficsplittinghavebeenproposedinthepast,andsomesimpleschemeshaveevenbeenimplementedincommercialproducts,theperformanceofsuchschemeshasnotbeenadequatelyevaluated.Thispaperpresentsthefirstcomprehensiveperformancestudyonawiderangeofhashing-basedschemes,usingrealpackettracesfrombackbonenetworks.

III.FRAMEWORK

Inthissection,wedescribethebehaviorofanidealtrafficsplitter,explaintherequirementsforapracticalsystem,anddefinetheperformancemetricsforassessingvariousschemes.

A.ReferenceModel

AloadbalancingsystemtypicallycomprisesatrafficsplitterandmultipleoutgoinglinksasshowninFigure1.Insuchasystem,thetrafficsplitterreceivesanincomingpacketfromahigher-speedlinkandforwardsittooneofthelower-speedoutgoinglinks.Agoodloadbalancingsystemshouldbeabletosplitthetraffictothemultipleoutgoinglinksevenlyorbysomepre-definedproportion.

In[7],ithasbeenobservedthatthereisacloserelationshipbetweenfairqueuingandloadbalancing.Wenowextendtheirobservationtoamathematicalmodeltoobtaintheconstraintsforidealtrafficsplitting.

Letusfirstlookatanidealfluidmodelwherethetrafficisinfinitelydivisible.Supposethatthereareoutgoinglinksintheloadbalancingsystem,andthecapacityoflinkIisui.LetSi(T,t)be

展开阅读全文
相关资源
猜你喜欢
相关搜索
资源标签

当前位置:首页 > 考试认证 > 从业资格考试

copyright@ 2008-2023 冰点文库 网站版权所有

经营许可证编号:鄂ICP备19020893号-2