简介：ＦＲＡＭＥＷＯＲＫＯＦＡＤＩＳＴＲＩＢＵＴＥＤＥＮＧＩＮＥＥＲＩＮＧＤＢＭＳＷａｎｇＴａｏ；ＬｉｎＺｏｎｇｋａｉ；ＧｕｏＹｕｃｈａｉＳＭＯＯＴＨＳＵＲＦＡＣＥＩＮＴＥＲＰＯＬＡＴＩＯＮＯＶＥＲＡＲＢＩＴＲＡＲＹＴＲＩＡＮＧＵＬＡＴＩＯＮＳＢＹＳＵＢＤ...

简介：NewexperimentsincludingthoseattheLHCwillrequireanalysisofverylargedatasetswhicharebesthandledwithdistributedcomputation.WepresentthedesignanddevelopmentofaprototypeframeworkusingJavaandObjectivity.Ourframeworksolvessuchanalysis-specificproblemsasselectingeventsamplesfromlargedistributeddatabases.producingvarialbedistributions,andnegotiatingbetweenmultipleanalysisserviceproviders.ExamplesfromthesuccessfulapplicationoftheprototypetotheanalysisofdatafromtheL3experimentwillalsobepresented.

简介：Theobjectiveofconformancetestingistodeterminewhetheranimplementationundertest(IUT)conformstoitsspecification.Indistributedtestarchitecturewheretherearemultipleremotetesters,theobjectivecanbecomplicatedbythefactthattestersmayencountercontrollabilityandobservabilityproblemsduringtheapplicationofatestsequence.Acertainamountofworkhasbeendoneintheareaofgeneratingtestsequencethatisfreefromtheseproblems.However,fewresearchersinvestigatethemfromtheaspectoftestexecution.Thisworkstudiesthetestexecutionphasewhentestsequencesareappliedtotheimple-mentationanditispointedoutthatcontrollabilityandobservabilityproblemscanberesolvedifandonlyifthetestsystemim-plementssometimingconstraints.Whendeterminingtheseconstraints,thedynamictimeinformationduringtestistakenintoaccount,whichreducesthetestexecutiontimeandimprovestestefficiencyfurther.

简介：ＤＧＰＳＬ：ＡＤＩＳＴＲＩＢＵＴＥＤＧＲＡＰＨＩＣＳＬＩＢＲＡＲＹＳｈｉＪｉａｏｙｉｎｇ；ＰａｎＺｈｉｇｅｎｇ；ＺｈｅｎｇＷｅｎｔｉｎｇＤＧＰＳＬ：ＡＤＩＳＴＲＩＢＵＴＥＤＧＲＡＰＨＩＣＳＬＩＢＲＡＲＹ￥ＳｈｉＪｉａｏｙｉｎｇ；ＰａｎＺｈｉｇｅｎｇ；...

简介：这份报纸学习在一个空间领域上演变的一个分布式的双线性的系统的地区性的稳定。为地区性的弱、强壮、指数的稳定的足够的条件被给。另外，我们讨论一个地区性的最佳的稳定问题。获得的结果被例子和模拟说明。

简介：ADistributedMonitoringSystem(NGOP)thatwillscaletotheanticipatedrequirementsforRUnIIcomputinghasbeenunderdevelopmentatFermilab.NGOP[1]providesaframeworktocreateMonitoringAgentsformonitoringtheoverallstateofcomputersandsoftwarethatarerunningonthem.SeveralMonitoringAgentsareavailablewithinNGOPthatarecapableofanalyzinglogfiles,andcheckingexistenceofsystemdaemons,CPUandmemoryutilization,etc,NGOPalsoprovidescustomizablegraphicalhierarchicalrepresentationsofthesemonitoredsystems.NGOPisabletogenerateeventswhenseriousproblemshaveoccurredaswellasraisingalarmswhenpotentialproblemshavebeendetected.NGOPallowsperformingcorrectivactionsorsendingnotifications,NGOPprovidespersistentstorageforcollectedevents,alarmsandactions.AfirstimplementationofNGOPwasrecentlydeployedatFermilab.Thisisafullyfunctionalprototypethatsatisfiesmostoftheexistingrequirements.ForthetimebeingtheNGOPprototypeismonitoring512nodes.DuringthefirstfewmonthsofrunningNGOPhasprovedtobeausefultool.Multipleproblemssuchasnoderesets,offlineCPUs,anddeadsystemdaemonshavebeendetected.NGOPprovidedsystemadministratorswithinformationrequiredforbettersystemtuningandconfiguration.Thecurrentstateofdeploymentandfuturestepstoimprovetheprototypeandtoimplementsomenewfeatureswillbepresented.

简介：Thedistributedcomputersystemdescribedinthispaperisasetofcomputernodesinterconnectedinaninterconnectionnetworkviapacket-switchinginterfaces.Thenodescommunicatewitheachotherbymeansofmessage-passingprotocols.Thispaperpresentstheimplementationofrendezvousfacilitiesashigh-levelprimitivesprovidedbyaparallelprogramminglanguagetosupportinterprocesscommunicationandsynchronization.

简介：

简介：ThisworkhasoriginallybeendoneintheP.R.China(WuhanTechnicalUniversityofSurveyingandMapping)andinHongKong(ChineseUniversityofHongKong).Prof.LinHuigaveusefuladvice.Mr.HanHaiyang,andMs.ChenLilihelpedworkwiththeimplementationofprototypesystem.WealsothanktheSwissNationalScienceFoundationforitsfinancialsupportandacknowledgethecommentsofDr.AndreasGeppertandProf.KlausR.Dittrichonanearlierversionofthispaper.

简介：Sequentialsimulationoflargecomplexphysicalsystemsisoftenregardedasacomputationallyexpensivetask.Inordertospeed-upcomplexdiscrete-eventsimulations,theparadigmofParallelandDistributedDiscreteEventSimulation(PDES)hasbeenintroducedsincethelate70s.Inthispaperwe'llanalyzetheapplicabilityofPDEStothemodelingandanalysisoflargecomputersystem;suchsystemsareincreasinglycommonintheareaofHighEnergyandNuclearPhysics,becausemanymodernexperimentsmakeuseoflarge"computefarms",Somefeasibilitytestshavebeenperformedonaprototypedistributedsiumulator.

简介：Adistributedcontrolsystemisdesignedforplasmasprayingequipmentandtheconfigurationsofsystemsoftwareandhardwareisdiscussed.Throughfoundinganexpertdatabase,thesprayingprocessparametersareworkedoutandtheinitializationandcontrolofsprayingprocessarerealized.Theplasmasprayingsystemwiththiscontrolconfigurationcansimplifythesprayingoperation,improveautomationlevelofsprayprocess,andapproachtheexperiencecriterionassoonaspossible.

简介：Wereportanindex-coupleddistributedfeedbackquantumcascadelaserbyemployinganequivalentphaseshift(EPS)ofquarter-waveintegratedwithadistributedBraggreflector(DBR)atλ~5.03μm.TheEPSisfabricatedthroughextendingonesamplingperiodby50%inthecenterofasampledBragggrating.ThekeyEPSandDBRpatternarefabricatedbyconventionalholographicexposurecombinedwiththeopticalphotolithographytechnology,whichleadstoimprovedflexibility,repeatability,andcost-effectiveness.Stablesingle-modeemissioncanbeobtainedbychangingtheinjectioncurrentorheatsinktemperatureevenundertheconditionoflargedrivingpulsewidth.

简介：BuiltspecificallyfortheSemanticWeb,triplestoresarerequiredtoaccommodatealargenumberofRDFtriplesandremainprimarilycentralized.Astriplestoresgrowandevolvewithtime,thereisademandingneedforscalabletechniquestoremoveresourceandperformancebottlenecksinsuchsystems.Tothisend,weproposeafullydecentralizedpeer-to-peerarchitectureforlargescaletriplestoresinwhichtriplesaremaintainedbyindividualstakeholders,andasemantics-directedsearchprotocol,mediatedbytopologyreorganization,forlocatingtriplesofinterest.Wetestourdesignthroughsimulationsandtheresultsshowanticipatedimprovementsoverexistingtechniquesfordistributedtriplestores.Inadditiontoengineeringfuturelargescaletriplestores,ourworkwillinparticularbenefitthefederationofstand-alonetriplestoresoftodaytoachievedesiredscalability.

简介：

简介：

简介：Inordertofacetheexpectedincreaseinstatisticsbetweennowand2005,theBabarexperimentatSLACisevolvingitscomputingmodeltowardadistributedmultitiersystem.ItisforeseenthatdatawillbespreadamongTier-AcentersanddeletedfromtheSLACcenter,Aunifromcomputingenvirommentisbeingdeployedinthecenters,thenetworkbandwidthiscontinuouslyincreasedanddatadistributiontoolshasbeendesignedinordertoreachatransferrateof-100TBofdataperyear,Inparallel,smallerTier-BandCsitesreceivesubsetsofdata,presentlyinKanga-Root[1]formatandlaterinObjectivity[2]format,GRIDtoolswillbeusedforremotejobsubmission.

简介：Theauthordesignedtwoalgorithmsfordistributedcooperationamongmultiplevideostreamssharingcommoncommunicationresources.Thealgorithmstakeadvantageofanoptimizationframeworkthatcharacterizesvideopacketssuchthatjointresourceallocationcanbeimplementednotonlyoverthepacketsofasinglestream,butalsoacrosspacketsofdifferentstreams.Thefirstalgorithmenablescollaborationamongmultiplevideosendersinan802.11CSMA/CAwirelessnetworksuchthattheirjointperformanceismaximized.Viathealgorithm,theuserscooperativelyestablishtransmissionprioritiesbasedontheassignedcharacterizationsoftheirvideopackets.Thesecondtechniqueallowsforlow-complexityjointbandwidthadaptationofmultiplevideostreamsatintermediatenetworknodesintheIntemetinordertomaximizetheoverallnetworkperformance.Theauthoranalyzestheadvantagesoftheproposedalgorithmsoverconventionalsolutionsemployedinsuchscenarios.Itisshownthatdependingonsystemparameterssuchasavailablenetworkdataratetheproposedtechniquescanprovidesubstantialgainsinend-to-endperformance.

简介：Aprototypesystemforagent-baseddistributeddynamicservicesthatwillbeappliedtothedevelopmentofDataGridsforhigh-energyphysicsispresented.Theagent-basedsystemswearedesigninganddevelogpinggather,disseminateandcoordinateconfiguration,time-dependentstateandotherinformationintheGridsystemasawhole.Thesesystemsarebeingdevelopedasanenablingtechnologyforworkflow-managementandotherformsofend-to-endGridsystemmonitoringandmanagement.ThisprototypeisbeingdevelopedinJavaandisbasedontheJINI,MobileAgents,Self-OrganizingNeuralNetworks.

简介：Theelectronicmailhasbecometheindispensablefunctionindailyjob,andtheserverstabilityandperformancearerequired.UsingDCEandDFSwehavebuiltthedistributedelectronicmailsever,thatis,serverssuchasSMPT,IMAParedistributedsymmetrically,andprovidstheseamlessaccess.

简介：Thisarticleisaboutapieceofmiddleware,allowingtoconvertadumptapebasedTertiaryStorageSystemintoamultipetabyterandomaccessdevicewiththousandsofchannels.Usingtypicalcachingmechanisms,thesoftwareoptimizestheaccesstotheunderlyingStorageSystemandmakesbetteruseofpossiblyexpensivedrivesandrobotsorallowstointegratecheapandslowdeviceswithoutintroducingunacceptableperformancedegadation.Inaddition,usingthestandardNFS2protocol,thedCacheprovidesauniqueviewintothestoragerepository,hidingthephysicallocationofthefiledata,cachedortapeonly.BulkdatatransferissupportedthroughthekerberizedFTPprotocolandaC-API,providingtheposixfileaccesssemantics,Datasetstaginganddiskspacemanagementisperformedinvisiblytothedataclients.TheprojectisaDESY,FermilabjointefforttoovercomelimitationsintheusageoftertiarystorageresourcescommontomanyHEPlabs.ThedistributedcachenodesmayrangefromhighperformanceSGImachinestocommodityCERNLinux-IDElikefileservermodels.Differentcachenodesareassumedtohavedifferentaffinitiestoparticularstoragegroupsorfilesets.AffinitiesmaybedefinedmanuallyorarecalculatedbythedCachebasedontopologyconsiderations.Cachenodesmayhavedifferentdiskspacemanagementpoliciestomatchthelargevarietyofapplicationsfromrawdatatouseranalysisdatapools.