简介:Wehaveprototypedandanalyzeddesignofanovelapproachforthehighthroughputcomputing-acoreelementfortheemergingHENPcomputationalgrid.IndependenteventprocessinginHENPiswellsutedforcomputinginparallel.Theprototypefacilitatedsuseofinexpensivemass-marketcomponentsbypovidingfaulttolerantresilienece(insteadoftheexpensivetotalsystemreliablity)viahighlyscalablemanagementcomponents.TheabilitytohandlebothhardwareandsoftwarefailuresonalargededicatedHENPfacilitylimitstheneedforuserintervention.ArobustdatamanagementisespeciallyimportantinHENPcomputingsincelargedata-flowsoccurbeforeand/oratfereachprocessingtask.Thearchitectureofouractiveobjectobjectcoordinationschemaimplementsamulti-levelhierarchicalagentmodel,Itprovidesfaulttolerancebysplittingalargeoveralltaskintoindependentatomicprocesses,performedbylowerlevelagentssynchronizingeachotherviaalocaldatabase.Necessarycontrolfunctionperformedbyhigherlevelagentsinteractwiththesamedatabasethusmanagingdistributeddataproduction.ThesystemhasbeentestedinproductionenvironmentforsimulationsintheSTARexperimentatRHIC.Ourarchitecturalprototypecontrolledprocessesonmorethanahundredprocessorsatatimeandhasrunforextendedperiodsoftime.Twentyterabytesofsimulateddatahavabeenproduced.ThegenericnatureofourtwolevelarchitecturalsolutionfaulttoleranceindistributedenvironmenthasbeendemonstratedbyistsuccessfultestforthegridfilereplicationservicesbetweenBNLandLBNL.
简介:Thispaperistoconstructa'digitallocal,regional,region'informationframeworkbasedonthetechnologyof'SIG'anditssignificanceandapplicationtotheregionalsustainabledevelopmentevaluationsystem.First,theconceptofthe'gridcomputing'and'SIG'isinterpretedanddiscussed,thentherelationshipbetweenthe'gridcomputing'and'digitalregion'isanalyzed,andtheframeworkofthe'digitalregion'isputforward.Finally,thesignificanceandapplicationof'gridcomputing'tothe'regionsustainabledevelopmentevaluationsystem'arediscussed.
简介:Artificalneuralnetworks(ANN)arenowwidelyusedsuccessfullyastoolsforhithenergyphysics.Thepapercoverstwoaspects.First,mappingANNsontotheproposedringandlinearsystolicarrayprovidesanefficientimplementationofVLSI-basedarchitecturessinceinthiscaseallconnectionsamongprocessingelementsarelocalandregular,Second.itisdiscussedalgorthmicorganizingofsuchstructuresonthebaseofmodularalgebrawhoseusecanprovideanessentialincreaseofsystemthroughput.
简介:Thispapergivesabackgroundofcrackers,thensomeattackeventsthathavehappenedinIHEPnetworksareenumeratedandintroduced.Atlastahighlyefficientdefencesystemthatintegratesauthor'sexperience,researchresultsandhaveputinpracticeinIHEPnetworksenvironmentisdescribedindetail,ThispaperalsogivesnetworkandinformationsecurityadviceandprocessforhighenergyphysicscomputingenvironmentintheInstituteofHighEnergyPhysicsthatwillimplementinthefuture.
简介:Knowledgeisoneofthemostimportantelementsinproductionandthefirstresourceinenterprise.Byintroducingknowledgetodecisionmaking,thispaperpresentsakindofknowledgedecisionmakingmodel(KDMforshort)basedongeneralizedcomputing.
简介:Inordertofacetheexpectedincreaseinstatisticsbetweennowand2005,theBabarexperimentatSLACisevolvingitscomputingmodeltowardadistributedmultitiersystem.ItisforeseenthatdatawillbespreadamongTier-AcentersanddeletedfromtheSLACcenter,Aunifromcomputingenvirommentisbeingdeployedinthecenters,thenetworkbandwidthiscontinuouslyincreasedanddatadistributiontoolshasbeendesignedinordertoreachatransferrateof-100TBofdataperyear,Inparallel,smallerTier-BandCsitesreceivesubsetsofdata,presentlyinKanga-Root[1]formatandlaterinObjectivity[2]format,GRIDtoolswillbeusedforremotejobsubmission.
简介:Granularcomputingisanewintelligentcomputingtheorybasedonpartitionofproblemconcepts.ItisanimportantprobleminRoughSettheorytoprocessincompleteinformationsystemsdirectly.Inthispaper,agranularcomputingmodelbasedontolerancerelationforprocessingincompleteinformationsystemsisdeveloped.Furthermore,acriteriaconditionforattributionnecessityisproposedinthismodel.
简介:TheColliderDetectoratFermilab(CDF)collaborationrecordsandanalysesprotonanti-protoninteractionswithacenter-of-massenergyof2TeVattheTevatron,Anewcolliderrun,RunII,oftheTevatronstartedinApril.DuringitsmorethantwoyeardurationtheCDFexperimentexpectstorecordabout1PetaByteofdata.Withitsmulti-purposedetectorandcenter-ofmassenergyatthefrontier,theexperimentalprogramislargeandversatile.Theover500scientistsofCDFwillengageinsearchesfornewparticles,liketheHiggsbosonorsupersymmetricparticles,precisionmeasurementofelectroweakparameters,likethemassoftheWboson,measurementoftopquarkparametersandalargespectrumofBphysics.Theexperimenthastakendataandanalysedtheminpreviousruns.ForRunII,however,thecomputingmodelwaschangedtoincorporatenewmethodologies.thefileformatswitched.andbothdatahandlingandanalysissystemredesignedtocopewiththeincreaseddemands.Thispaper(4-036atChep2001)givesanoverviewoftheCDFRunIIcomputesystemwithemphasizeonareaswherethecurrentsystemdoesnotmatchinitialestimatesandprojections.Forthedatahandlingandanalysissystemamoredetaileddescriptionisgiven.
简介:Theemergenceofhighspeedwideareanetworksmakesgridcomputingareality.HowevergridapplicationsthatneedreliabledatatransferstillhavedifficultiestoachieveoptimalTCPperformanceduetometworktuningofTCPwindowsizetoimporvethebandwidthandtoreducelatencyonahighspeedwideareanetwork.ThispaperpresentsapureJavapackagecalledJPARSS(javaParallelSecureStream)thatdividesdataintopartitionsthataresentoverseveralparallelJavaStreamssimultaneouslyandallowsJavaorWebapplicationstoachieveoptimalTCPperformanceinagirdenvironmentwithoutthenecessityoftuningtheTCPwindowsize.SeveralexperimentalresultsareprovidedtoshowthatusingparallelstreamismoreeffectivethantuningTCPwindowsize.InadditionX.509certificatebasedsinglesign-onmechanismandSSLbasedconnectionestablishmentareintegratedintothispackage,Finallyafewapplicationsusingthispackagewillbediscussed.