简介:Forworkflow-basedservicecompositionapproach,therelationsbetweentheWebserviceQoSandenvironmentsareusuallynotconsidered,sothattheinformationaboutQoSforcompositeserviceselectionisinaccurate.Itmakestheselectedcompositeserviceinefficient,orevenunexecutable.Toaddressthisproblem,anovelservicecompositionapproachbasedonproductionQoSrulesisproposedinthispaper.Generally,itisverydifficulttodirectlyanalyzehowdifferentkindsofenvironmentfactorsinfluencetheWebserviceQoS.Weadopt'black-box'analysismethodofoptimizingcompositeservices,discoveringtheknowledgesuchas'theQoSofoneWebservicewillbehigherinspecificenvironments'.Inourapproach,theexecutioninformationofthecompositeserviceisrecordedintoalogfirst,whichwillbetakenasthebasisofthesubsequentstatisticalanalysisanddatamining.Then,thetimelyQoSvaluesoftheWebservicesareestimatedandtheproductionQoSrulesbeingusedtoqualitativelyexpressthedifferentperformancesoftheWebserviceQoSindifferentenvironmentsaremined.Atlast,weemploytheminedQoSknowledgeoftheWebservicestooptimizethecompositeserviceselection.Extensiveexperimentalresultsshowthatourapproachcanimprovetheperformanceofselectedcompositeservicesonthepremiseofassuringtheselectingcomputationcost.
简介:Inthispaper,aquantitativeQoSmanagementimplementmechanisminIP-DiffServispresented.Theapproachhasthreemerits:(1)itoptimizesboththerouteselectedforserviceclassbytrafficengineeringinnetworklayerandtheserviceclassselectedfortheuserservicesbynetworkplaninservicelayer,(2)itensuresquantitativeQoSfortheuserserviceinDiffServ,(3)itachieveshighresourceutilization,andachievesbetterperformancewithlesscostfortheuserserviceandselectedroute.Simulationhasverifiedthesemerits.
简介:Mobileadhocnetworks(MANETs)aregainingalotofattentioninresearchlatelyduetotheirimportanceinenablingmobilewirelessnodestocommunicatewithoutanyexistingwiredorpredeterminedinfrastructures.Furthermore,inordertosupportthegrowingneedformultimediaandrealtimeapplications,qualityofservice(QoS)supportbythenetworkingprotocolisrequired.SeveralimportantQoSparametersthatareneededbysuchapplicationscanbeidentified.Theyincludebandwidth,end-to-enddelay,delayjitter,andbiterrorrate.Agoodamountofresearchhasbeendevelopedinthisareacoveringdifferentissuesandchallengessuchasdevelopingroutingprotocolsthatsupportbandwidthreservationanddelaymanagement.Inthispaper,thecurrentstateofresearchforQoSsupportinTDMA-basedMANETsatdifferentlayersofthenetworkingmodelispresentedandcategorized.Inaddition,thecurrentissuesandfuturechallengesinvolvedinthisexcitingareaofresearcharealsoincluded.
简介:随着Internet、移动网络和高性能网络的不断发展,在网络和不确定参数下具有多约未的QoS多播路由优化技术已成为网络及分布式系统领域的一个重要的研究课题,这也是下一代Internet和高性能网络的难题。本文探讨了具有QoS多播路由问题,逦描述了一种适应于研究QoS多播路由的网络模型。本文在网络环境及不确定参数下,提出了一种在网络规模和可行性方面为Internet、移动网络和其他高性能网络基于Tabu搜索的QoS多播路由优化算法(TQMRA)。TQMRA在网络环境及不确定参数下能够优化网络资源。仿真实验结果表明,TQMRA为QoS多播路由提供了一种新的有效途径。
简介:本文从网络发展从当前的Best-Effort到QoS的必然趋势出发,介绍了两种主要的QoS控制机制:IntServ和DiffServ体系结构,并通过对两者的不同特点的分析,对两种体系结构的有机结合提出初步的模块设计,并做出展望分析.