網(wǎng)絡(luò)鏈路傳輸控制的隨機(jī)Petri網(wǎng)建模與性能分析.pdf

1、分類號————UDC密級——單何代碼!Q151網(wǎng)絡(luò)鏈路傳輸控制的隨機(jī)Petri網(wǎng)建模與性指導(dǎo)教師李忠淮能分析姜云漢劉朝斌學(xué)何授予單位職稱教授副教授人連海事大學(xué)申請學(xué)位級別工學(xué)碩十學(xué)科與專業(yè)計算機(jī)應(yīng)用技術(shù)論文完成日期2007年3月論文答辯日期2007年3月答辯委員會主席叢坦生英文摘要ModelingandPerformanceAnalysisofNetworkLinkTransfersControlBasedonStochasticPet

2、riNetsAbstractSelfsimilarityllatBreofnetworktraffichassevereeffectonpackctlossrate，throughputandqueuecapacityofnetworklinktransfercontr01Tllisleadstosignificantresultsdifferencebetweenthetraditionalmodelandselfsimilarity

3、trafficinnetworkperformanceTherefore，thetraditionalMarkovianmodelisnolongerapplicabletoanalyzeperformanceofselfsimilaritytrafficThereisagreatneedtoadoptnewmodelsandtoolstostudytheperformanceofnetworklinktransfercontrolwi

4、thselfsimilaritynatureBasedonde印analysisofthecharacteristicsofdifferenttrafficflowinnetworklinktransfercontrol，stochasticPetrinetmodelsfordifferenttrafficflowarepresentedAsfarastheeffectsofselfsimilarityonnetworkcommunic

5、ationareconcerneddiversifiedtransfersourcestochasticPetrinetmodelsareputforwardinthispaperThenanonmarkovianstochasticPetrinetmethodisadoptedtoanalyzetheproposedmodelswithsimulationandverificacionbasedonSPNPt001Simulation

6、resultsshowthatthroughput，packetlossrateandretransmissionrateofnetworktransfercontrolgrowsslowlyforselfsimilaritytrafficflowbuttheaveragequeuecapacityincreasesquicklycharacterizedwithbursts，thereforemorebandwidthisneeded

7、OurworkprovidesaquantitativeanalysisapproachforimprovingtheperformanceofnetworklinkcontrolindifferentconditionsandpresentsanonmarkvianmodeltoanalyzetheperformanceKeyWords：selfsimilarity；stochasticPetriNets：liOn—markovlan