ATM上的視頻業(yè)務(wù)建模與性能分析.pdf

1、分婁號UDC嶄扭二￡竺、70單何代碼]Q!!ATM上的視頻業(yè)務(wù)建模與性能分析李妍指導(dǎo)教師李肚祖職稱副教授學(xué)位授了r單位大連海事大學(xué)申請學(xué)位級別碩學(xué)科與專業(yè)通信與信息系統(tǒng)論文完成同期20062論文答辯同期2006，3答辯委員會主席ModelingandCharacteristicAnalysisofVideoTrafficonATMAbstractInBISDNbasedonATMavarietyoftrafficdataisdivide

2、dintocellsofconstantlengthandcarriesoutvariousprocessionssuchastransmissionandexchangeonnetworksInthesamenetworkavarietyoftraffictypesinintegratedfashionistheadvantagethatATMbringson，butitscostisthatismustbeusedtohandlew

3、ithvariousburstytraffictypesonnetworksAburstisstochasticpropertiesofvarioustrafficcells缸e鋤sSohowtocharacterizetheperformanceofeellstreamsonnetworksisaproblemItisnecessarytomodeltoreflectthestatisticalpropertiesofvideosou

4、rcesaccuratelyinordertostudyvideostreamstransmissionWestudyvideomodelsinthecontextofstatisticalmultiplexingInthispaperwediscussthreetypesofvideosourcemodels：aautoregressivemodel，aMarkov—modulatedfluidflowmodel，andaMarkov

5、modulatedPoissonprocessmodel(MMPP)TheautoregressivemodelisusefulinprovidingasimplerepresentationofcodedvideosourcestatisticsforsimulationsofnetworkbehaviorwithvideotramcappliedHoweverbecausetheautoregressivemodelforvideo

6、cannotreadilybeappliedtotheanalyticaldeterminationofbuffersizeortheequivalentdelaydeterminationsinbroadband，cell一，orpacketbasednetworks，OthermodelssuchasthefluidflowmodelorMMPPmodelsmustbeintroducedtoanalyzetheimpactofvi

7、deotramconnetworksTheapproachtomatchfirstandsecondmomentsofcompositeMarkovchainmodelandtllemeasuredstatisticisusedtotheFwstmodelstoanalyzethecharacteristicoflOSSprobabilitywhilethehistogramapproximationisusedtoMMPPmodelN

8、everthelessnodemodelcanadapttoalltheregionsofthebuffersizeSoweanalysisthesmallandlargebufferregion，usingtheterminology“cell’’and‘burst’region，respectivelyThehistogramapproximationofMMPPmodelisusedtocellregion，andthefluid