关于ip骨干网络的流量矩阵估计方法分析

《关于ip骨干网络的流量矩阵估计方法分析》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、基于lP骨干网络的流量矩阵估计方法研究AbstractWiththerapiddevelopmentofInternet，thescaleofthenetworkhasbecomemoreandmorelarge，andthestructurehasbecomemoreandmorecomplex，andthenumberofIntemetusershasgrownexponentially．However，thenon-criticalbusinessinthenetworkhasalsolargelyconsumedthebandwid

2、thresourcesofnetwork，affectingtherunningofothercriticalbusiness．Thesehaveledtoitsmonitoringandmanagementincreasinglydifficult．Inordertocarryoutbetternetworkplanning，networkdesign，networkmanagement，networkmonitoring，routingconfiguration，networktrafficen#neeringandnetworksim

3、ulation，thetrafficinformationofnetworkisneededurgently．Trafficmatrix，whichisanoverviewofthewholenetwork，isacompletedescriptionoftrafficflowsanditsdistribution．Combinedwithnetworkroutinginformation，itcanalsoclearlyreflecttrafficcomponentofeachlinkinthenetwork．Itisakeyinputp

4、arameterofnetworktrafficengineeringandnetworkmanagement·However，large．scale，trafficmatrixiSdifficulttoobtainthroughdirectmeasurementmethodinthecomplexnetwork．Currently，estimatingtrafficmatrixthroughthelimitedmeasuredinformationhasbecomethemainmethod．Theproblemoftrafficmatr

5、ixestimationiSanill—posedlinearinverseproblem．Thisarticledescribesthedevelopmentprocessoftrafficmatrixestimation．Representativeofeverystagesoftrafficmatrixestimationmethodsaredescribedindetail．Weanalyzetheadvantagesanddisadvantagesofeachmethod．Theinnovativeachievementsofth

6、isarticleareinthefollowingtwoaspects．Fori11．posedcharacteristicsoftrafficmatrixestimation，weavoidthetraditionalideaofestimationalgorithm，namely，increasingtheconstraintconditionbymodelingtheorigin．destinationflowstoovercometheconstraintsofill—posedcharacteristics．Throughsta

7、tisticalanalysisinlargequantitiesoftheactualmeasuredtrafficmatrixdata，weassumpethattraffiCmatrixhasthecharacteristicofspatialself-similarity．Weproposedlinearmappingmethodbasedonspatialsimilarity．Experimentsshowthatthealgorithmcalculatesquicklyandaccurately．Complexnatureofn

8、etworktrafficmakesthecurrentresearcherstendtousemoresophisticatedandcomplexmodelfortraffi