摘要
Offlinenetworktrafficanalysisisveryimportantforanin-depthstudyupontheunderstandingofnetworkconditionsandcharacteristics,suchasuserbehaviorandabnormaltraffic.WiththerapidgrowthoftheamountofinformationontheInternet,thetraditionalstand-aloneanalysistoolsfacegreatchallengesinstoragecapacityandcomputingefficiency,butwhichistheadvantagesforHadoopcluster.Inthispaper,wedesignedanofflinetrafficanalysissystembasedonHadoop(OTASH),andproposedaMapReduce-basedalgorithmforTopNuserstatistics.Inaddition,westudiedthecomputingperformanceandfailuretoleranceinOTASH.FromtheexperimentswedrewtheconclusionthatOTASHissuitableforhandlinglargeamountsofflowdata,andarecompetenttocalculateinthecaseofsinglenodefailure.
出版日期
2013年05月15日(中国期刊网平台首次上网日期,不代表论文的发表时间)