predicting number of zombies in a ddos attacks using isotonic regression

《predicting number of zombies in a ddos attacks using isotonic regression》由会员上传分享，免费在线阅读，更多相关内容在工程资料-天天文库。

1、PredictingNumberofZombiesinaDDoSAttacksUsingIsotonicRegressionB.B.GuptaandNadeemJamaliAbstractAnomalybasedDDoSdetectionsystemsconstructprofileofthetrafficnormallyseeninthenetwork,andidentifyanomalieswhenevertrafficdeviatefromnormalprofilebeyondathreshold.Thisdeviationintrafficbeyondthresholdisusedinthep

2、astforDDoSdetectionbutnotforfindingnumberofzombies.Thischapterpresentsanapproachthatutilizesthisdeviationintraffictopredictnumberofzom-biesusingisotonicregressionmodel.Arelationshipisestablishedbetweennumberofzombiesandobserveddeviationinsampleentropy.Internettypetopologiesusedforsimulationaregenerat

3、edusingTransit-StubmodelofGT-ITMtopologygen-erator.NS-2networksimulatoronLinuxplatformisusedassimulationtestbedforlaunchingDDoSattackswithvariednumberofzombies.Variousstatisticalper-formancemeasuresareusedtomeasuretheperformanceoftheregressionmodel.Thesimulationresultsarepromisingasweareabletopredi

4、ctnumberofzombiesefficientlywithverylesserrorrateusingisotonicregressionmodel.KeywordsDDoSattack·Intrusiondetection·Isotonicregression·Zombies·Entropy1IntroductionDDoSattacksareanimpendingthreattoInternetrelatedapplications.Itposesagravedangertousers,organizationsandinfrastructuresoftheInternet.ADDo

5、Sattackerattemptstodisruptatarget,inmostcasesawebserver,byfloodingitwithillegitimatepackets,usurpingitsbandwidthandovertaxingittopreventle-gitimateinquiriesfromgettingthrough[1,2].InanomalybasedDDoSdetectionmechanisms,theprofileofthetrafficnormallyseeninthenetworkisconstructedB.B.Gupta(B)IndianInstitu

6、teofTechnologyRoorkee,Roorkee,Indiae-mail:gupta.brij@gmail.comN.JamaliUniversityofSaskatchewan,Saskatoon,Canadae-mail:jamali@agents.usask.caT.Özyeretal.(eds.),MiningSocialNetworksandSecurityInformatics,145LectureNotesinSocialNetworks,DOI10.1007/978-94-007-6359-3_8,©SpringerScience+BusinessMediaDord

7、recht2013146B.B.GuptaandN.Jamaliandanomaliesareidentifiedwhenevertrafficdeviatesfromnormalprofilebeyondathreshold[3].Proposedapproachutilizesthisdeviationintrafficbeyondthresholdtopredictnumberofzombiesusingiso