PH D THESIS

-

Documents
276 pages
Obtenez un accès à la bibliothèque pour le consulter en ligne
En savoir plus

Description

Niveau: Supérieur, Doctorat, Bac+8
PH.D. THESIS presented at Louis Pasteur University, Strasbourg Department of Computer Science LSIIT Laboratory, UMR CNRS-ULP N:7005 For obtaining the degree: Louis Pasteur University Doctor of Philosophy (Ph.D) in Computer Science by Emil Ivov Optimizing Real-Time Communications over the Internet Protocol Public defense on April 1st 2008 with the following jury : Andrzej Duda, External evaluator Professor at ENSIMAG, Grenoble Eckhart Koerner, External evaluator Professor at the University of Mannheim Thomas Noel, Thesis advisor Professor at Louis Pasteur University, Strasbourg Jean-Jacques Pansiot, Internal evaluator Professor at Louis Pasteur University, Strasbourg David Simplot-Ryl, Examinator Professor at the Lille University of Science and Technology

  • grenoble eckhart

  • during discus- sions

  • external evaluator

  • effort

  • therefore like

  • many aspects

  • ve shared

  • efforts during

  • ipv6 mobility


Sujets

Informations

Publié par
Nombre de lectures 63
Langue English
Signaler un problème

PH.D.THESIS
presentedat
LouisPasteurUniversity,Strasbourg
DepartmentofComputerScience
LSIITLaboratory,UMRCNRS-ULPN:7005
Forobtainingthedegree:
LouisPasteurUniversity
DoctorofPhilosophy(Ph.D)in
ComputerScience
by
EmilIvov
OptimizingReal-TimeCommunications
overtheInternetProtocol
PublicdefenseonApril1st2008withthefollowingjury:
AndrzejDuda,Externalevaluator
ProfessoratENSIMAG,Grenoble
EckhartKoerner,Externalevaluator
ProfessorattheUniversityofMannheim
ThomasNoel,Thesisadvisor
ProfessoratLouisPasteurUniversity,Strasbourg
Jean-JacquesPansiot ,Internalevaluator
ProfessoratLouisPasteurUniversity,Strasbourg
DavidSimplot-Ryl ,Examinator
ProfessorattheLilleUniversityofScienceandTechnology2Acknowledgements
GettingtotheendofaPh.D.isatoughmatterandonecouldhardlygetthroughitalone.
ThroughoutmylastyearsintheLouisPasteurUniversityIhavebeenfortunatetomeet
andreceivehelpfrommanypeople. Oneofthepleasuresoffinallyfinishingthistaskis
theopportunitytothankthem.
First and foremost, I would like to express my gratitude to Thomas Noel. I could
hardly overstate his role in my life during the last years: not only would I have never
even thought of starting a Ph.D. if it hadn’t been him but I would have certainly not
finished it. Thank you, Thomas, for guiding me when I most needed it, for helping me
stayontrack,andforbeingagreatfriendduringthelastyears.
My work on the field of peer-to-peer real-time communication has evolved in a
context of tight collaboration with Enrico Marocco from Telecom Italia Labs, Turin. I
would like to express my most sincere gratitude to him for sharing his vast experience
inthefieldsofSIPandP2Poverlays.
I have started my Ph.D. together with Julien Montavont and we have worked to-
gether on subjects related to seamless IPv6 mobility. We have shared many common
problems and many of the ideas presented in this thesis have came up during discus-
sions with him. In addition to helping me out with his experience in geolocation as-
sisted mobility, he was also the first of us to finish his thesis and he has provided me
withgenerousadvicewhileIwaspreparingmine.
ThroughmyentirelifeIhavealwaysbeenabletocountontheunconditionalsupport
of my family. Mom, Dad, Tsveti, thank you very much for always being there for me
andfortoleratingallthewhimsandtheweirdeffectsthataPh.D.thesismayhaveona
student:).
Experimentation has been a very important part of my Ph.D. and work on the SIP
Communicatorandfmipv6.orgprojectshastakenasubstantialamountoftheeffortthat
I’ve spent on it. I would therefore like to thank Yana Stamcheva and Martin Andre for
workingwithmeontheseprojects. Ireallyenjoyedworkingwithyou! Thankyou!
iOne of the most difficult things for a Ph.D. student is having to learn and adopt a
properresearchattitude. Knowingwhattotakeforgranted,andwhattoprove,whento
leanonpreviouswork,andwhentoquestionitarechoicesthatarefarfrombeingeasy.
I esteem myself particularly lucky in this respect because I have had the possibility of
observing and learning this research attitude from a true scientist. I would hereby like
toexpressmyprofoundadmirationforJean-JacquesPansiot,whorepresentsinmyeyes
theveryimageofaresearcherandthankhimforalltheinterstingdiscussions: nomatter
howminortheymayhaveseemed,tometheyhavealwaysbeenasourceofinspiration
andadisplayofamodeltofollow.
Manythankstothemembersofmyjury: AndrzejDuda,EckhartKoerner,andDavid
Simplot-Ryl for spending the effort of reading the whole lot of pages contained in the
manuscriptofmythesis. Thankyou,Idoappreciatetheeffort!
IhavesharedthetimespentintheULPNetworkResearchTeamwithmanystudents,
engineersandinterns,allofwhichIammorethangladtohavemet. Iwouldliketotake
this chance and also say thanks to all of them for all the fun we’ve shared. Alexander
Pelov, Antoine Gallais, Arnaud Frey, Christophe Jelger, Cristina Tabacaru, Guillaume
Schreiner, Jean Lorchat, Jean-Marc Muller, Koshiro Mitsuya, Mickael Hoerdt, Nicolas
Dichtel,NicolasMontavont,PascalMerindol,RomainKuntz,SymphorienWanko,Tom
Remoleur,andVincentLucas,guys,ithasbeenapleasureworkingwithyou!
Aconsiderablepartofmyeffortsduringthepastyearshasbeendedicatedtodevel-
opment. One of the people that I have most learned from and that I consider in many
aspectstobemymentorisMudumbaiRanganathanfromtheAdvancedNetworkTech-
nologiesDivisionoftheNationalInstituteofStandardsandTechnologyinUSA.Ranga,
thanks for always being there and always taking the time to discuss all the matters that
Ineededhelpon!
Lastbutcertainlynotleast,IwouldliketoprofoundlythankmybelovedVéronique
Dupont,forsharingmylifesincetheverystartofmyPh.D.adventures;),forgivingme
all necessary support and reassuring me every time I needed it while writing this thesis
(andthatmeansmanytimesindeed). Thankyou,Véro!
iiTomyfamilyivContents
Acknowledgements i
1 Introduction 1
1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Managingmicromobility . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 UsingPeer-to-Peerinreal-timecommunication . . . . . . . . . . . . . 5
1.4 Therestofthispaper... . . . . . . . . . . . . . . . . . . . . . . . . . . 6
I Transparentmobilityforreal-timecommunication 9
2 IntroductionandstateoftheartinthefieldofIPmobility 11
2.1 MobilityBackground . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2 Layer2mobilityinIEEE802.11WLANnetworks . . . . . . . . . . . 13
2.2.1 Scanning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2.2 Authentication . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2.3 Association . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.3 Layer3mobilityinIPv6networks . . . . . . . . . . . . . . . . . . . . 17
2.3.1 TheMobileIPv6protocol . . . . . . . . . . . . . . . . . . . . 17
2.3.2 Networkmobility(NEMO). . . . . . . . . . . . . . . . . . . . 21
2.4 Applicationlayer . . . . . . . . . . . . . . . . . . . . . . . . 23
2.4.1 SessionInitiationProtocolBasics . . . . . . . . . . . . . . . . 24
2.5 Limitationsofexistingstandards . . . . . . . . . . . . . . . . . . . . . 27
2.5.1 Layer2handoverlatency . . . . . . . . . . . . . . . . . . . . . 27
v2.5.2 Layer3handoverlatency . . . . . . . . . . . . . . . . . . . . . 28
2.6 WorkonoptimizingtheLayer2handoverprocedure . . . . . . . . . . 29
2.6.1 IAPPandContextCachingusingNeighborGraphs . . . . . . . 30
2.6.2 SelectivescanningandAPcache . . . . . . . . . . . . . . . . . 32
2.6.3 SynchronizedBeacons . . . . . . . . . . . . . . . . . . . . . . 34
2.6.4 ReverseengineeringCiscoSystem’sWirelessDomainService . 36
2.7 Layer3MovementDetectionOptimizations . . . . . . . . . . . . . . . 41
2.7.1 PeriodicRouterAdvertisementBeaconing(DenseRAs) . . . . 42
2.7.2 FastRouterAdvertisement(FastRA) . . . . . . . . . . . . . . . 42
2.7.3 RAcachinginAccessPoints(FastRouterDiscovery) . . . . . . 43
2.7.4 LinkLayertriggersontheMobileNode . . . . . . . . . . . . . 44
2.7.5 RALinkIdentificationforMobileIPv6MovementDetection. . 45
2.8 OptimizingDAD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.8.1 OptimisticDuplicateAddressDetection . . . . . . . . . . . . . 46
2.8.2 Duplicate Address Detection Optimization using IPv6 Multi-
castListenerDiscovery . . . . . . . . . . . . . . . . . . . . . . 46
2.9 WorkonoptimizingtheLayer3handoverprocedure . . . . . . . . . . 47
2.9.1 HierarchicalMobileIPv6 . . . . . . . . . . . . . . . . . . . . . 47
2.9.2 TheFMIPv6protocol . . . . . . . . . . . . . . . . . . . . . . . 49
2.9.3 S-MIP:ASeamlessHandoffArchitectureforMobileIP . . . . 52
2.9.4 Bi-casting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
2.10 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3 OptimizingVoIPmobilityattheapplicationlayer 61
3.1 Analyticalevaluationofapplicationlayerhandovers . . . . . . . . . . . 61
3.1.1 Descriptionoftheoptimization . . . . . . . . . . . . . . . . . 63
3.2 Implementationandtesting . . . . . . . . . . . . . . . . . . . . . . . . 64
3.2.1 Testbed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.2.2 Experimentalevaluationresults . . . . . . . . . . . . . . . . . 64
3.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
vi4 Evaluatingagenericlayer3solution 69
4.1 PreviousevaluationsoftheFMIPv6protocol . . . . . . . . . . . . . . 69
4.2 TestbedandTestScenarios . . . . . . . . . . . . . . . . . . . . . . . . 71
4.3 FMIPv6evaluationresultsandanalysis . . . . . . . . . . . . . . . . . 73
4.3.1 PredictiveHandovers . . . . . . . . . . . . . . . . . . . . . . . 73
4.3.2 Bufferingissues . . . . . . . . . . . . . . . . . . . . . . . . . 74
4.3.3 Reactivehandovers . . . . . . . . . . . . . . . . . . . . . . . . 75
4.3.4 CandidateAccessPointDiscovery . . . . . . . . . . . . . . . . 76
4.3.5 Resultssummary . . . . . . . . . . . . . . . . . . . . . . . . . 78
4.4 Drawingsomeconclusionsfromthisevaluation . . . . . . . . . . . . . 78
5 OptimizingIPv6mobility 81
5.1 MultipleinterfacesforIEEE802.11nodes . . . . . . . . . . . . . . . . 82
5.1.1 Analysisandevaluation . . . . . . . . . . . . . . . . . . . . . 84
5.1.2 Simulationresults . . . . . . . . . . . . . . . . . . . . . . . . 87
5.1.3 Necessaryfurtheranalysis . . . . . . . . . . . . . . . . . . . . 91
5.1.4 Summaryconclusiononthesofthandoverapproach . . . . . . 92
5.2 DoubleWirelessNetworkInterfaceswithFMIPv6 . . . . . . . . . . . 93
5.2.1 SolutionDescription . . . . . . . . . . . . . . . . . . . . . . . 93
5.2.2 Anexperimentalperformanceevaluation . . . . . . . . . . . . 94
5.2.3 Conclusionsandfuturework . . . . . . . . . . . . . . . . . . . 103
5.3 OptimizingFMIPv6withgeographicalpositioninginformation . . . . . 104
5.3.1 DatabaseextensionforAccessRouters . . . . . . . . . . . . . 105
5.3.2 Mobilitycache . . . . . . . . . . . . . . . . . . . . . . . . . . 105
5.3.3 SelectingthenextAccessPoint . . . . . . . . . . . . . . . . . 106
5.3.4 Handovermanagement . . . . . . . . . . . . . . . . . . . . . . 108
5.3.5 Experimentation . . . . . . . . . . . . . . . . . . . . . . . . . 110
5.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
6 Concludingourworkonmobility... andmovingon 117
6.0.1 Possiblefurtherimprovement . . . . . . . . . . . . . . . . . . 118
vii6.0.2 Wheretonow... . . . . . . . . . . . . . . . . . . . . . . . . . 119
II Real-timecommunicationsusingpeer-to-peeroverlays 121
7 Introduction: Usingpeer-to-peermethodsinreal-timecommunications 123
7.1 Commonarchitectures . . . . . . . . . . . . . . . . . . . . . . . . . . 124
7.1.1 Hostedproviders . . . . . . . . . . . . . . . . . . . . . . . . . 124
7.1.2 Privateinstallations . . . . . . . . . . . . . . . . . . . . . . . . 125
7.2 Problemswithexistingtopologies . . . . . . . . . . . . . . . . . . . . 125
7.2.1 Difficultscalabilityandlimitedflexibility . . . . . . . . . . . . 125
7.2.2 Repetitiveservices . . . . . . . . . . . . . . . . . . . . . . . . 126
7.2.3 Federation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
7.2.4 IPlayermobility . . . . . . . . . . . . . . . . . . . . . . . . . 127
7.3 ThePeer-to-PeerPromise . . . . . . . . . . . . . . . . . . . . . . . . . 128
7.3.1 Scalability . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
7.3.2 Abundanceofnew,customservices . . . . . . . . . . . . . . . 129
7.3.3 Federation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
7.3.4 IPlayermobility . . . . . . . . . . . . . . . . . . . . . . . . . 129
7.3.5 IPv4/IPv6interoperability . . . . . . . . . . . . . . . . . . . . 130
7.3.6 SharingmorethanCPUandbandwidth . . . . . . . . . . . . . 130
8 Stateoftheartinthefieldofpeer-to-peeroverlays 131
8.1 BasicconceptsofP2Parchitectures . . . . . . . . . . . . . . . . . . . 131
8.2 Peer-to-peerSecurity . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
8.2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
8.2.2 Admissioncontrol . . . . . . . . . . . . . . . . . . . . . . . . 135
8.2.3 Determiningthepositionintheoverlay . . . . . . . . . . . . . 136
8.2.4 Identificationanddissemination . . . . . . . . . . . . . . . . . 137
8.2.5 IntegrityinP2Pnetworks . . . . . . . . . . . . . . . . . . . . . 139
8.2.6 Theattackers . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
8.2.7 P2PinReal-TimeCommunication . . . . . . . . . . . . . . . . 141
viii