Các phương pháp tiết kiệm năng lượng sử dụng công nghệ mạng điều khiển bằng phần mềm trong môi trường điện toán đám mây

MINISTRY OF EDUCATION AND TRAININGHANOIUNIVERSITYOFSCIENCEANDTECHN OLOGY TRANMANHNAM CÁC PHƯƠNG PHÁP TIẾT KIỆM NĂNG LƯỢNG SỬ DỤNGCƠNGNGHỆMẠNGĐIỀUKHIỂNBẰNGPHẦNMỀMTRO NG MƠITRƯỜNGĐIỆNTỐNĐÁM MÂY SDN-BASED ENERGY-EFFICIENT NETWORKING INCLOUDCOMPUTINGENVIRONMENTS DOCTORALTHESISOFTELECOMMUNICATIONSENGINEERING HANOI-2018 MINISTRY OF EDUCATION AND TRAININGHANOIUNIVERSITYOFSCIENCEANDTECHN OLOGY TRANMANHNAM CÁC PHƯƠNG PHÁP TIẾT KIỆM NĂNG LƯỢNG SỬ DỤNGCƠNGNGHỆMẠNGĐIỀUKHIỂNBẰNGPHẦNMỀMTRONGM ƠI TRƯỜNGĐIỆNTỐNĐÁMMÂY SDN-BASED ENERGY-EFFICIENT NETWORKING IN CLOUDCOMPUTINGENVIRONMENTS Specialization: Telecommunications EngineeringCodeNo: 62520208 DOCTORALTHESISOFTELECOMMUNICATIONSENGINEERING Supervisor:Assoc.Prof.NguyenHuuThanh HANOI-2018 PREFACE I hereby assure that the results presented in this dissertation are my work under theguidanceofmysupervisor.Thedataandresultspresentedinthedissertationarecompletelyhonest and havenotbeendisclosedinanypreviousworks.Thereferenceshavebeenfullycitedand in accordancewith theregulations Tôi xin cam đoan kết trình bày luận án cơng trình nghiên cứu tơidưới hướng dẫn giáo viên hướng dẫn Các số liệu, kết trình bày luận án làhoàn toàn trung thực chưa cơng bố cơng trình trước Các kếtquảsửdụngtham khảo đềuđãđượctrích dẫn đầyđủtheo đúngquyđịnh HàNội, Ngày19 tháng01năm 2018 Tácgiả TrầnMạnhNam ACKNOWLEDGEMENTS First and foremost, I would like to thank my advisor, Associate Prof Dr Nguyen HuuThanh, for providing an excellent researching atmosphere, for his valuable comments,constant support and motivation His guidance helped me in all the time and also in writingthisdissertation.IcouldnothavethoughtofhavingabetteradvisorandmentorformyPhD Moreover, I would like to thank Associate Prof Dr Pham Ngoc Nam, Dr Truong ThuHuongfortheiradvicesandfeedbacks,alsoformanyeducationalandinspiringdiscussions Mysinceregratitudegoestothemembers(presentandformer)oftheFutureInternetLab,Schoolof`Elec tronicsandTelecommunications,HanoiUniversityofScienceandTechnology Without their support and friendship it would have been difficult for me tocomplete myPhD studies Finally, I would like to express my deepest gratitude to my family They are alwayssupporting me and encouraging me with their best wishes, standing by me throughout mylife Hanoi, 19thJan 2018 CONTENTS LISTOFFIGURES viii LISTOFTABLES .x INTRODUCTION CHAPTER1 AN OVERVIEW OF ENERGYEFFICIENTAPPROACHESINCLOUDCOMPUTINGENVIRONMENTS.6 1.1 Today'sInternet 1.1.1 CloudComputingServicesandInfrastructures .6 1.1.2 Energyconsumptionproblems 1.2 AnOverviewofEnergy-EfficientApproaches 1.2.1 Energyconsumptioncharacteristics 1.2.2 Energy-EfficientApproaches'Classification 1.3 Software-definedNetworking(SDN)technology 10 1.3.1 SDNArchitecture 10 1.3.2 SDNSouthboundAPI-OpenFlowProtocol 11 1.3.3 SDNControllers .12 1.4 DifficultiesonNetworkEnergyEfficiencyandMotivations 13 1.5 Dissertation’sContributions 14 1.5.1 ProposinganenergyawareandflexibledatacenternetworkthatisbasedontheSDNtechnology 14 1.5.2 Proposingenergyefficientapproa ches inanetwork virtualization forcloudenviron ments 14 1.5.3 Proposinga n e n e r g y a w a r e d a t a c e n t e r v i r t u a l i z a t i o n f o r c l o u d environments 15 CHAPTER2 SDN-BASEDENERGY-AWARE DATACENTERNETWORK 16 2.1 BackgroundTechnologies 16 2.1.1 DCNtechniqueandarchitecture .16 2.1.2 Existingsystem .22 2.2 Power-ControlSystemofaDCNetwork 22 2.2.1 Energymodelingofanetwork 23 2.2.2 TheDiagramof thePower-ControlSystem 25 2.3 Energy-AwareRoutingbasedonPowerProfileof DevicesinDataCenterNetworksusingSDN 29 2.3.1 Energy-AwareRoutingandTopologyOptimizationAlgorithm 30 2.3.2 Performanceevaluation 36 2.4 GreenDataCenterusingcentralizedPowercontroloftheNetworkandservers 39 2.4.1 ExtendedPower-ControlSystem 40 2.4.2 Usecase 41 2.4.3 Topology-awareVMmigrationalgorithm 43 2.4.4 VMMigrationcostandPowermodelingof aServer .45 2.4.5 ExperimentalResults 45 2.5 Conclusion .48 CHAPTER3 ENERGYEFFICIENTN E T W O R K VIRTUALIZATIONFORCLOUDENVIRONMENTS 49 3.1 Network VirtualizationandVirtualNetworkEmbedding 51 3.2 ConstructingEnergy-AwareSDNbasedNetworkVirtualization System 51 3.2.1 System’sDiagram 52 3.2.2 System’sworkflow 53 3.3 ModelingandProblemFormulation 54 3.3.1 VNEModeling 54 3.3.2 ObjectiveandConstraints 55 3.3.3 Time-basedEmbeddingStrategies 57 3.4 Energy-efficientVNEalgorithms .58 3.4.1 Energy-costCoefficientofCapacity 58 3.4.2 VirtualNodeMappingalgorithms 59 3.4.3 VirtualLinkMapping(VLiM)Algorithm .62 3.5 PerformanceEvaluation 63 3.6 Conclusion .67 CHAPTER4 AN ENERGY-AWARE DATACENTERVIRTUALIZATIONFORCLOUDENVIRONMENTS 68 4.1 VirtualDCTechnologies 69 4.1.1 Virtualdatacenterembedding 69 4.1.2 Virtualmachinemigrationandserverconsolidation 71 4.1.3 Discussion 71 4.2 DesignObjectives 73 4.3 ProblemFormulation .74 4.3.1 DataCenterModeling 74 4.3.2 EnergyModelingofDCComponents 75 4.3.3 Energy-EfficientProblemFormulation 76 4.4 ANew Conceptfor VDCEmbedding 77 4.4.1 Energy-awareVDCarchitecture .77 4.4.2 Energy-awareVDCembeddingalgorithm 78 4.4.3 JointVDCEmbeddingandVMMigrationAlgorithms .81 4.5 PerformanceEvaluation 84 4.5.1 Performancecriteria .84 4.5.2 Numericalresults 85 4.6 Conclusion .91 CHAPTER5 5.1 5.2 CONCLUSIONANDFUTUREWORK 92 Majorcontributions 92 Futureresearchdirections .93 LISTOFPUBLICATIONS 94 REFERENCES .96 ABBREVIATIONS APCI APEX ASIC BAU BFS CAPEX DC DCN D-ITG EA-NV EA-VDC ECO FM FPGA GH HEA-E HEE IaaS ICT ISP MoA MST NaaS NFV NV OLD OPEX PaaS PCS PM POD PSnEP RMD-EE SaaS SDSN SN SNMP TCAM AdvancedConfiguration&PowerInterface Capitalexpenditure Applicationspecificintegratedcircuits Business-as-usual Breadth-firstSearch CapitalExpenditure Datacenter Datacenternetwork Distributedinternettrafficgenerator Energy-awarenetworkvirtualization Energy-awareVirtualDataCenter Ecosustainable Fullmigration Fieldprogrammablegatearrays GreenHead HeuristicEnergy-awareVDCEmbedding Heuristicenergy-efficient Infrastructure-as-a-service Informationandcommunicationtechnologies Internetserviceprovider Migrateonarrival Minimumspanningtree Network-as-a-service Networkfunctionvirtualization Networkvirtualization OpenDayLight Operatingexpenses Platform-as-a-service Power-ControlSystem Partialmigration Optimizeddatacenters Powerscalingand energy-profile-aware Reducingmiddle nodeenergyefficiency Software-as-a-service Software-DefinedSubstrateNetwork SecondNet Simplenetworkmanagementprotocol Ternarycontent-addressablememory VDC VDCE VLiM VM VmM VNE VNoM VNR Virtualdatacenter Virtualdatacenterembedding Virtuallinkmapping VirtualMachine Virtualmachinemapping Virtualnetworkembedding Virtualnode mapping Virtualnetworkrequests LISTOFFIGURES Figure 1.1: Estimate of the global carbon footprint of ICT (including PCs, telcos’ networksanddevices,printersand datacenters)[15] Figure1.2:EnergyconsumptionestimationfortheEuropeantelcos’networkinfrastructuresinthe”Busi ness-As-Usual”(BAU)andintheEcosustainable(ECO)scenarios,andcumulativeenergysavingsbetweenthetwoscenarios[16] Figure1.3:OperatingExpenses(OPEX)estimationrelatedtoenergycostsfortheEuropeantelcos’networki nfrastructuresinthe”Business-As-Usual”(BAU)andintheEcosustainable(ECO)scenarios,andcumulativesavingsbetweenthetwo scenarios[17]8 Figure1.4:SDN Architecture .11 Figure1.5:OpenFlowcontroller andswitches 12 Figure2.1:DCNArchitecture[43] 18 Figure2.2:Three-tierDCN Architecture[45] .18 Figure 2.3:Fat-treeDCN Topology .19 Figure2.4:DcellDCNArchitecture[53] .19 Figure2.5:BCubeDCNArchitecture[54] .20 Figure2.6:Fat-treearchitecture withk =4 21 Figure2.7:Diagramof theElasticTreesystem[57] 22 Figure2.8:Energy–Utilizationrelationofanetwork[58] .23 Figure2.9:Power-controlSystem of aNetwork 26 Figure 2.10:Fat-treetopology with MinimumSpanning Tree 28 Figure 2.11:Power Scaling Algorithm 32 Figure2 : P o w e r S c a l i n g a n d E n e r g y - P r o f i l e - A w a r e P S n E P a l g o r i t h m ( P r o p o s e d Algorithm 1).The flowchartdescribesthe processbetweenEdgeandAggregationswitches .3 Figure2.13:use-casewith PSnEP algorithm in aDCN 35 Figure2.14:PSnEPvsPower scaling (PS)with k=6 Fat-tree,mixscenario 38 Figure 2.15: Energy-saving level ratio of the PSnEP algorithm to the PS algorithmindifferentsizes 39 Figure2.16:Extended Power-Controlsystem(Ext-PCS) 40 Figure2.17:Example 42 Figure2.18:First-fitMigration[67]Algorithm .42