Management of Data Center Networks

Nadjib Aitsaadi, PhD, is a Full Professor in Networks and Telecommunications at UVSQ Paris-Saclay University in France. He is a member of DAVID Laboratory and leads the Next Generation Networks Team. Prof. Aitsaadi earned a PhD in 2010 from Sorbonne University in Networks & Telecommunications and graduated in 2016 with a “Habilitation” diploma from University Paris Est (UPE). His main research fields are the security and QoS optimization of cellular networks (5G, 6G, HAP), IoT, DCN, V2X, MEC, NFV/SDN, and more. The results have been published in many major journals such as IEEE JSAC, IEEE TVT, Elsevier ComNet, Elsevier ComCom, etc. and major conferences such IEEE SECON, IEEE LCN, IEEE MASS, ACM MSWiM, IEEE/IFIP NOMS/IM, IEEE ICC, IEEE GLOBECOM, etc. Prof. Aitsaadi chairs many tracks in IEEE/IFIP conferences such as IEEE GLOBECOM, IEEE/IFIP IM, IEEE/IFIP CIoT, etc. and he is very active in IEEE Technical Committees TCIIN.

About the Editor xi

Contributors xiii

Acronyms xv

Introduction xvii

1 Architectures of Data Center Networks: Overview 1
Boutheina Dab, Ilhem Fajjari, Dallal Belabed, and Nadjib Aitsaadi

1.1 Taxonomy of DCN Architectures 1

1.1.1 Classification of DCN Architectures 2

1.1.2 Switch-Centric DCN Architectures Overview 3

1.1.2.1 Tree-Based DCN 3

1.1.2.2 Hierarchical DCN Architecture 4

1.1.2.3 Flat DCN Architecture 6

1.1.3 Server-Centric DCN Architectures Overview 7

1.1.4 Enhanced DCN Architectures Overview 10

1.1.4.1 Optical DCN Architecture 10

1.1.4.2 Wireless DCN Architecture 12

1.2 Comparison Between DCN Architectures 15

1.3 Proposed HDCN Architecture 15

1.3.1 HDCN Architecture Based on MSDC Model 19

1.3.1.1 ECMP Protocol 19

1.3.2 60GHz Technology in HDCN 20

1.3.3 Beamforming Technique in HDCN 21

1.4 Conclusion 23

References 23

2 Data Center Optimization Techniques 29
Dallal Belabed

2.1 Ethernet Switching and Routing 29

2.2 Data Center Optimization Techniques 38

2.2.1 Virtual Network Embedding 38

2.2.2 Server Consolidation 40

2.2.3 Traffic Engineering 43

2.2.3.1 Link-State Traffic Engineering 44

2.2.3.2 MPLS Traffic Engineering 44

2.2.3.3 TCP Proportional Fairness Model 46

2.3 Conclusion 49

Bibliography 51

3 Resource Management in Hybrid (Wired/Wireless) Data Center Networks 57
Boutheina Dab, Ilhem Fajjari, and Nadjib Aitsaadi

3.1 Routing and Wireless Channel Allocation Problematic in HDCN 58

3.1.1 Routing and Wireless Channel Assignment Challenges in HDCN 59

3.1.2 Routing and Wireless Channel Assignment Criteria in HDCN 61

3.2 Wireless Channel Allocation Strategies for One-Hop Communications in HDCN 62

3.2.1 Channel Allocation Problem in Wireless Networks 63

3.2.2 Omni-Directional Antennas Based Strategies 63

3.2.3 Beamforming-Based Strategies 67

3.3 Online Joint Routing and Wireless Channel Allocation Strategies in HDCN 69

3.3.1 Joint Routing and Channel Assignment in Mesh Networks 70

3.3.2 Online Joint Routing and Channel Assignment Strategies in HDCN 71

3.4 Joint Batch Routing and Channel Allocation Strategies in HDCN 75

3.5 Joint Batch Routing and Channel Allocation Strategies in HDCN 75

3.6 Summary 77

3.7 Conclusion 80

References 80

4 Inter-Data Center Networks: Routing and Reliability in Virtual Network Backbone 85
Oussama Soualah, Ilhem Fajjari, and Nadjib Aitsaadi

4.1 Overview of Basic Virtual Network Embedding Without Reliability Constraint 85

4.1.1 Online Approaches 86

4.1.2 Batch Approaches 87

4.2 Overview of Virtual Network Embedding with Reliability Constraint 89

4.2.1 Distributed Approaches 89

4.2.2 Centralized Approaches 91

4.2.2.1 Substrate Router Failures 91

4.2.2.2 Substrate Link Failures 92

4.2.2.3 Substrate Router and Link Failures 94

4.2.2.4 Regional Failures 95

4.2.3 Summary 101

4.3 Conclusion 101

References 101

5 An Evaluation Method of Optimal Cost Saving in a Data Center with Proactive Management 105
Ruben Milocco, Pascale Minet, Éric Renault, and Selma Boumerdassi

5.1 Introduction 106

5.2 RelatedWork 108

5.3 Framework for DC Modeling 111

5.3.1 Notations and Assumptions 111

5.3.2 Energy Computation 111

5.3.2.1 Single-Resource Case 111

5.3.2.2 Extension to the Multi-resource Case 114

5.4 Cost Formulation 114

5.4.1 Example 115

5.4.2 Methodology 116

5.4.3 Relative Energy Cost Saving 116

5.4.4 Upper Bound Computation 118

5.5 Application to a Real DC 118

5.5.1 Generalities 119

5.5.1.1 Selection of the Sampling Interval 119

5.5.1.2 Selection of Possible Values for the Costs 119

5.5.1.3 Dynamic Capacity Provisioning Based on Energy Prediction 119

5.5.2 Application to a Google Dataset 120

5.5.2.1 Energy Computation 120

5.5.2.2 Evaluation of the Upper Bound 122

5.5.2.3 Computation of the Relative Energy Cost Saving 123

5.5.2.4 Discussion of Results 124

5.6 Conclusion 124

References 125

Index 129