A Survey on Coordinated Power Management in Multi-Tenant Data Centers (eBook)

Thant Zin Oo received the B.Eng. degree in electrical systems and electronics at Myanmar Maritime University, Thanlyin, Myanmar in 2008 and the B.S. degree in computing and information system from London Metropolitan University, U.K., in 2008, for which he received grant from the British Council. He is currently working towards Ph.D. degree in computer science and engineering from Kyung Hee University, Korea, for which he was awarded a scholarship in 2010. His research interests include wireless communications, network virtualization, data centers, and sustainable energy. Nguyen H. Tran received the BS degree from Hochiminh City University of Technology and Ph.D. degree from Kyung Hee University, in electrical and computer engineering, in 2005 and 2011, respectively. Since 2012, he has been an Assistant Professor with Department of Computer Science and Engineering, Kyung Hee University. His research interest is to applying analytic techniques of optimization, game theory, and stochastic modeling to cutting-edge applications such as cloud and mobile-edge computing, data centers, heterogeneous wireless networks, and big data for networks. He received the best KHU thesis award in engineering in 2011 and best paper award at IEEE ICC 2016. He is the Editor of IEEE Transactions on Green Communications and Networking.  Shaolei Ren is an Assistant Professor of Electrical and Computer Engineering at University of California, Riverside. He received his B.E. from Tsinghua University in 2006, M.Phil. from Hong Kong University of Science and Technology in 2008, and Ph.D. from University of California, Los Angeles in 2012, all in electrical and computer engineering. His research interests include cloud computing, data centers, and network economics. He was a recipient of the U.S. NSF Faculty Early Career Development (CAREER) Award in 2015. He also received best paper awards from several conferences, including ACM e-Energy'16, IEEE ICC'16 and IEEE ICC'09. Choong Seon Hong received the B.S. and M.S. degrees in electronic engineering from Kyung Hee University, Seoul, South Korea, in 1983 and 1985, respectively, and the Ph.D. degree from Keio University, Minato, Japan, in 1997. In 1988, he joined Korea Telecom, where he worked on broadband networks as a Member of Technical Staff. In September 1993, he joined Keio University. He worked for the Telecommunications Network Laboratory, Korea Telecom, as a Senior Member of Technical Staff and the Director of the Networking Research Team until August 1999. Since September 1999, he has been a Professor with the Department of Computer Science and Engineering, Kyung Hee University. His research interests include future Internet, ad hoc networks, network management, and network security. He is a member of ACM, IEICE, IPSJ, KIISE, KICS, KIPS, and OSIA. He has served as the General Chair, a TPC Chair/Member, or an Organizing Committee Member for international conferences such as NOMS, IM, APNOMS, E2EMON, CCNC, ADSN, ICPP, DIM, WISA, BcN, TINA, SAINT, and ICOIN. In addition, he is currently an Associate Editor of the IEEE Transactions on Network and Service Management, International Journal of Network Management, and Journal of Communications and Networks and an Associate Technical Editor of the IEEE Communications Magazine.

Preface 5
Acknowledgement 6
Contents 7
Part I Introduction 11
1 Overview 12
1.1 Importance of Multi-Tenant Data Centers 14
1.1.1 Operator's Perspective 14
1.1.2 Tenants' Perspective 14
1.2 State-of-the-Art Research on Data Centers 16
1.3 Potential of Coordinated Power Management 17
1.3.1 Importance of Coordinated Power Management 17
1.3.2 Coordinated Power Management in Multi-Tenant Data Centers 18
1.4 Research Directions for Multi-Tenant Data Centers 18
1.5 Sustainable Multi-Tenant Data Centers 18
1.6 Multi-Tenant Data Center Demand Response 19
1.6.1 What is Data Center Demand Response? 19
1.6.2 Why Multi-Tenant Data Center Demand Response? 20
2 Preliminaries 21
2.1 Multi-Tenant Data Center 21
2.2 Electrical Systems 22
2.2.1 Power Usage Effectiveness 23
2.2.2 Electricity Supply 23
2.2.3 Electricity Demand 24
2.2.4 Electricity Bill 24
2.3 Carbon Footprint 25
2.4 Inconvenience to Tenants 27
2.4.1 Delay Performance Cost 27
2.4.2 Other Costs 27
Part II Sustainable Multi-Tenant Data Center 29
3 Background 30
3.1 Motivation 30
3.2 Issues 31
3.3 Challenges 31
3.4 Uncertainty 32
3.5 On-line Coordination 32
4 System Model 33
4.1 Problem Formulation 33
4.1.1 Minimizing the Operating Cost 33
4.1.2 Minimizing the Energy Consumption 34
5 Solutions 36
5.1 Reducing Cost via Rewards 37
5.1.1 Feedback-Based On-Line Optimization 38
5.1.2 Simulation and Results 39
5.1.2.1 Simulation Settings 39
5.1.2.2 Tenants' Response 39
5.1.2.3 Simulation Results 40
5.1.3 Experiment 44
5.1.3.1 Colocation Testbed 44
5.1.3.2 Tenants' Response 45
5.1.3.3 Experimental Results 47
5.2 Minimizing Carbon Footprint in Colocation Data Center (GreenColo) 48
5.2.1 Simulation and Results 50
5.2.1.1 Simulation Settings 50
5.2.1.2 Execution of GreenColo 52
5.2.1.3 Tenant Costs 52
5.2.1.4 Carbon Footprint Reduction 54
5.3 Randomization for Pricing and Auction 55
5.3.1 Randomized Pricing Approach 55
5.3.1.1 An Off-Line Approximation Algorithm 55
5.3.1.2 An On-Line Algorithm 56
5.3.2 Randomized Auction Approach 57
5.3.2.1 An On-Line Algorithm 57
5.3.2.2 A More Intelligent On-Line Algorithm 59
5.3.3 Randomized Truthful Auction Mechanism 59
5.3.4 Simulation and Results 63
5.3.4.1 Simulation Setup 63
5.3.4.2 Algorithms for Pricing Approach 64
5.3.4.3 Algorithms for Auction Approach 67
6 Summary 71
Part III Multi-Tenant Data Center Demand Response 72
7 Background 73
7.1 Motivation 73
7.2 Issues 74
7.3 Challenges 75
8 System Model 77
8.1 Problem Formulations 77
8.1.1 Maximizing the Total Energy Demand Reduction 77
8.1.2 Minimizing the Social Cost of a Multi-Tenant Data Center 78
8.1.2.1 Mixed Integer Programming Problem Chen2015PER, Zhang2015INFOCOM 78
8.1.2.2 Linear Programming Problem Guo2015 78
8.1.3 Maximizing the Social Welfare 79
8.1.4 Maximizing the Social Cost Savings 80
8.1.5 Thermal-Aware Minimization via Backup Energy Storage 81
8.1.5.1 Temperature and Heat Recirculation Model 81
8.1.5.2 Auction Model 82
8.1.6 Contract Design Formulation 83
8.1.7 Stackelberg Game Formulation 84
8.1.8 Minimizing Social Cost for Geo-Distributed Multi-Tenant Data Centers 85
9 Solutions 87
9.1 Incentivizing Colocation Tenants for Demand Response (iCODE) 87
9.1.1 Simulations and Results 89
9.1.1.1 Simulation Setup 89
9.1.1.2 Comparison Between iCODE and NDR 90
9.1.1.3 Impact of Workload Over-Prediction 91
9.1.1.4 Impact of Greediness 92
9.2 Truthful Incentive Mechanism (Truth-DR) 93
9.2.1 2-Approximation Algorithm 94
9.2.2 The Randomized Auction 96
(1) Optimal Fractional Solution 97
(2) Convex Decomposition 97
(3) Winner Determination and Payment 98
9.2.3 Simulations and Results 99
9.2.3.1 Simulation Setup 99
9.2.3.2 Close-to-Minimum Social Cost 100
9.2.3.3 Satisfying Energy Reduction Target 101
9.2.3.4 Tenants' Non-Negative Utilities 101
9.2.3.5 Social Cost Reduction Compared to ``Backup Energy Storage Only'' 102
9.3 Greening Multi-Tenant Data Center Demand Response (ColoEDR) 103
9.3.1 Price-Taking Tenants 105
9.3.2 Price-Anticipating Tenants 106
9.4 Fair Rewarding in EDR (FairDR) 108
9.4.1 Auction Algorithm (FairDR) 108
9.4.1.1 Eligible Tenant Initialization 108
9.4.1.2 Tenant Rewarding 109
9.4.1.3 Winner Determination and Energy Reduction Allocation 109
9.4.2 Theoretical Analysis 111
9.4.3 Simulations and Results 112
9.4.3.1 Simulation Setups 112
9.4.3.2 Fairness in rewarding 113
9.4.3.3 Social Cost and Social Cost Savings 114
9.4.3.4 Competitive Ratio in Social Cost Savings 115
9.5 Thermal-Aware Cost Efficient Mechanism for EDR (TECH) 115
9.5.1 Reverse Auction Mechanism (TECH) 117
9.5.1.1 Winner Selection 118
9.5.1.2 BES Energy Calculation 119
9.5.1.3 Payment Determination 119
9.5.2 TECH-EH 120
9.5.2.1 Winner Selection 120
9.5.2.2 Payment Determination 121
9.5.3 Simulations and Results 122
9.5.3.1 Simulation Setup 123
9.5.3.2 Energy Reduction 124
9.5.3.3 Cost Efficiency 125
9.5.3.4 Temperature of Supplied Cooling Air 125
9.5.3.5 Tenant's Energy Reduction 126
9.5.3.6 Number of Chosen Servers 127
9.6 Nash Bargaining Solution 127
9.6.1 One-to-One Bargaining 130
9.6.2 Concurrent Bargaining 132
9.7 Contract Design Approach (Contract-DR) 134
9.7.1 Contract Design with Complete Information 134
9.7.2 Contract Design with Incomplete Information 135
9.7.3 Feasibility and Optimality Proofs 136
9.7.4 Simulations and Results 136
9.7.4.1 Simulation Settings 136
9.7.4.2 Comparison of Different Contract Designs 137
9.7.4.3 Comparison with Non-demand Response Approach 140
9.7.4.4 Impact of Energy Storage Device Cost 142
9.7.4.5 Workload Over-Prediction of Tenants 143
9.8 Stackelberg Game Approach 144
9.8.1 Stackelberg Equilibrium: Analysis and Algorithm 144
9.8.2 Simulations and Results 147
9.8.2.1 Simulation Settings 147
9.8.2.2 Impact of Utility Functions 147
9.8.2.3 Impact of ?3 150
9.9 Alternating Direction Method of Multipliers 150
9.9.1 Case Study 153
9.10 EDR in Geo-Distributed Multi-Tenant Data Centers (BatchEDR) 154
9.10.1 On-Line Algorithm Framework 154
9.10.2 Auction Mechanism 157
10 Summary 159
11 Concluding Remarks 161
11.1 Open Issues 161
11.1.1 Making the Multi-Tenant Data Center Efficient and Green 161
11.1.2 Multi-Tenant Data Center Demand Response 162
11.2 Conclusion 163
References 164
Index 176