This book addresses the urgent issue of massive and inefficient energy consumption by data centers, which have become the largest co-located computing systems in the world and process trillions of megabytes of data every second. Dynamic provisioning algorithms have the potential to be the most viable and convenient of approaches to reducing data center energy consumption by turning off unnecessary servers, but they incur additional costs from being unable to properly predict future workload demands that have only recently been mitigated by advances in machine-learned predictions. This book explores whether it is possible to design effective online dynamic provisioning algorithms that require zero future workload information while still achieving close-to-optimal performance. It also examines whether characterizing the benefits of utilizing the future workload information can then improve the design of online algorithms with predictions in dynamic provisioning. The book specifically develops online dynamic provisioning algorithms with and without the available future workload information. Readers will discover the elegant structure of the online dynamic provisioning problem in a way that reveals the optimal solution through divide-and-conquer tactics. The book teaches readers to exploit this insight by showing the design of two online competitive algorithms with competitive ratios characterized by the normalized size of a look-ahead window in which exact workload prediction is available.
With the increasing popularity of cloud computing service, IT companies are building and expanding their datacenters nationwide or worldwide. With hundreds of thousands of servers, a commercial datacenter consumes many mega-watts of power annually, and imposes significant electricity costs to its operator. In this dissertation, we focus on reducing energy cost for Internet-scale datacenters (IDCs) to promote more efficient datacenter operation. To reduce IDC energy consumption, load-aware capacity provisioning schemes have been considered where capacity can be reduced (by turning off servers or scheduling a low CPU frequency) when its load is lower. To reduce IDC energy cost, electricity price-aware load shifting is considered, where IDCs with a lower price/cost serve more load. Our work is related to the load-aware capacity provisioning and electricity price-aware load shifting. However, different from most existing work, we consider practical traffic and load models. We first consider traffic dynamics at each IDC. We consider both large time scale and small time scale traffic variation. We further explicitly differentiate the load demand. We consider both delay sensitive jobs which have strict quality of service requirements, and delay tolerant jobs. We propose joint capacity allocation and load shifting schemes for distributed IDCs with dynamic and differentiated load demand. Our work is summarized as follows. In datacenters, traffic demand varies both in large and small time scales. A datacenter with dynamic traffic often needs toover-provision active servers to meet the peak load demand, which incurs significant energy cost. In the dissertation, our first goal is to achieve an optimal tradeoff between energy efficiency and service performance over a set of distributed IDCs with dynamic demand. In particular, we consider the overload probability as the QoS metric, where overload is defined as service demand exceeding the capacity of an IDC. We require the overload probability at each IDC to be smaller than some predefined threshold. Our goal is thus to minimize total energy cost over all IDCs, subject to the overload constraint. We achieve the goal by dynamically adjusting server capacity and performing load shifting in different time scales. We propose three different load-shifting and joint capacity allocation schemes with different complexity and performance. Our load shifting schemes leverage electricity-price diversity. Meanwhile, they reduce the traffic burstiness such that a smaller capacity is required to satisfy the overload probability constraint. Thus, our schemes reduce energy consumption/cost even when all IDCs have the same electricity price. We use both simulated load traces and real traffic traces to evaluate the performance of the proposed schemes. Our results show that our proposed schemes are efficient in reducing energy cost, and robust in QoS provisioning. Our first work uses load-aware capacity provisioning schemes, where servers are turned on/off according to the load. Similar to most existing work, we only consider instantaneous (Internet) requests, which are explicitly or implicitly assumed to be delay-sensitive. On the other hand, in datacenters, there are many delay-tolerant jobs, such as background/maintainance jobs. In our second work, we explicitly differentiate delay-sensitive jobs and delay tolerant jobs. We focus on the problem of using delay-tolerant jobs to fill the extra capacity of datacenters, referred to as trough/valley filling. Giving a higher priority to delay-sensitive jobs, our schemes improve most existing demand-proportional resource provisioning schemes. Our goal is then to design intelligent trough filling mechanisms that are energy efficient and also achieve good delay performance. Specifically, we propose two joint dynamic speed scaling and traffic shifting schemes, one subgradient-based and the other queue-based. Our schemes assume little statistical information about the system, which is usually difficult to obtain in practice. In both schemes, energy cost saving comes from dynamic speed scaling, statistical multiplexing, electricity price diversity, and service efficiency diversity. In addition, good delay performance is achieved in the queue-based scheme via load shifting and capacity allocation based on queue conditions. Practical issues that may arise in datacenter networks are considered, including capacity and bandwidth constraints, service agility constraints, and load shifting cost. We use both simulated and real datacenter traces to evaluate the proposed schemes. Last, we also propose joint capacity provisioning and load shifting schemes for mixed traffic with delay sensitive and delay tolerant jobs. In these schemes, we introduce a delay constraint as the QoS requirement for delay sensitive jobs. We employ a priority queue to derive the capacity demand for delay sensitive jobs under given delay constraints, and further consider bandwidth and capacity sharing for delay tolerant jobs. We follow similar approaches in the second work to reduce energy cost by delay sensitive jobs and delay tolerant jobs, while guaranteeing the delay requirement for delay sensitive jobs and good delay performance for delay tolerant jobs.
Recent increase in energy prices has led researchers to find better ways for capacity provisioning in data centers to reduce energy wastage due to the variation in workload. This paper explores the opportunity for cost saving and proposes a novel approach for capacity provisioning under bounded latency requirements for the workload. We investigate how many servers to be kept active and how much workload to be delayed for energy saving while meeting every deadline. We present an offline LP formulation for capacity provisioning by dynamic deferral and give two online algorithms to determine the capacity of the data center and the assignment of workload to servers dynamically. We prove the feasibility of the online algorithms and show that their worst case performance are bounded by a constant factor with respect to the offline formulation. We validate our algorithms on synthetic workload generated from two real HTTP traces and show that they actually perform much better in practice than the worst case, resulting in 20-40% cost-savings.
Data centers consume roughly 1% of the total electricity demand, while ICT as a whole consumes around 10%. Demand is growing exponentially and, left unchecked, will grow to an estimated increase of 20% or more by 2030. This book covers the energy consumption and minimization of the different data center components when running real workloads, taking into account the types of instructions executed by the servers. It presents the different air- and liquid-cooled technologies for servers and data centers with some real examples, including waste heat reuse through adsorption chillers, as well as the hardware and software used to measure, model and control energy. It computes and compares the Power Usage Effectiveness and the Total Cost of Ownership of new and existing data centers with different cooling designs, including free cooling and waste heat reuse leading to the Energy Reuse Effectiveness. The book concludes by demonstrating how a well-designed data center reusing waste heat to produce chilled water can reduce energy consumption by roughly 50%, and how renewable energy can be used to create net-zero energy data centers.
This book constitutes the thoroughly refereed post-conference proceedings of the Second International Workshop on Energy Efficient Data Centers, E2DC 2013, held in Berkeley, CA, USA, in May 2013; co-located with SIGCOMM e-Energy 2013. The 8 revised full papers presented were carefully reviewed and selected from numerous submissions. The papers are organized in topical sections on energy and workload measurement; energy management; simulators and control.
This book highlights recent advances in Cybernetics, Machine Learning and Cognitive Science applied to Communications Engineering and Technologies, and presents high-quality research conducted by experts in this area. It provides a valuable reference guide for students, researchers and industry practitioners who want to keep abreast of the latest developments in this dynamic, exciting and interesting research field of communication engineering, driven by next-generation IT-enabled techniques. The book will also benefit practitioners whose work involves the development of communication systems using advanced cybernetics, data processing, swarm intelligence and cyber-physical systems; applied mathematicians; and developers of embedded and real-time systems. Moreover, it shares insights into applying concepts from Machine Learning, Cognitive Science, Cybernetics and other areas of artificial intelligence to wireless and mobile systems, control systems and biomedical engineering.
This book constitutes the thoroughly refereed post-conference proceedings of the First International Workshop on Energy Efficient Data Centers (E2DC 2012) held in Madrid, Spain, in May 2012. The 13 revised full papers presented were carefully selected from 32 submissions. The papers cover topics from information and communication technologies of green data centers to business models and GreenSLA solutions. The first section presents contributions in form of position and short papers, related to various European projects. The other two sections comprise papers with more in-depth technical details. The topics covered include energy-efficient data center management and service delivery as well as energy monitoring and optimization techniques for data centers.
This book investigates the coordinated power management of multi-tenant data centers that account for a large portion of the data center industry. The authors include discussion of their quick growth and their electricity consumption, which has huge economic and environmental impacts. This book covers the various coordinated management solutions in the existing literature focusing on efficiency, sustainability, and demand response aspects. First, the authors provide a background on the multi-tenant data center covering the stake holders, components, power infrastructure, and energy usage. Then, each power management mechanism is described in terms of motivation, problem formulation, challenges and solution.
Surge growth of numerous cloud services, Internet of Things, and edge computing promotes continuous increasing demand for data centers worldwide. Significant electricity consumption of data centers has tremendous implications on both operating and capital expense. The power infrastructure, along with the cooling system cost a multi-million or even billion dollar project to add new data center capacities. Given the high cost of large-scale data centers, it is important to fully utilize the capacity of data centers to reduce the Total Cost of Ownership. The data center is designed with a space budget and power budget. With the adoption of high-density rack designs, the capacity of a modern data center is usually limited by the power budget. So the core of the challenge is scaling up power infrastructure capacity. However, resizing the initial power capacity for an existing data center can be a task as difficult as building a new data center because of a non-scalable centralized power provisioning scheme. Thus, how to maximize the power utilization and optimize the performance per power budget is critical for data centers to deliver enough computation ability. To explore and attack the challenges of improving the power utilization, we have planned to work on different levels of data center, including server level, row level, and data center level. For server level, we take advantage of modern hardware to maximize power efficiency of each server. For rack level, we propose Pelican, a new power scheduling system for large-scale data centers with heterogeneous workloads. For row level, we present Ampere, a new approach to improve throughput per watt by provisioning extra servers. By combining these studies on different levels, we will provide comprehensive energy efficient system designs for data center.
Energy systems worldwide are undergoing major transformation as a consequence of the transition towards the widespread use of clean and sustainable energy sources. Basically, this involves massive changes in technical and organizational levels together with tremendous technological upgrades in different sectors ranging from energy generation and transmission systems down to distribution systems. These actions generate huge science and engineering challenges and demands for expert knowledge in the field to create solutions for a sustainable energy system that is economically, environmentally, and socially viable while meeting high security requirements. This book covers these promising and dynamic areas of research and development, and presents contributions in sustainable energy systems planning, integration, and management. Moreover, the book elaborates on a variety of topics, ranging from design and planning of small- to large-scale energy systems to the operation and control of energy networks in different sectors, namely electricity, heat, and transport.
