Search

The dramatic increasing of data and demands for computing capabilities may result in excessive use of resources in cloud data centers, which not only causes the raising of energy consumption, but also leads to the violation of Service Level Agreement (SLA). Dynamic consolidation of virtual machines (VMs) is proven to be an efficient way to tackle this issue. In this paper, we present an Adaptive Deep Reinforcement Learning (DRL)-based Virtual Machine Consolidation (ADVMC) framework for energy-efficient cloud data centers. ADVMC has two phases. In the first phase, Influence Coefficient is introduced to measure the impact of a VM on producing host overload, and a dynamic Influence Coefficient-based VM selection algorithm (ICVMS) is proposed to preferentially choose those VMs with the greatest impact for migration in order to remove the excessive workloads of the overloaded host quickly and accurately. In the second phase, a Prediction Aware DRL-based VM placement method (PADRL) is further proposed to automatically find suitable hosts for VMs to be migrated, in which a state prediction network is designed based on LSTM to provide DRL-based model more reasonable environment states so as to accelerate the convergence of DRL. Simulation experiments on the real-world workload provided by Google Cluster Trace have shown that our ADVMC approach can largely cut down system energy consumption and reduce SLA violation of users as compared to many other VM consolidation policies. [ABSTRACT FROM AUTHOR]

A growing number of organizations are hosting their software applications in distributed data centers (DCs) in the cloud, for faster response time and higher energy efficiency. The dramatic increase of user tasks, however, poses a significant challenge on DC providers to retain users’ expectations on both aspects. To tackle this challenge, this work first formulates the problem into a constrained biobjective optimization problem. A biobjective algorithm, named simulated-annealing-based adaptive differential evolution (SADE), is presented to simultaneously reduce both the response time of tasks and energy cost. Meanwhile, a method of minimal Manhattan distance is adopted to search for a final knee, for achieving a good balance between response time minimization and energy cost reduction. Experimental results on real-life datasets, i.e., the electricity prices and tasks collected from a Google cluster trace, have proved that SADE yields less task response time and lower energy cost compared with state-of-the-art algorithms. [ABSTRACT FROM AUTHOR]

A disaster may not necessarily demolish the telecommunications infrastructure, but instead it might affect the national grid and cause blackouts, consequently disrupting the network operation unless there is an alternative power source(s). In disaster-resilient networks, fiber cut, datacenter destruction, and node isolation have been studied before with different scenarios, but the power outage impact has not been investigated before. In this paper, power outages are considered, and the telecommunication network performance is evaluated during a blackout. A mixed Integer Linear Programming (MILP) model is developed to evaluate the network performance for a single node blackout under two scenarios: minimization of blocking and minimization of renewable and battery energy consumption. Insights analyzed from the MILP model results have demonstrated the trade-off between the two evaluated optimization cost functions and shown that the proposed scheme can extend the network lifetime while minimizing the required amount of backup energy. [ABSTRACT FROM AUTHOR]

Given recent algorithm, software, and hardware innovation, computing has enabled a plethora of new applications. As computing becomes increasingly ubiquitous, however, so does its environmental impact. This article brings the issue to the attention of computer-systems researchers. Our analysis, built on industry-reported characterization, quantifies the environmental effects of computing in terms of carbon emissions. Broadly, carbon emissions have two sources: operational energy consumption, and hardware manufacturing and infrastructure. Although carbon emissions from the former are decreasing, thanks to algorithmic, software, and hardware innovations that boost performance and power efficiency, the overall carbon footprint of computer systems continues to grow. This work quantifies the carbon output of computer systems to show that most emissions related to modern mobile and data-center equipment come from hardware manufacturing and infrastructure. We, therefore, outline future directions for minimizing the environmental impact of computing systems. [ABSTRACT FROM AUTHOR]

Mobile edge computing (MEC) is an enabling technology for low-latency AI applications, by caching AI services originally deployed in remote data centers to 5G base stations in network edge. Due to limited computing resource of 5G base stations, not all services can be cached in base stations to meet the resource demands of user requests. Also, if the workload of a 5G base station reaches to its resource capacity, the energy consumption of the base station will be pushed up exponentially. To reduce the energy consumption and overcome resource limitations on base stations, an alternative is to allow the base stations to collaborate with each other to admit user requests. In this paper, we investigate the problem of collaborative service caching and request offloading between a 5G-enabled MEC and remote data centers, while meeting the quality of service (QoS) requirements of users, and resource capacities on base stations that are operated by multiple selfish network service providers. We aim to maximize the total payoff of all base stations. To this end, we first propose a two-stage optimization framework: In the first stage, we develop a mechanism that adopts a best-reply rule for dynamically distributed coalition formation. In the second stage, we propose a near-optimal payoff allocation method by devising a randomized algorithm with a provable approximation ratio. We then evaluate the performance of the proposed optimization framework by extensive experimental simulations. Simulation results show that the proposed framework outperforms its counterparts by achieving at least 30% higher payoff and 20% lower energy consumption of base stations. [ABSTRACT FROM AUTHOR]

The data center market grows rapidly with the increase of data and its corresponding applications (e.g., machine learning, cloud storage, Internet of Things, and so on). The growth is boosted recently due to the shift of activities online during the COVID-19 pandemic. Reducing the energy consumption of data centers faces various challenges that are further aggravated by the tropical conditions with high temperature and humidity in the tropics like Singapore. The prevailing siloed approach of operating the information technology (IT) and the facility systems separately has resulted in wasteful over-provisioning. The recently proposed approaches for energy usage minimization under various constraints including thermal safety scale poorly with the data center size and often result in non-optimal solutions. To advance the state of the art, we apply deep reinforcement learning (DRL) to address the scalability problem and achieve optimality over a long time horizon in reducing data center energy usage. In particular, we deploy the data-driven deep model and physical rule based model in lieu of the physical data center during the training and validation phases to manage the thermal safety risks caused by DRL's strategy of learning from errors. [ABSTRACT FROM AUTHOR]

With the explosive growth in the scale of cloud computing infrastructures, reliability and energy efficiency have become important concerns considering the great complexity of cloud data centers. There is an urgent need for efficient task assignment that can dispatch tasks to appropriate cloud data center servers, which is critical to achieve reliability and energy efficiency in current cloud data centers. Most of the research on task assignment focuses on only one of the objectives of reliability and energy efficiency, while the two objectives are intrinsically conflicting with each other. In this paper, we deal with the problem of task assignment in data centers, with the objective of minimizing the energy consumption while providing failure tolerance to task execution failure. We propose a reliability-aware and energy-efficient task replica assignment algorithm based on running task replicas at a low speed and enabling multiple task replicas to share the same server resources. Each task in a job processed by the cloud computing platform has two instances: main task and task replica (shadow). Each main task runs on an individual server, and the task replica associated with the main task is assigned on a different server. The main tasks run at the full server speed, while the task replicas run at a lower rate than the main tasks. The task replicas can be mapped onto dedicated backup servers or be assigned to the servers on which the main tasks are running. Multiple task replicas can share the same server resources to reduce the number of servers required. We conduct experiments through simulations. Experimental results demonstrate that the proposed algorithm can effectively reduce the energy consumption, while achieving a good balance between the number of servers used and job completion time. [ABSTRACT FROM AUTHOR]

The frequent data movement between the processor and the memory has become a severe performance bottleneck for deep neural network (DNN) training workloads in data centers. To solve this off-chip memory access challenge, the 3-D stacking processing-in-memory (3D-PIM) architecture provides a viable solution. However, existing 3D-PIM designs for DNN training suffer from the limited memory bandwidth in the base logic die. To overcome this obstacle, integrating the DNN related logic near each memory bank becomes a promising yet challenging solution, since naively implementing the floating-point (FP) unit and the cache in the memory die incurs a large area overhead. To address these problems, we propose DLUX, a high performance and energy-efficient 3D-PIM accelerator for DNN training using the near-bank architecture. From the hardware perspective, to support the FP multiplier with low area overhead, an in-DRAM lookup table (LUT) mechanism is invented. Then, we propose to use a small scratchpad buffer together with a lightweight transformation engine to exploit the locality and enable flexible data layout without the expensive cache. From the software aspect, we split the mapping/scheduling tasks during DNN training into intralayer and interlayer phases. During the intralayer phase, to maximize data reuse in the LUT buffer and the scratchpad buffer, achieve high concurrency, and reduce data movement among banks, a 3D-PIM customized loop tiling technique is adopted. During the interlayer phase, efficient techniques are invented to ensure the input–output data layout consistency and realize the forward–backward layout transposition. Experiment results show that DLUX can reduce FP32 multiplier area overhead by 60% against the direct implementation. Compared with a Tesla V100 GPU, end-to-end evaluations show that DLUX can provide on average $6.3\times $ speedup and $42\times $ energy efficiency improvement. [ABSTRACT FROM AUTHOR]

With the development of the Internet of Things, the data that needs to be processed is increasing rapidly. Therefore, the collaboration of cloud and edge emerges as the times require. Edge nodes are mainly responsible for collecting data, and decide to process the data locally or offload to cloud data centers. Cloud data centers are suitable for data analysis, model training, and managing edge nodes. In this article, we focus on the task assignment problems in collaborative edge and cloud environments and study it in a distributed, non-cooperative environment. An M/M/1 queueing model is established to characterize the task transmission. Because of the multi-core processors, we set an M/M/C queueing model to characterize the task computation. We consider the problem from the perspective of game theory and formulate it into a non-cooperative game among multi-agents (multiple edge data centers) in which each agent is informed with incomplete information (allocation strategies) of others. For each agent, we define a function of the expected cost of tasks as the disutility function, and minimize it subject to the QoS constraint. We analyze the existence of Nash equilibrium and develop a Greedy Energy-aware Algorithm (GEA) to choose active servers using the Limit Searching Algorithm (LSA) to find the ceiling utilization. Then we propose the Best Response Algorithm (BRA) to optimize the utility function. The convergence of the BRA algorithm has been discussed. Finally, the results demonstrate that the BRA algorithm can get a solution close to Nash equilibrium and reach it quickly. [ABSTRACT FROM AUTHOR]

The significant growth in the number and types of tasks of heterogeneous applications in green cloud data centers (GCDCs) dramatically increases their providers’ revenue from users as well as energy consumption. It is a big challenge to maximize such revenue, while minimizing energy cost in a market where prices of electricity, availability of renewable power generation, and behind-the-meter renewable generation contract models differ among the geographical sites of the GCDCs. A multiobjective optimization method that investigates such spatial differences in the GCDCs is for the first time proposed to trade off such two objectives by cost-effectively executing all tasks while meeting their delay constraints. In each time slot, a constrained biobjective optimization problem is formulated and solved by an improved multiobjective evolutionary algorithm based on decomposition. Realistic data-based simulations prove that the proposed method achieves a larger total profit in faster convergence speed than the two state-of-the-art algorithms. Note to Practitioners—This article considers the tradeoff between profit maximization and energy cost minimization for the green cloud data center (GCDC) providers while meeting the delay constraints of all tasks. Current task-scheduling methods fail to take the advantage of spatial variations in many factors, e.g., prices of electricity and availability of renewable power generation at geographically distributed GCDC locations. As a result, they fail to execute all tasks of heterogeneous applications within their delay bounds in a high-revenue and low-energy-cost manner. In this article, a multiobjective optimization method that addresses the disadvantages of the existing methods is proposed. It is realized by a proposed intelligent optimization algorithm. Simulations demonstrate that in comparison with the two state-of-the-art scheduling algorithms, the proposed one increases the profit and reduces the convergence time. It can be readily implemented and integrated into actual industrial GCDCs. [ABSTRACT FROM AUTHOR]

A microscopic fuel consumption model with respect to comprehensive fuel consumption key impact factors that can reflect dynamic engine operations and various real-world driving conditions is essential for fuel efficient model-based vehicle and powertrain control (MB-VPC). Existing microscopic fuel consumption models are mostly steady state models with limited vehicle parameters as dynamic variables. Data-driven solutions can make the model comprehensive by introducing broad impact factors. However, without the required accessible online computational capability and fast model adaptation mechanisms, those data-driven solutions cannot quickly adapt to unseen driving conditions and engine condition changes based on real-world vehicle data to support MB-VPC. In this paper, a connected vehicle-based data mining (CV-DM) framework is proposed to achieve online adaptive dynamic fuel consumption modeling through knowledge sharing over CVs and the CV remote data center. Based on the CV-DM framework, CV-supported Distributed Meta-regression (CV-DMR) algorithms are developed to realize a fast few-shot adaptation with limited training data. Extensive proof-of-concept experiments are conducted with steady-state and transient vehicle engine data. Compared to the baseline physical model and the existing non-adaptive grey-box model, prediction accuracy is improved by 47%–85% and 38%–80% respectively with only limited training data needed for the subject vehicle. Accordingly, a fuel savings of up to 9.4% can be achieved owing to the improvement of prediction accuracy. [ABSTRACT FROM AUTHOR]

To boost energy saving for the general delay-tolerant IoT networks, a two-stage, and single-relay queueing communication scheme is investigated. Concretely, a traffic-aware $N$ -threshold and gated-service policy are applied at the relay. As two fundamental and significant performance metrics, the mean waiting time and long-term expected power consumption are explicitly derived and related with the queueing and service parameters, such as packet arrival rate, service threshold and channel statistics. Besides, we take into account the electrical circuit energy consumptions when the relay server and access point (AP) are in different modes and energy costs for mode transitions, whereby the power consumption model is more practical. The expected power minimization problem under the mean waiting time constraint is formulated. Tight closed-form bounds are adopted to obtain tractable analytical formulae with less computational complexity. The optimal energy-saving service threshold that can flexibly adjust to packet arrival rate is determined. In addition, numerical results reveal that: 1) sacrificing the mean waiting time not necessarily facilitates power savings; 2) a higher arrival rate leads to a greater optimal service threshold; and 3) our policy performs better than the current state-of-the-art. [ABSTRACT FROM AUTHOR]

The MapReduce programming paradigm has been increasingly adopted to implement data-intensive applications processing both small and large scale datasets. As most jobs in data centers have a data footprint in the order of gigabytes, emerging high-end scale-up machines are capable of running most data center processing tasks, thus significantly improving power and server density. However, this approach provides limited performance and energy efficiency because of inefficient utilization of the memory subsystem and serial execution within the MapReduce programming model. Recent work has proposed a distributed hardware acceleration architecture, called CASM, which augments each core in a scale-up machine with a lightweight compute engine. The CASM's network of accelerators operates concurrently with the cores in executing MapReduce stages and reduces significantly traffic to/from storage. In this article, we study the benefits and applicability of CASM, by offering an extensive analysis of design parameters and of its scalable performance on a wide range of applications, and exploring its applicability to incremental data aggregation tasks. Our experimental evaluation indicates that CASM reduces off-chip traffic by four times on average over a chip multiprocessor solution, while scaling well with the number of cores in the system, and it is highly effective in providing incremental results that approximate final outcomes. [ABSTRACT FROM AUTHOR]

Direct liquid cooling enables the development of IC, especially processors in servers, toward higher density, which demands more efficient means of heat removal. To exploit the capability, the control system that dynamically balances between cooling power and energy consumption is critical to operational safety and cost. This article presents the design, implementation, and characteristic of a dynamic control system for server processor direct liquid cooling in a high-density data center. The controller was trained and deployed in an online–offline structure. The loss function to train this model was built based on the model predictive control (MPC) principle. To simplify the complex physical model, a hybrid model with a nonlinear steady-state term and a linearized transient term was extracted. We used both averaged overheat and ratio of pumping power consumption to evaluate the performance of controllers. Benchmark tests on open- and closed-loop controllers were conducted. The results showed the advantages of the deep explicit MPC (DEMPC) in terms of cooling effectiveness and energy efficiency. The normalized pumping energy consumption of DEMPC reached $2.3\times 10^{-4}$ with the mean overheat level of −0.96 °C and the overheat time percentage of 6.5%. Comparing with open-loop control that has similar normalized energy costs, the mean overheat level was reduced by 1.2 °C and the time percentage of overheat was reduced by eight times. Meanwhile, normalized energy consumption was reduced by 21.7% than closed-loop proportional control who has a similar average overheat level and overheat time percentage. [ABSTRACT FROM AUTHOR]

We study the problem of the energy-efficient networking in cloud services with geographically distributed data centers for industrial Internet-of-Things (IIoT) networks, specially for multimedia IIoT networks. This is significantly challenged by dynamic end-to-end request demands and unbalanced link energy efficiency, unbalanced and time-varying link utilization, and bandwidth and delay constraints for service requirements. To solve these issues, we propose a multi-constraint optimization model for the energy efficiency optimization in cloud computing services where data centers are geographically distributed and are interconnected by cloud networks. Our model jointly optimizes energy efficiency in data centers and cloud networks. An intelligent heuristic algorithm is presented to solve this model for dynamic request demands between different data centers and between data centers and users. This is implemented by combining the niche genetic algorithm and the random depth-first search. Simulation results for energy-efficient networking show that better gains in network energy efficiency can be achieved by our joint optimization. Joint optimization between industrial data centers and industrial cloud networks can further improve energy savings and link utilization for time-varying requests. [ABSTRACT FROM AUTHOR]

Custom accelerators improve the energy efficiency, area efficiency, and performance of deep neural network (DNN) inference. This article presents a scalable DNN accelerator consisting of 36 chips connected in a mesh network on a multi-chip-module (MCM) using ground-referenced signaling (GRS). While previous accelerators fabricated on a single monolithic chip are optimal for specific network sizes, the proposed architecture enables flexible scaling for efficient inference on a wide range of DNNs, from mobile to data center domains. Communication energy is minimized with large on-chip distributed weight storage and a hierarchical network-on-chip and network-on-package, and inference energy is minimized through extensive data reuse. The 16-nm prototype achieves 1.29-TOPS/mm2 area efficiency, 0.11 pJ/op (9.5 TOPS/W) energy efficiency, 4.01-TOPS peak performance for a one-chip system, and 127.8 peak TOPS and 1903 images/s ResNet-50 batch-1 inference for a 36-chip system. [ABSTRACT FROM AUTHOR]

Energy efficiency is a critical issue in the management of data centers, which form the backbone of cloud computing. Virtual resource management has a significant impact on improving the energy efficiency of data centers. Despite the progress in this area, virtual resource management has been considered mainly at two separate levels: application assignment and virtual machine placement. It has not been well-investigated in a unified framework for both levels, limiting further improvement in the energy efficiency of data centers. To address this issue, this paper proposes the virtual resource management problem for energy efficiency as a constrained optimization problem. Then, this paper simplifies the problem through profile-guided task classification and problem decomposition for complexity reduction and improved energy efficiency. After that, a three-phase framework and algorithms are presented for profiling and profile updating, task classification and application assignment, and successive virtual machine placement. Experimental studies show energy savings of 8–12% by the three-phase framework compared to the existing technique. [ABSTRACT FROM AUTHOR]

The advancement of silicon photonics promises integrated optical switches to provide high-bandwidth, low-latency, and low-power communications in data centers. An optical switch’s loss limits its scale and affects the energy efficiency of the switch system. In this paper, we present cross-layer optical switch optimization (CLOSO), a cross-layer optimization (CLO) framework, based on not only photonic device models at the physical layer but also optical switch models at the fabric layer. With the proposed framework, optimal losses of optical switches can be evaluated efficiently, and the corresponding losses and design parameters of photonic devices can be obtained. Using CLOSO, we optimize four categories of integrated optical switches, Crossbar, PILOSS, DRAGON, and FODON, and compare them regarding their optimal worst-case loss with variation of the switch scale and data rate of signals. Furthermore, system-level evaluations of the optimized optical switches are performed, demonstrating a significant improvement of energy efficiency from the CLO. For instance, CLOSO helps to reduce the energy consumption of a 64-port DRAGON and FODON to as low as 6 pJ/bit and that of a 128-port DRAGON and FODON to as low as 10 pJ/bit. The investigation of 128-port switches also shows the necessity of adaptive power control on lasers for high-radix integrated optical switches. Through quantitative analyses and comparisons, CLOSO shows the capability of facilitating initial design exploration of optical switches and paves the way to fair evaluations and comparisons of switch systems in data centers. [ABSTRACT FROM AUTHOR]

Multipath TCP (MPTCP) has recently been suggested as a promising transport protocol to boost the utilization of underlaying datacenter networks, yet it also increases the host CPU power consumption. It remains unclear whether datacenters can indeed benefit from using MPTCP from the perspective of energy efficiency. Through realworld measurement of MPTCP, we show that the energy efficiency of MPTCP is largely related to the flow completion time and the existence of link-sharing subflows. In particular, we find that the link-sharing subflows in MPTCP will significantly elevate the CPUs’ power consumption on hosts. To make the matter worse, it will also reduce the transmission efficiency for both throughput-sensitive long flows and latency-sensitive short flows. To address such a problem, we present MPTCP-D, an energy-efficient enhancement of MPTCP in datacenter networks. MPTCP-D incorporates a novel congestion control algorithm that improves energy efficiency by minimizing the flow completion time. It also has a build-in subflow elimination mechanism that precludes link-sharing subflows from increasing the host CPU power consumption. We implement MPTCP-D in the Linux kernel, analyze the parameter selection in the algorithm and study its performance through packet-level simulation and on Amazon EC2. Our results show that, without degrading the performance of the long flow throughput and the short flow completion time, MPTCP-D reduces the long flow energy consumption by up to 72% compared to DCTCP for data transfers, and reduces the short flow power consumption by up to 46% compared to MPTCP with link-sharing subflows. [ABSTRACT FROM AUTHOR]

In this article, we discuss the design and implementation of a novel DCN system, which utilizes a knowledge-defined NO-M to operate a HOEDCN cost-effectively and energy-efficiently. The motivations behind the proposed HOE-DCN system are the urgent need to address the scalability, energy, and manageability issues in existing DCN systems. To realize the knowledge-defined NO-M, we follow the principle of predictive analytics in the human brain to design three artificial intelligence modules based on deep learning and make them operate collaboratively. The proposed HOE-DCN system is implemented in a network testbed, and we conduct experiments that involve both control and data plane operations to demonstrate its advantages. The experimental results show that the HOE-DCN simultaneously achieves high-performance service provisioning and improved energy efficiency. Furthermore, by analyzing the pros and cons of the HOE-DCN system, we also point out several directions to work on in the future. [ABSTRACT FROM AUTHOR]

In recent years we have experienced a wide adoption of novel distributed processing frameworks such as Apache Spark for handling batch and stream processing big data applications. An important aspect that has not been examined in these systems yet, is the energy consumption during the applications’ execution. Reducing the energy consumption of modern datacenters is a necessity, as datacenters contribute over 2 percent of the total US electric usage. However, efficiently scheduling applications in distributed processing systems can be challenging as there is a trade-off between minimizing the datacenter's energy usage and satisfying the application performance requirements. In this work we propose, ExpREsS, a scheduler for orchestrating the execution of Spark applications in a way that enables us to minimize the energy consumption while ensuring that the applications’ performance requirements are met. Our approach exploits time-series segmentation for capturing the applications’ energy usage and execution times, and then applies a novel DVFS technique to minimize the energy consumption. In order to tackle the limited number of application's profiling runs, we exploit regression techniques to predict the applications’ execution times and power consumption. Our detailed experimental evaluation using realistic workloads on our local cluster illustrates the working and benefits of our approach. [ABSTRACT FROM AUTHOR]

As the IT sector is rapidly growing all over the world, more and more energy is consumed by data centers, which leads to the continuous increase of operating costs. Among other factors, the load management among uninterruptible power supply (UPS) nodes should be considered in the energy management model of the data center since its energy efficiency is correlated with the load ratio. In this article, an energy management scheme of the data center, which focuses on the workload scheduling among UPS nodes, is proposed to improve its operation efficiency. The flexible service requirement of batch workloads is also modeled in the proposed scheme to illustrate the cost of flexibility. In the proposed scheme, the efficiency of the UPS node is modeled as a function of the load ratio. Meanwhile, the cost of inconvenience is introduced to represent the flexible service requirements of batch workloads. A stochastic formulation is established to address the uncertainties involved in the scheduling process. Simulated case studies are provided, and the results demonstrate the effectiveness of the proposed approach. [ABSTRACT FROM AUTHOR]

Analyzing data centers with thermal-aware optimization techniques is a viable approach to reduce energy consumption of data centers. By taking into account thermal consequences of job placements among the servers of a data center, it is possible to reduce the amount of cooling necessary to keep the servers below a given safe temperature threshold. We set up an optimization problem to analyze and characterize the optimal set points for the workload distribution and the supply temperature of the cooling equipment. Furthermore, under mild assumptions, we design and analyze controllers that regulate the system to the optimal state without knowledge of the current total workload to be handled by the data center. The response of our controller is validated by simulations and convergence to the optimal set points is achieved under varying workload conditions. [ABSTRACT FROM AUTHOR]

Internet data centers are booming in the last few years, leading to gigantic amount of energy consumption and greenhouse gas emission. In this paper, a data center microgrid including conventional units, energy storage, and renewable generation is modeled to optimize its electricity bill and carbon footprint. Considering the randomness of electricity price, renewable output and workload distribution, the volatility risk is incorporated to measure the associated operation risk. Then, a day-ahead emission-aware stochastic resource planning scheme is formulated to decide the strategy on power procurement, energy storage operation, batch workload allocation, and unit commitment of conventional units. Simulation results show the effectiveness of the proposed approach. [ABSTRACT FROM AUTHOR]

Designing a cost-effective network for data centers that can deliver sufficient bandwidth and provide high availability has drawn tremendous attentions recently. In this paper, we propose a novel server-centric network structure called RCube, which is energy efficient and can deploy a redundancy scheme to improve the availability of data centers. Moreover, RCube shares many good properties with BCube, a well known server-centric network structure, yet its network size can be adjusted more conveniently. We also present a routing algorithm to find paths in RCube and an algorithm to find multiple parallel paths between any pair of source and destination servers. In addition, we theoretically analyze the power efficiency of the network and availability of RCube under server failure. Our comprehensive simulations demonstrate that RCube provides higher availability and flexibility to make trade-off among many factors, such as power consumption and aggregate throughput, than BCube, while delivering similar performance to BCube in many critical metrics, such as average path length, path distribution and graceful degradation, which makes RCube a very promising empirical structure for an enterprise data center network product. [ABSTRACT FROM AUTHOR]

Previous work on optimizing resource provisioning in virtualized environments focused either on mapping virtual machines (VMs) to physical machines (PMs) or mapping application components to VMs. In this paper, we argue that these two optimization problems influence each other significantly and in a highly non-trivial way. We define a sophisticated problem formulation for the joint optimization of the two mappings, taking into account sizing aspects, colocation constraints, license costs, and hardware affinity relations. As demonstrated by the empirical evaluation on a real-world workload trace, the combined optimization leads to significantly better overall results than considering the two problems in isolation. [ABSTRACT FROM PUBLISHER]

Aisle containment separates hot and cold aisles in air-cooled data centers to prevent hot air recirculation into cold aisle and decrease temperature nonuniformity among the servers. Uniform server inlet temperatures allow the cooling system to operate more efficiently at a higher evaporator temperature. Lower cooling power consumption can be achieved by optimizing the combined power consumption of the chillers and the computer room air handlers (CRAHs) fans. CRAH bypass (BP) is proposed, in which additional fans at the tiles induce a fraction of tile flow from the room into the plenum through low-resistance ports or leakage paths, thus decreasing the airflow passing through the high pressure resistance of the CRAH heat exchangers and filters and associated fan power usage. A further advantage of the proposed induced CRAH BP is the elimination of leakage as a result of reversing the room-plenum pressure difference. However, in order to keep the enclosed-aisle temperature below acceptable limits (typically ≤27 °C), the chiller needs to operate at a lower, less efficient temperature. This paper presents the experimental verification of a flow network model (FNM) in a data center test cell demonstrating the proposed induced CRAH BP concept. Exercise of the verified FNM in conjunction with a thermodynamic model of the cooling infrastructure (TDM) confirms that there is an optimum BP fraction that minimizes the combined chiller and CRAH fan power consumption. [ABSTRACT FROM AUTHOR]

An important challenge of modern data centers is to reduce energy consumption, of which a substantial proportion is due to the network. Switches and NICs supporting the recent energy efficient Ethernet (EEE) standard are now available, but current practice is to disable EEE in production use, since its effect on real world application performance is poorly understood. This paper contributes to this discussion by analyzing the impact of EEE on MapReduce workloads, in terms of performance overheads and energy savings. MapReduce is the central programming model of Apache Hadoop, one of the most widely used application frameworks in modern data centers. We find that, while 1GbE links (edge links) achieve good energy savings using the standard EEE implementation, optimum energy savings in the 10 GbE links (aggregation and core links) are only possible, if these links employ packet coalescing. Packet coalescing must, however, be carefully configured in order to avoid excessive performance degradation. With our new analysis of how the static parameters of packet coalescing perform under different cluster loads, we were able to cover both idle and heavy load periods that can exist on this type of environment. Finally, we evaluate our recommendation for packet coalescing for 10 GbE links using the energy-delay metric. This paper is an extension of our previous work [1], which was published in the Proceedings of the 40th Annual IEEE Conference on Local Computer Networks (LCN 2015). [ABSTRACT FROM PUBLISHER]

As big data applications proliferate, datacenters today are increasingly looking to adopt a scale-out model. Nevertheless, power capacity has become an important bottleneck that restricts horizontal scaling of servers, especially in datacenters that oversubscribe power infrastructure. When a datacenter hits its ceiling for power provisioning, conventionally the owner has to either build another facility or upgrade existing infrastructure—both approaches add huge cost, require significant time, and can further increase carbon footprint. This paper proposes Oasis, a novel datacenter expansion strategy that enables power-/carbon- constrained servers to scale out economically and sustainably. The basic structure of Oasis, called Oasis Node, naturally supports incremental capacity expansion with near-zero environmental impact since it leverages modular solar panels and distributed battery systems to power newly added servers. To optimize the operation of newly added nodes, we further propose a management framework called Ozone. It allows Oasis to jointly perform power supply switching and server speed scaling to improve efficiency locally and globally. We implement a prototype of Oasis and use it as a research platform for evaluating the design tradeoffs of green scale-out datacenters. With Oasis, a green datacenter could gradually double its capacity with near-oracle performance, extended battery lifetime, and 26 percent cost savings. [ABSTRACT FROM PUBLISHER]

Geographical load balancing can optimize the utilization of green energy and the cost of electricity by taking advantage of greenness and price diversities at geographically dispersed data centers. However, higher green energy utilization or lower electricity cost may actually increase the total energy consumption, and is not necessarily the best option. The achievable energy trade-offs can be captured by taking into consideration a defined service efficiency parameter for geo-dispersed data centers. [ABSTRACT FROM PUBLISHER]

This brief conceives a learning system for implementing self-optimization-based dynamic server resource provisioning (DSRP) of data centers under deregulated electricity markets. We formulate the DSRP problem as a constrained Markov decision process to minimize the electricity cost subject to a constraint on the queue delay. Instead of applying conventional Q-learning to solve this problem, a postdecision state learning-based DSRP algorithm having fast convergence is proposed by estimating and exploiting the workload arrival distribution. We further discuss the offline optimization of the DSRP problem, which is used as the performance benchmark of the proposed method. Finally, we evaluate the performance of the proposed scheme by using real workloads and electricity prices. [ABSTRACT FROM PUBLISHER]

Due to the increasing usage of cloud computing applications, it is important to minimize energy cost consumed by a data center, and simultaneously, to improve quality of service via data center management. One promising approach is to switch some servers in a data center to the idle mode for saving energy while to keep a suitable number of servers in the active mode for providing timely service. In this paper, we design both online and offline algorithms for this problem. For the offline algorithm, we formulate data center management as a cost minimization problem by considering energy cost, delay cost (to measure service quality), and switching cost (to change servers’s active/idle mode). Then, we analyze certain properties of an optimal solution which lead to a dynamic programming based algorithm. Moreover, by revising the solution procedure, we successfully eliminate the recursive procedure and achieve an optimal offline algorithm with a polynomial complexity. For the online algorithm, We design it by considering the worst case scenario for future workload. In simulation, we show this online algorithm can always provide near-optimal solutions. [ABSTRACT FROM PUBLISHER]

This work presents an on-line, energy- and communication-aware scheduling strategy for SaaS applications in data centers. The applications are composed of various services and represented as workflows. Each workflow consists of tasks related to each other by precedence constraints and represented by Directed Acyclic Graphs (DAGs). The proposed scheduling strategy combines advantages of state-of-the-art workflow scheduling strategies with energy-aware independent task scheduling approaches. The process of scheduling consists of two phases. In the first phase, virtual deadlines of individual tasks are set in the central scheduler. These deadlines are determined using a novel strategy that favors tasks which are less dependent on other tasks. During the second phase, tasks are dynamically assigned to computing servers based on the current load of network links and servers in a data center. The proposed approach, called Minimum Dependencies Energy-efficient DAG (MinD+ED) scheduling, has been implemented in the GreenCloud simulator. It outperforms other approaches in terms of energy efficiency, while keeping a satisfiable level of tardiness. [ABSTRACT FROM AUTHOR]

Data centers hosting distributed computing systems consume huge amounts of electrical energy, contributing to high operational costs, whereas the utilization of data centers continues to be very low. Moreover, a data center generally consists of heterogeneous servers with different performance and energy. Failure to fully consider the heterogeneity of servers will lead to both sub-optimal energy saving and performance. In this study, we employ game theoretic approaches to model the problem of minimizing energy consumption as a Stackelberg game. In our model, the system monitor, who plays the role of the leader, can maximize profit by adjusting resource provisioning, whereas scheduler agents, who act as followers, can select resources to obtain optimal performance. In addition, we model the problem of minimizing average response time of tasks as a noncooperative game among decentralized scheduler agents as they compete with one another in the sharing resources. Several algorithms are presented to implement the game models. Simulation results demonstrate that the proposed technique has immense potential to improve energy efficiency under dynamic work scenarios without compromising service level agreements. [ABSTRACT FROM AUTHOR]

We study the problem of job assignment in a large-scale realistically dimensioned server farm comprising multiple processor-sharing servers with different service rates, energy consumption rates, and buffer sizes. Our aim is to optimize the energy efficiency of such a server farm by effectively controlling carried load on networked servers. To this end, we propose a job assignment policy, called Most energy-efficient available server first Accounting for Idle Power (MAIP), which is both scalable and near optimal. MAIP focuses on reducing the productive power used to support the processing service rate. Using the framework of semi-Markov decision process, we show that, with exponentially distributed job sizes, MAIP is equivalent to the well-known Whittle’s index policy. This equivalence and the methodology of Weber and Weiss enable us to prove that, in server farms where a loss of jobs happens if and only if all buffers are full, MAIP is asymptotically optimal, as the number of servers tends to infinity under certain conditions associated with the large number of servers, as we have in a real server farm. Through extensive numerical simulations, we demonstrate the effectiveness of MAIP and its robustness to different job-size distributions, and observe that significant improvement in energy efficiency can be achieved by utilizing the knowledge of energy consumption rate of idle servers. [ABSTRACT FROM PUBLISHER]

Finding the best way to map virtual machines (VMs) to physical machines (PMs) in a cloud data center is an important optimization problem, with significant impact on costs, performance, and energy consumption. In most situations, the computational capacity of PMs and the computational load of VMs are a vital aspect to consider in the VM-to-PM mapping. Previous work modeled computational capacity and load as one-dimensional quantities. However, today's PMs have multiple processor cores, all of which can be shared by cores of multiple multicore VMs, leading to complex scheduling issues within a single PM, which the one-dimensional problem formulation cannot capture. In this paper, we argue that at least a simplified model of these scheduling issues should be taken into account during VM placement. We show how constraint programming techniques can be used to solve this problem, leading to significant improvement over non-multicore-aware VM placement. Several ways are presented to hybridize an exact constraint solver with common packing heuristics to derive an effective and scalable algorithm. [ABSTRACT FROM AUTHOR]

Recently, datacenter carbon emission has become an emerging concern for the cloud service providers. Previous works are limited on cutting down the power consumption of datacenters to defuse such a concern. In this paper, we show how the spatial and temporal variabilities of the electricity carbon footprint can be fully exploited to further green the cloud running on top of geographically distributed datacenters. Specifically, we first verify that electricity cost minimization conflicts with carbon emission minimization, based on an empirical study of several representative geo-distributed cloud services. We then jointly consider the electricity cost, service level agreement (SLA) requirement, and emission reduction budget. To navigate such a three-way tradeoff, we take advantage of Lyapunov optimization techniques to design and analyze a carbon-aware control framework, which makes online decisions on geographical load balancing, capacity right-sizing, and server speed scaling. Results from rigorous mathematical analysis and real-world trace-driven evaluation demonstrate the effectiveness of our framework in reducing both electricity cost and carbon emission. [ABSTRACT FROM PUBLISHER]

Motivated by applications in data centers, we consider a scenario where multiple classes of requests arrive at a dispatcher at time-varying rates which historically has daily or weekly patterns. We assume that the underlying environment is such that at all times the load from each class is very high and a large number of servers are necessary which, for example, is fairly common in many data centers. In addition, each server can host one or more classes. Design, control and performance analysis under such heterogeneous and transient conditions is extremely difficult. To address this shortcoming we have suggested a holistic approach that includes a combination of sizing, assignment, and routing in an integrated fashion. Our proposed approach decomposes a multi-dimensional and non-stationary problem into a one-dimensional, simpler and stationary one, and achieves time-stability by introducing an insignificant number of dummy requests. Based on time-stability, our suggested approach can provide performance bounds and guarantees for time-varying and transient system. Moreover, we can operate the data centers in an energy-efficient manner via suggested approach. [ABSTRACT FROM PUBLISHER]

Companies operating large datacenters are focusing on how to reduce the electrical energy costs of operating datacenters. A common way of cost reduction is to perform a dynamic voltage and frequency scaling (DVFS), thereby matching the CPU’s performance and power level to incoming workloads. Another power saving technique is CPU consolidation, which uses the minimum number of CPUs necessary to meet the service request demands and turns OFF the remaining unused CPUs. DVFS has been already extensively studied and verified its effectiveness. On the other hand, it is necessary to study more about the effectiveness of CPU consolidation. Key questions that must be answered are how effectively the CPU consolidation improves the energy efficiency and how to maximize the improvement. These questions are addressed in this paper. After understanding modern power management techniques and developing an appropriate power model, this paper provides an extensive set of hardware-based experimental results and makes suggestions about how to maximize energy efficiency improvement through CPU consolidation. In addition, this paper also presents new online CPU consolidation algorithms, which reduce the energy-delay product up to 13% compared with the Linux default DVFS algorithm. [ABSTRACT FROM AUTHOR]

The paper proposes an adaptive framework for modern data centers that jointly manages application- and system-level (dynamic voltage and frequency scaling) adaptation to improve energy efficiency of multicore servers. The results of applying this framework show that significant power savings and performance improvements are possible with respect to current data center management techniques. [ABSTRACT FROM AUTHOR]

Energy efficiency of server farms is an important design consideration of the green datacenter initiative. One effective approach is to optimize power consumption of server farms by controlling the carried load on the networked servers. In this paper, we propose a robust heuristic policy called E* for stochastic job assignment in a server farm, aiming to improve the energy efficiency by maximizing the ratio of job throughput to power consumption. Our model of the server farm considers a parallel system of finite-buffer processor-sharing queues with heterogeneous server speeds and energy consumption rates. We devise E* as an insensitive policy so that the stationary distribution of the number of jobs in the system depends on the job size distribution only through its mean. We provide a rigorous analysis of E* and compare it with a baseline approach, known as most energy-efficient server first (MEESF), that greedily chooses the most energy-efficient servers for job assignment. We show that E* has always a higher job throughput than that of MEESF, and derive realistic conditions under which E* is guaranteed to outperform MEESF in energy efficiency. Extensive numerical results are presented and demonstrate that E* can improve the energy efficiency by up to 100%. [ABSTRACT FROM PUBLISHER]

This work studies the online electricity cost minimization problem at a co-location data center, which serves multiple tenants who rent the physical infrastructure within the data center to run their respective cloud computing services. The co-location operator has no direct control over power consumption of its tenants, and an efficient mechanism is desired for eliciting desirable consumption patterns from the tenants. Electricity billing faced by a data center is nowadays based on both the total volume consumed and the peak consumption rate. This leads to an interesting new combinatorial optimization structure on the electricity cost optimization problem, which also exhibits an online nature due to the definition of peak consumption. We model and solve the problem through two approaches: the pricing approach and the auction approach, and design online algorithms with small competitive ratios. [ABSTRACT FROM PUBLISHER]

In this work, we present an exhaustive empirical characterization of the power requirements of multiple components of data center servers. To do so, we devise different experiments to stress these components, taking into account the multiple available frequencies and the fact that we are working with multicore servers. In these experiments, we measure energy consumption of server components and identify their optimal operational points. Our study proves that the curve defining the minimal CPU power utilization, as a function of the load in active cycles per second, is neither concave nor purely convex. Instead, it definitively shows a super-linear dependence on the load. Similarly, we present results on how to improve the efficiency of network cards and disks. Finally, we validate the accuracy of the model derived from our characterization by comparing the real energy consumed by two Hadoop applications—PageRank and WordCount—with the estimation from our model, obtaining errors below 4.1% on average. [ABSTRACT FROM PUBLISHER]

The explosive expansion of data center sizes aggravates the power consumption and carbon footprint, which has restricted the sustainable growth of cloud services and seriously troubled data center operators. In recent years, plenty of advanced data center network architectures have been proposed. They usually employ richly-connected topologies and multi-path routing to provide high network capacity. Unfortunately, they also undergo inefficient network energy usage during the traffic valley time. To address the problem, many energy-aware flow scheduling algorithms are proposed recently, primarily considering how to aggregate traffic by flexibly choosing the routing paths, with flows fairly sharing the link bandwidths. In this paper, we leverage software defined network (SDN) technique and explore a new solution to energy-aware flow scheduling, i.e., scheduling flows in the time dimension and using exclusive routing (EXR) for each flow, i.e., a flow always exclusively utilizes the links of its routing path. The key insight is that exclusive occupation of link resources usually results in higher link utilization in high-radix data center networks, since each flow does not need to compete for the link bandwidths with others. When scheduling the flows, EXR leaves flexibility to operators to define the priorities of flows, e.g., based on flow size, flow deadline, etc. Extensive simulations and testbed experiments both show that EXR can effectively save network energy compared with the regular fair-sharing routing (FSR), and significantly reduce the average flow completion time if assigning higher scheduling priorities to smaller flows. [ABSTRACT FROM PUBLISHER]

The past decade witnessed the tremendous growth of online services and applications. Together with the increase of cloud computing, more and more computation are hosted by Internet data centers (IDCs). Today’s IDCs are achieving significant advances in communication and computation capabilities. However, along with the increasing demand from IDC clients, power consumption for powering up and cooling these IDCs has been skyrocketing. Most existing works optimize the power consumption of either servers or Computer Room Air Conditioners (CRACs), and overlook the correlation between the power consumption of these two types of equipment. In this paper, we propose an adaptive power control method which leverages the correlation between the power consumption of servers and CRACs. To capture the workload uncertainties and thermal dynamics, we exploit Recursive-Least Square based Model Predictive Control (MPC) to solve the power control problem. Performance evaluations shows the effective power peak reduction using our approach. [ABSTRACT FROM AUTHOR]

Distributed renewable energy has emerged as a promising resource because of its environmental friendliness and economic considerations. However, most renewable energy sources are unreliable and may require considerable effort to be efficiently utilized in a computing center for providing services. In this paper, we exploit distributed renewable energy (e.g., solar energy) and peer-to-peer (P2P) technologies to aggregate distributed computing resources to provide an infrastructure called solar-based energy-efficient distributed server (SEEDS) farm for distributed computing and distributed storage. Energy-efficient devices (e.g., embedded devices) powered by solar energy form a P2P computing system to provide their computing resources to end-users. Specifically, this paper uses a Web-based service as a case study. A group of solar-powered embedded devices acts as a front-end cooperative caching system for Web servers. Web objects may be accessed through the distributed caching system without going through servers, and thus we can reduce brown energy consumption of the servers. This paper also develops an analytical model to evaluate the total energy consumption of Web-based services with and without SEEDS. Theoretical and simulation results show that the SEEDS system can support services and achieve significant improvements in energy efficiency by aggregating distributed energy resources. [ABSTRACT FROM AUTHOR]

The effectiveness of heat exchangers strongly influences the thermal performance of data center cooling systems. Rear door heat exchangers, in-row and overhead coolers, and fully contained water cooled cabinets are some examples of liquid and hybrid cooling systems used in data centers. Modeling the dynamic behavior of heat exchangers is important for the design of control strategies to improve energy efficiency. In this paper, an existing 2-D model of the transient temperature response of an unmixed-unmixed cross flow heat exchanger, of the type that is widely used in data center cooling equipment, is solved numerically. A detailed analysis of the transient response to a step, ramp, or exponential change in the hot fluid inlet temperature is conducted. The heat capacity rate ratio ( $E$ ), thermal resistance ( $R$ ), capacitance ratio ( $V$ ), and number of transfer units are varied over a wide range to determine their influence on the heat exchanger dynamic thermal performance under different transient input conditions. In addition, the thermal response to the magnitude and time period of the transient input functions are evaluated. The modeling results are used to analyze specific data center cooling scenarios, and provide a means for predicting the transient behavior of heat exchangers used in data center cooling equipment. [ABSTRACT FROM AUTHOR]

Why do dynamic power-management technologies that dramatically improve datacenter server energy efficiency continue to go unleveraged? [ABSTRACT FROM PUBLISHER]

We study the problem of stochastic job assignment in a server farm comprising multiple processor-sharing servers with various speeds and finite buffer sizes. We consider two types of assignment policies: without jockeying, where an arriving job is assigned only once to an available server, and with jockeying, where a job may be reassigned at any time. We also require that the underlying Markov process under each policy is insensitive. Namely, the stationary distribution of the number of jobs in the system is independent of the job size distribution except for its mean. For the case without jockeying, we derive two insensitive heuristic policies: One aims at maximizing job throughput, and the other trades off job throughput for energy efficiency. For the case with jockeying, we formulate the optimal assignment problem as a semi-Markov decision process and derive optimal policies with respect to various optimization criteria. We further derive two simple insensitive heuristic policies with jockeying: One maximizes job throughput, and the other aims at maximizing energy efficiency. Numerical examples demonstrate that, under a wide range of system parameters, the latter policy performs very close to the optimal policy. Numerical examples also demonstrate energy/throughput tradeoffs for the various policies and, in the case with jockeying, they show a potential of substantial energy savings relative to a policy that optimizes throughput. [ABSTRACT FROM PUBLISHER]

Software-defined networking is an emerging method for providing flexible and scalable network connectivity in both intra- and inter-datacenter interconnects with regard to various requirements, including traffic-awareness, quality of service, energy efficiency, and renewable-power intermittency. This article investigates issues and solutions for the software-defined planning of a virtual slice that involves multiple virtual machines with interdependent constraints spanning a network of distributed datacenters. A flexible and optimized virtual-slice assignment that considers server consolidation and multipath forwarding can address large-scale cloud computing services. Simulations on large-scale testbeds, such as the GreenStar Network and the Green Telco Cloud, showed that the proposed model achieves good performance in terms of flexibility and energy efficiency. [ABSTRACT FROM AUTHOR]