Improving Energy Efficiency in Data Centers by Controlling Task Distribution and Cooling

2018 ◽  
Author(s):  
Yusuke Nakajo ◽  
Jayati Athavale ◽  
Minami Yoda ◽  
Yogendra Joshi ◽  
Hiroaki Nishi
Keyword(s):  
Energy Efficiency ◽  
Load Balancing ◽  
Rapid Growth ◽  
Data Centers ◽  
Temperature Response ◽  
Total Power ◽  
Raspberry Pi ◽  
Total Power Consumption ◽  
Institute Of Technology ◽  
Global Energy Supply

The rapid growth in cloud computing, the Internet of Things (IoT), and data processing via Machine Learning (ML), have greatly increased our need for computing resources. Given this rapid growth, it is expected that data centers will consume more and more of our global energy supply. Improving their energy efficiency is therefore crucial. One of the biggest sources of energy consumption is the energy required to cool the data centers, and ensure that the servers stay within their intended operating temperature range. Indeed, about 40% of a data center’s total power consumption is for air conditioning[1]. Here, we study how the server air inlet and outlet, as well as the CPU, temperatures depend upon server loads typical of real Internet Protocol (IP) traces. The trace data used here are from Google clusters and include the times, job and task ID, as well as the number and usage of CPU cores. The resulting IT loads are distributed using standard load-balancing methods such as Round Robin (RR) and the CPU utilization method. Experiments are conducted in the Data Center Laboratory (DCL) at the Georgia Institute of Technology to monitor the server outlet air temperature, as well as real-time CPU temperatures for servers at different heights within the rack. Server temperatures were measured by on-line temperature monitoring with Xbee, Raspberry PI, Arduino, and hot-wire anemometers. Given that the temperature response varies with server position, in part due to spatial variations in the cooling airflow over the rack inlet and the server fan speeds, a new load-balancing approach that accounts for spatially varying temperature response within a rack is tested and validated in this paper.

Spatially-Aware Optimization of Energy Consumption in Consolidated Data Center Systems

2011 ◽  
Author(s):  
Hui Chen ◽  
Mukil Kesavan ◽  
Karsten Schwan ◽  
Ada Gavrilovska ◽  
Pramod Kumar ◽  
...  
Keyword(s):  
Energy Efficiency ◽  
Data Center ◽  
Power Distribution ◽  
Total Power ◽  
Inlet Temperature ◽  
Workload Management ◽  
Thermal Distribution ◽  
Total Power Consumption ◽  
Application Requirements ◽  
Online Methods

Energy efficiency in data center operation depends on many factors, including power distribution, thermal load and consequent cooling costs, and IT management in terms of how and where IT load is placed and moved under changing request loads. Current methods provided by vendors consolidate IT loads onto the smallest number of machines needed to meet application requirements. This paper’s goal is to gain further improvements in energy efficiency by also making such methods ‘spatially aware’, so that load is placed onto machines in ways that respect the efficiency of both cooling and power usage, across and within racks. To help implement spatially aware load placement, we propose a model-based reinforcement learning method to learn and then predict the thermal distribution of different placements for incoming workloads. The method is trained with actual data captured in a fully instrumented data center facility. Experimental results showing notable differences in total power consumption for representative application loads indicate the utility of a two-level spatially-aware workload management (SpAWM) technique in which (i) load is distributed across racks in ways that recognize differences in cooling efficiencies and (ii) within racks, load is distributed so as to take into account cooling effectiveness due to local air flow. The technique is being implemented using online methods that continuously monitor current power and resource usage within and across racks, sense BladeCenter-level inlet temperatures, understand and manage IT load according to an environment’s thermal map. Specifically, at data center level, monitoring informs SpAWM about power usage and thermal distribution across racks. At rack-level, SpAWM workload distribution is based on power caps provided by maximum inlet temperatures determined by CRAC speeds and supply air temperature. SpAWM can be realized as a set of management methods running in VMWare’s ESXServer virtualization infrastructure. Its use has the potential of attaining up to 32% improvements on the CRAC supply temperature requirement compared to non-spatially aware techniques, which can lower the inlet temperature 2∼3°C, that is to say we can increase the CRAC supply temperature 2∼3°C to save nearly 13% −18% cooling energy.

Performance Research on the All Aluminum Heat Exchanger of Family Air-Condition

2011 ◽  
Vol 71-78 ◽  
pp. 2068-2072
Author(s):  
Rui Wang ◽  
Yi Chun Wang ◽  
Chao Qing Feng ◽  
Huo Ming Zhan ◽  
Hua Jun Li
Keyword(s):  
Energy Efficiency ◽  
Heat Exchanger ◽  
Small Volume ◽  
Total Power ◽  
Efficiency Ratio ◽  
Total Power Consumption ◽  
The Common ◽  
Performance Research ◽  
Pipe Heat ◽  
The Ideal

Air enthalpy method is used in the contrastive experiment of the new condenser and the common wing-pipe heat exchanger of family air-condition. The refrigerating capacity and EER (Energy Efficiency Ratio) are obtained by the experiment. The conclusion of the experiment shows that the new condenser with small volume and diathermanous area can create more refrigerating capacity, but the total power consumption is basically unchanged, so the EER improved. This kind of all aluminum heat exchanger is the ideal substitute of family air-condition’s wing-pipe heat exchanger.

Power Reduction Techniques in Cloud Data Centers

2014 ◽  
Vol 1061-1062 ◽  
pp. 1070-1073
Author(s):  
Lei Tang ◽  
Zheng Ce Cai ◽  
Guo Long Chen ◽  
Xian Wei Li
Keyword(s):  
Cloud Computing ◽  
Data Centers ◽  
Large Fraction ◽  
Power Reduction ◽  
Cloud Services ◽  
Total Power ◽  
Cloud Data ◽  
Reduction Techniques ◽  
Total Power Consumption ◽  
Cloud Data Centers

In recent years, cloud computing has received much attention from both academia and engineering areas. With more and more companies beginning to provide cloud services, more and more data centers are being built. Recent studies show that the energy consumed by cloud data centers accounts for a large fraction of the total power consumption today. This motivates us to survey power reduction techniques in cloud data centers.

scholarly journals Joint Optimization of Pico-Base-Station Density and Transmit Power for an Energy-Efficient Heterogeneous Cellular Network

2019 ◽  
Vol 11 (10) ◽  
pp. 208
Author(s):  
Jie Yang ◽  
Ziyu Pan ◽  
Hengfei Xu ◽  
Han Hu
Keyword(s):  
Energy Efficiency ◽  
Focal Point ◽  
Transmission Rate ◽  
Optimization Method ◽  
Base Station ◽  
Base Stations ◽  
Joint Optimization ◽  
Total Power ◽  
Transmit Power ◽  
Total Power Consumption

Heterogeneous cellular networks (HCNs) have emerged as the primary solution for explosive data traffic. However, an increase in the number of base stations (BSs) inevitably leads to an increase in energy consumption. Energy efficiency (EE) has become a focal point in HCNs. In this paper, we apply tools from stochastic geometry to investigate and optimize the energy efficiency (EE) for a two-tier HCN. The average achievable transmission rate and the total power consumption of all the BSs in a two-tier HCN is derived, and then the EE is formulated. In order to maximize EE, a one-dimensional optimization algorithm is used to optimize picocell BS density and transmit power. Based on this, an alternating optimization method aimed at maximizing EE is proposed to jointly optimize transmit power and density of picocell BSs. Simulation results validate the accuracy of the theoretical analysis and demonstrate that the proposed joint optimization method can obviously improve EE.

scholarly journals Analysis and Design of Colpitts Oscillator for Short-range WSN

2016 ◽  
Vol 62 (3) ◽  
pp. 279-282
Author(s):  
Mousa Yousefi
Keyword(s):  
Energy Efficiency ◽  
Output Power ◽  
Short Range ◽  
Cmos Technology ◽  
Paper Analysis ◽  
Total Power ◽  
Colpitts Oscillator ◽  
Analysis And Design ◽  
Total Power Consumption ◽  
On Chip

Abstract In this paper, analysis and design of colpitts oscillator with ability to transmit data at low output power with application in short-range wireless sensor networks such as MICS is described. Reducing the area required to implement the transmitter, on-chip implementation and appropriate energy efficiency are the advantages of this structure that makes it suitable for the design of short-range transmitter in biomedical applications. The proposed OOK transmitter works at 405 MHz with 10 Mbps data rate. Output power and total power consumption are 25 µW and 726 µW, respectively. Energy efficiency is 72.6 pJ/bit. The transmitter has been designed and simulated in 0.18 µm CMOS technology.

A hierarchical energy efficiency optimization control strategy for distributed drive electric vehicles

2018 ◽  
Vol 233 (3) ◽  
pp. 605-621 ◽  
Author(s):  
Min Hua ◽  
Guoying Chen ◽  
Buyang Zhang ◽  
Yanjun Huang
Keyword(s):  
Energy Efficiency ◽  
Electric Vehicle ◽  
Sliding Mode ◽  
Fixed Ratio ◽  
Tracking Performance ◽  
Total Power ◽  
Control Allocation ◽  
Motor Efficiency ◽  
Total Power Consumption ◽  
Distributed Drive Electric Vehicle

Distributed drive electric vehicle with four in-wheel motors is widespread with various characteristics, such as performance potentials for independent wheel drive control and energy efficiency. However, in future, one of the biggest obstacles for its success in the automotive industry would be its limited energy storage. This paper proposes a hierarchical control method that involves a high-level motion controller that uses sliding mode control to calculate the total desired force and yaw moment and a low-level allocation controller in which an optimal energy-efficient control allocation scheme is presented to provide optimally distributed torques of four in-wheel motors in all the normal cases. A practicable motor energy efficiency model as a motor actuator is proposed by incorporating the electric motor efficiency map based on measured data into the motor efficiency experiment and a current closed-loop motor model. Moreover, both tracking performance and energy-saving are carried out in this research and evaluated via a co-simulation approach using MATLAB/Simulink and CarSim. A ramp maneuver at a constant speed and New European Driving Cycle and Urban Dynamometer Driving Schedule maneuvers have been conducted. To conclude, it is demonstrated that distributed drive electric vehicle with four in-wheel motors can reduce total power consumption and enhance tracking performance compared with a simple control allocation in which the torques are the fixed ratio distribution.

scholarly journals Design and Implementation of Low-Cost Real-Time Energy Logger for Industrial and Home Applications

2021 ◽  
Author(s):  
Pooja R. Khanna ◽  
Gareth Howells ◽  
Pavlos I. Lazaridis
Keyword(s):  
Power Consumption ◽  
Real Time ◽  
Web Application ◽  
Single Phase ◽  
Total Power ◽  
Raspberry Pi ◽  
Energy Usage ◽  
Measurement Device ◽  
Design And Implementation ◽  
Total Power Consumption

AbstractWith the significant increase in energy demands in the last decade, the issues of unnecessary energy usage have increased rapidly. Therefore, there is an immediate need to provide a cheap and easily accessible monitoring tool for the energy consumed by an appliance used in homes and industries. Instead of monitoring the total power consumption of the houses and/or industries, it is useful to monitor the power consumption of the individual appliance, which in turn, helps in saving the overall energy usage and thereby makes it cost-effective. This paper presents a cost-efficient design and implementation of a monitoring system that can precisely measure the current and voltage of each appliance. The design provides tracking of device activity in a real-time environment for the industries and helps in adopting to the green initiative. The design comprises of Arduino based micro-controller and Raspberry Pi, that performs precise measurements of current and voltage of the device, followed by measuring the power consumed by the device. This paper presents two different system designs, one for the single-phase measurements and the other for the DC measurements. The single-phase measurement device comprises of 10-bit ADC whereas, the 24 V DC measurement device comprises of a 12-bit ADC, which provides higher measurement accuracy compared to other systems available in the market. The implemented design uses the EmonCMS web application to accumulate and envision the monitored data. It provides a flexible and user-friendly solution to monitor the measured data easily on any android or iOS devices.

scholarly journals Energy Efficiency in 5G Communications – Conventional to Machine Learning Approaches

2020 ◽  
Vol 4 ◽  
pp. 1-9
Author(s):  
Muhammad Khalil Shahid ◽  
Filmon Debretsion ◽  
Aman Eyob ◽  
Irfan Ahmed ◽  
Tarig Faisal
Keyword(s):  
Machine Learning ◽  
Energy Efficiency ◽  
Data Exchange ◽  
Mixed Reality ◽  
Mimo Systems ◽  
Base Station ◽  
Small Cells ◽  
Total Power ◽  
5G Networks ◽  
Total Power Consumption

Demand for wireless and mobile data is increasing along with development of virtual reality (VR), augmented reality (AR), mixed reality (MR), and extended reality (ER) applications. In order to handle ultra-high data exchange rates while offering low latency levels, fifth generation (5G) networks have been proposed. Energy efficiency is one of the key objectives of 5G networks. The notion is defined as the ratio of throughput and total power consumption, and is measured using the number of transmission bits per Joule. In this paper, we review state-of-the-art techniques ensuring good energy efficiency in 5G wireless networks. We cover the base-station on/off technique, simultaneous wireless information and power transfer, small cells, coexistence of long term evolution (LTE) and 5G, signal processing algorithms, and the latest machine learning techniques. Finally, a comparison of a few recent research papers focusing on energy-efficient hybrid beamforming designs in massive multiple-input multiple-output (MIMO) systems is presented. Results show that machine learningbased designs may replace best performing conventional techniques thanks to a reduced complexity machine learning encoder

Sign in / Sign up

Export Citation Format

Share Document