scholarly journals Heating Homes with Servers: Workload Scheduling for Heat Reuse in Distributed Data Centers

Sensors ◽  
2021 ◽  
Vol 21 (8) ◽  
pp. 2879
Author(s):  
Marcel Antal ◽  
Andrei-Alexandru Cristea ◽  
Victor-Alexandru Pădurean ◽  
Tudor Cioara ◽  
Ionut Anghel ◽  
...  
Keyword(s):  
Machine Learning ◽  
Mathematical Model ◽  
Data Center ◽  
Data Centers ◽  
Gradient Boosting ◽  
Distributed Data ◽  
Power Demand ◽  
Heat Demand ◽  
Workload Scheduling ◽  
Correlation Accuracy

Data centers consume lots of energy to execute their computational workload and generate heat that is mostly wasted. In this paper, we address this problem by considering heat reuse in the case of a distributed data center that features IT equipment (i.e., servers) installed in residential homes to be used as a primary source of heat. We propose a workload scheduling solution for distributed data centers based on a constraint satisfaction model to optimally allocate workload on servers to reach and maintain the desired home temperature setpoint by reusing residual heat. We have defined two models to correlate the heat demand with the amount of workload to be executed by the servers: a mathematical model derived from thermodynamic laws calibrated with monitored data and a machine learning model able to predict the amount of workload to be executed by a server to reach a desired ambient temperature setpoint. The proposed solution was validated using the monitored data of an operational distributed data center. The server heat and power demand mathematical model achieve a correlation accuracy of 11.98% while in the case of machine learning models, the best correlation accuracy of 4.74% is obtained for a Gradient Boosting Regressor algorithm. Also, our solution manages to distribute the workload so that the temperature setpoint is met in a reasonable time, while the server power demand is accurately following the heat demand.

Water Utilization in Data Center Infrastructure

2010 ◽  
Author(s):  
Ratnesh Sharma ◽  
Rocky Shih ◽  
Alan McReynolds ◽  
Cullen Bash ◽  
Chandrakant Patel ◽  
...  
Keyword(s):  
Power Generation ◽  
Water Quality ◽  
Water Consumption ◽  
Data Center ◽  
Data Centers ◽  
Cooling Towers ◽  
Power Demand ◽  
Water Usage ◽  
Water Treatment Plants ◽  
End Use

Fresh water is one of the few resources which is scarce and has no replacement; it is also closely coupled to energy consumption. Fresh water usage for power generation and other cooling applications is well known and accounts for 40% of total freshwater withdrawal in the U. S[1]. A significant amount of energy is embedded in the consumption of water for conveyance, treatment and distribution of water. Waste water treatment plants also consume a significant amount of energy. For example, water distribution systems and water treatment plants consume 1.3MWh and 0.5MWh[2], respectively, for every million gallons of water processed. Water consumption in data centers is often overlooked due to low cost impact compared to energy and other consumables. With the current trend towards local onsite generation[3], the role of water in data centers is more crucial than ever. Apart from actual water consumption, the impact of embedded energy in water is only beginning to be considered in water end-use analyses conducted by major utilities[4]. From a data center end-use perspective, water usage can be characterized as direct, for cooling tower operation, and indirect, for power generation to operate the IT equipment and cooling infrastructure[5]. In the past, authors have proposed and implemented metrics to evaluate direct and indirect water usage using an energy-based metric. These metrics allow assessment of water consumption at various power consumption levels in the IT infrastructure and enable comparison with other energy efficiency metrics within a data center or among several data centers[6]. Water consumption in data centers is a function of power demand, outside air temperature and water quality. While power demand affects both direct and indirect water consumption, water quality and outside air conditions affect direct water consumption. Water from data center infrastructure is directly discharged in various forms such as water vapor and effluent from cooling towers. Classification of direct water consumption is one of the first steps towards optimization of water usage. Subsequently, data center processes can be managed to reduce water intake and discharge. In this paper, we analyze water consumption from data center cooling towers and propose techniques to reuse and reduce water in the data center.

scholarly journals Power consumption prediction in cloud data center using machine learning

2020 ◽  
Vol 10 (2) ◽  
pp. 1524 ◽  
Author(s):  
Deepika T. ◽  
Prakash P.
Keyword(s):  
Machine Learning ◽  
Cloud Computing ◽  
Power Consumption ◽  
Data Center ◽  
Data Centers ◽  
Virtual Machines ◽  
Service Providers ◽  
Business World ◽  
Cloud Data ◽  
Cloud Data Centers

The flourishing development of the cloud computing paradigm provides several services in the industrial business world. Power consumption by cloud data centers is one of the crucial issues for service providers in the domain of cloud computing. Pursuant to the rapid technology enhancements in cloud environments and data centers augmentations, power utilization in data centers is expected to grow unabated. A diverse set of numerous connected devices, engaged with the ubiquitous cloud, results in unprecedented power utilization by the data centers, accompanied by increased carbon footprints. Nearly a million physical machines (PM) are running all over the data centers, along with (5 – 6) million virtual machines (VM). In the next five years, the power needs of this domain are expected to spiral up to 5% of global power production. The virtual machine power consumption reduction impacts the diminishing of the PM’s power, however further changing in power consumption of data center year by year, to aid the cloud vendors using prediction methods. The sudden fluctuation in power utilization will cause power outage in the cloud data centers. This paper aims to forecast the VM power consumption with the help of regressive predictive analysis, one of the Machine Learning (ML) techniques. The potency of this approach to make better predictions of future value, using Multi-layer Perceptron (MLP) regressor which provides 91% of accuracy during the prediction process.

The EIDA federator – a one-stop access to EIDA seismic data holdings

2021 ◽  
Author(s):  
Philipp Kaestli ◽  
Daniel Armbruster ◽  
The EIDA Technical Committee
Keyword(s):  
Data Center ◽  
Data Centers ◽  
Data Access ◽  
Distributed Data ◽  
Data Archive ◽  
Waveform Data ◽  
Seismic Waveform ◽  
One Stop ◽  
Single Data ◽  
Seismic Networks

<p>With the setup of EIDA (the European Integrated Data Archive https://www.orfeus-eu.org/data/eida/) in the framework of ORFEUS, and the implementation of FDSN-standardized web services, seismic waveform data and instrumentation metadata of most seismic networks and data centers in Europe became accessible in a homogeneous way. EIDA has augmented this with the WFcatalog service for waveform quality metadata, and a routing service to find out which data center offers data of which network, region, and type. However, while a distributed data archive has clear advantages for maintenance and quality control of the holdings, it complicates the life of researchers who wish to collect data archived across different data centers. To tackle this, EIDA has implemented the “federator” as a one-stop transparent gateway service to access the entire data holdings of EIDA.</p><p>To its users the federator acts just like a standard FDSN dataselect, station, or EIDA WFcatalog service, except for the fact that it can (due to a fully qualified internal routing cache) directly answer data requests on virtual networks.</p><p>Technically, the federator fulfills a user request by decomposing it into single stream epoch requests targeted at a single data center, collecting them, and re-assemble them to a single result.</p><p>This implementation has several technical advantages:</p><ul><li>It avoids response size limitations of EIDA member services, reducing limitations to those imposed by assembling cache space of the federator instance itself.</li> <li>It allows easy merging of partial responses using request sorting and concatenation, and reducing needs to interpret them. This reduces computational needs of the federator and allows high throughput of parallel user requests.</li> <li>It reduces the variability of requests to end member services. Thus, the federator can implement a reverse loopback cache and protect end node services from delivering redundant information and reducing their load.</li> <li>As partial results are quick, and delivered in small subunits, they can be streamed to the user more or less continuously, avoiding both service timeouts and throughput bottlenecks.</li> </ul><p>The advantage of having a one-stop data access for entire EIDA still comes with some limitations and shortcomings. Having requests which ultimately map to a single data center performed by the federator can be slower by that data center directly. FDSN-defined standard error codes sent by end member services have limited utility as they refer to a part of the request only. Finally, the federator currently does not provide access to restricted data.</p><p>Nevertheless, we believe that the one-stop data access compensates these shortcomings in many use cases.</p><p>Further documentation of the service is available with ORFEUS at http://www.orfeus-eu.org/data/eida/nodes/FEDERATOR/</p>

United in Variety: The EarthServer Datacube Federation

2020 ◽  
Author(s):  
Peter Baumann
Keyword(s):  
Data Center ◽  
Data Centers ◽  
Distributed Data ◽  
Temporal Data ◽  
Atmospheric Data ◽  
Elevation Data ◽  
Technology Services ◽  
Spatio Temporal ◽  
Data Location ◽  
Array Databases

<p>Datacubes form an accepted cornerstone for analysis (and visualization) ready spatio-temporal data offerings. Beyond the multi-dimensional data structure, the paradigm also suggests rich services, abstracting away from the untractable zillions of files and products - actionable datacubes as established by Array Databases enable users to ask "any query, any time" without programming. The principle of location-transparent federations establishes a single, coherent information space.</p><p>The EarthServer federation is a large, growing data center network offering Petabytes of a critical variety, such as radar and optical satellite data, atmospheric data, elevation data, and thematic cubes like global sea ice. Around CODE-DE and DIASs an ecosystem of data has been established that is available to users as a single pool, in particular for efficient distributed data fusion irrespective of data location.</p><p>In our talk we present technology, services, and governance of this unique intercontinental line-up of data centers. A live demo will show dist<br>ributed datacube fusion.</p><p> </p>

scholarly journals A Cartesian Genetic Programming Based Parallel Neuroevolutionary Model for Cloud Server’s CPU Usage Prediction

Electronics ◽  
2021 ◽  
Vol 10 (1) ◽  
pp. 67
Author(s):  
Qazi Zia Ullah ◽  
Gul Muhammad Khan ◽  
Shahzad Hassan ◽  
Asif Iqbal ◽  
Farman Ullah ◽  
...  
Keyword(s):  
Machine Learning ◽  
Genetic Programming ◽  
Resource Utilization ◽  
Data Center ◽  
Data Centers ◽  
Electrical Energy ◽  
Learning Models ◽  
Cartesian Genetic Programming ◽  
Cloud Data ◽  
Machine Learning Models

Cloud computing use is exponentially increasing with the advent of industrial revolution 4.0 technologies such as the Internet of Things, artificial intelligence, and digital transformations. These technologies require cloud data centers to process massive volumes of workloads. As a result, the data centers consume gigantic amounts of electrical energy, and a large portion of data center electrical energy comes from fossil fuels. It causes greenhouse gas emissions and thus ensuing in global warming. An adaptive resource utilization mechanism of cloud data center resources is vital to get by with this huge problem. The adaptive system will estimate the resource utilization and then adjust the resources accordingly. Cloud resource utilization estimation is a two-fold challenging task. First, the cloud workloads are sundry, and second, clients’ requests are uneven. In the literature, several machine learning models have estimated cloud resources, of which artificial neural networks (ANNs) have shown better performance. Conventional ANNs have a fixed topology and allow only to train their weights either by back-propagation or neuroevolution such as a genetic algorithm. In this paper, we propose Cartesian genetic programming (CGP) neural network (CGPNN). The CGPNN enhances the performance of conventional ANN by allowing training of both its parameters and topology, and it uses a built-in sliding window. We have trained CGPNN with parallel neuroevolution that searches for global optimum through numerous directions. The resource utilization traces of the Bitbrains data center is used for validation of the proposed CGPNN and compared results with machine learning models from the literature on the same data set. The proposed method has outstripped the machine learning models from the literature and resulted in 97% prediction accuracy.

Techniques for Controlling Airflow Distribution in Raised-Floor Data Centers

2003 ◽  
Author(s):  
Kailash C. Karki ◽  
Suhas V. Patankar ◽  
Amir Radmehr
Keyword(s):  
Mathematical Model ◽  
Data Center ◽  
Heat Load ◽  
Data Centers ◽  
Open Area ◽  
Flow Rates ◽  
Airflow Distribution ◽  
Wide Range ◽  
Optimum Utilization ◽  
Load Pattern

In raised-floor data centers, the airflow rates through the perforated tiles must meet the cooling requirements of the computer servers placed next to the tiles. The data centers house a wide range of equipment, and the heat load pattern on the floor can be quite arbitrary and changes as the data center evolves. To achieve optimum utilization of the floor space and the flexibility for rearrangement and retrofitting, the designers and managers of data centers must be able to modify the airflow rates through the perforated tiles. The airflow rates through the perforated tiles are governed primarily by the pressure distribution under the raised floor. Thus, the key to modifying the flow rates is to influence the flow field in the plenum. This paper discusses a number of techniques that can be used for controlling airflow distribution. These techniques involve changing the plenum height and open area of perforated tiles, and installing thin (solid and perforated) partitions in the plenum. A number of case studies, using a mathematical model, are presented to demonstrate the effectiveness of these techniques.

scholarly journals BHyberCube: A MapReduce aware heterogeneous architecture for data center

2017 ◽  
Vol 14 (3) ◽  
pp. 611-627
Author(s):  
Tao Jiang ◽  
Huaxi Gu ◽  
Kun Wang ◽  
Xiaoshan Yu ◽  
Yunfeng Lu
Keyword(s):  
Mathematical Model ◽  
Data Processing ◽  
Data Center ◽  
Heterogeneous Network ◽  
Distributed Data ◽  
Heterogeneous Architecture ◽  
Processing Application ◽  
Distributed Data Processing ◽  
Network Topologies ◽  
Heterogeneous Nodes

Some applications, like MapReduce, ask for heterogeneous network in data center network. However, the traditional network topologies, like fat tree and BCube, are homogeneous. MapReduce is a distributed data processing application. In this paper, we propose a BHyberCube network (BHC), which is a new heterogeneous network for MapReduce. Heterogeneous nodes and scalability issues are addressed considering the implementation of MapReduce in the existing topologies. Mathematical model is established to demonstrate the procedure of building a BHC. Comparisons of BHC and other topologies show the good properties BHC possesses for MapReduce. We also do simulations of BHC in multi-job injection and different probability of worker servers? communications scenarios respectively. The result and analysis show that the BHC could be a viable interconnection topology in today?s data center for MapReduce.

scholarly journals Managing Energy Consumption in Distributed Data Centers using Genetic algorithm

2019 ◽  
Vol 8 (4) ◽  
pp. 6594-6597
Keyword(s):  
Cloud Computing ◽  
Data Center ◽  
Data Centers ◽  
Operating Cost ◽  
Electrical Energy ◽  
Green Energy ◽  
Energy Utilization ◽  
Distributed Data ◽  
Huge Amount ◽  
Very High

This work shows a multi-target approach for planning vitality utilization in server farms thinking about customary and environmentally friendly power vitality information sources. Cloud computing is a developing innovation. Cloud computing offers administrations such as IaaS, SaaS, PaaS and it gives computing resources through virtualization over data network. Data center consumes huge amount of electrical energy in which it releases very high amount of carbon-di-oxide. The foremost critical challenge in cloud computing is to implement green cloud computing with the help of optimizing energy utilization. The carbon footprint is lowered while minimizing the operating cost. We know that renewable energies that are produced on-site are highly variable and unpredictable but usage of green energy is very important for the mankind using huge amount of single sourced brown energy is not suggested, so our algorithm which evolves genetically and gives practical solution in order to use renewable energy

scholarly journals Modeling Performance of Butterfly Valves Using Machine Learning Methods

2021 ◽  
Vol 13 (24) ◽  
pp. 13545
Author(s):  
Alex Ekster ◽  
Vasiliy Alchakov ◽  
Ivan Meleshin ◽  
Alexandr Larionenko
Keyword(s):  
Machine Learning ◽  
Mathematical Model ◽  
Activated Sludge ◽  
Empirical Model ◽  
Field Testing ◽  
Gradient Boosting ◽  
Machine Model ◽  
Machine Learning Methods ◽  
Gradient Boosting Machine ◽  
Semi Empirical

Control of airflow of activated sludge systems has significant challenges due to the non-linearity of the control element (butterfly valve). To overcome this challenge, some valve manufacturers developed valves with linear characteristics. However, these valves are 10–100 times more expensive than butterfly valves. By developing models for butterfly valves installed characteristics and utilizing these models for real-time airflow control, the authors of this paper aimed to achieve the same accuracy of control using butterfly valves as achieved using valves with linear characteristics. Several approaches were tested to model the installed valve’s characteristics, such as a formal mathematical model utilizing Simscape/Matlab software, a semi-empirical model, and several machine learning methods (MLM), including regression, support vector machine, Gaussian process, decision tree, and deep learning. Several versions of the airflow-valve position models were developed using each machine learning method listed above. The one with the smallest forecast error was selected for field testing at the 55.5×103 m3/day 12 MGD City of Chico activated sludge system. Field testing of the formal mathematical model, semi-empirical model, and the regularized gradient boosting machine model (the best among MLMs) showed that the regularized gradient boosting machine model (RGBMM) provided the best accuracy. The use of the RGBMMs in airflow control loops since 2019 at the City of Chico wastewater treatment plant showed that these models are robust and accurate (2.9% median error).

Sign in / Sign up

Export Citation Format

Share Document