Skip to main content
main-content

Über dieses Buch

This book constitutes the thoroughly refereed post-conference proceedings of the Third International Workshop on Energy Efficient Data Centers, E2DC 2014, held in Cambridge, UK, in June 2014. The 10 revised full papers presented were carefully selected from numerous submissions. They are organized in three topical sections named: energy optimization algorithms and models, the future role of data centres in Europe and energy efficiency metrics for data centres.

Inhaltsverzeichnis

Frontmatter

Energy Optimization Algorithms and Models

Frontmatter

Agile Traffic Merging for DCNs

Abstract
Data center networks (DCNs) have been growing in size and their power consumption is becoming a matter of concern. Many recent papers, including ElasticTree and CARPO, propose new near-energy-proportional DCNs, aiming at reducing the power consumption by dynamically powering off idle network switches and links. In this paper, we examine the power optimization model for DCNs, and present a scalable heuristic algorithm that finds a near-optimal subset of network switches and links that satisfies a given traffic load and consumes minimal power. Furthermore, we apply merge networks to each switch in order to power off the idle interfaces of the active switches, thus further reducing the energy consumption of active switches and achieving greater energy savings than ElasticTree. We finish by simulating large-scale fat-tree DCNs and comparing the energy cost of our techniques versus the ElasticTree method. The results demonstrate that our solution is more energy efficient.
Qing Yi, Suresh Singh

Performance and Energy Efficiency of Parallel Processing in Data Center Environments

Abstract
A novel approach is presented for the analysis of parallel processing of stochastic workload by multi-processor/multi-core processing resources in data center environments. The method is based on job workload descriptions by task graphs with generally-distributed task execution times and task scheduling under consideration of prescribed precedence and synchronization constraints. For the analytic performance evaluation, task graphs are restricted to the analysis of directed acyclic graphs which are reduced by stepwise aggregations of tasks. The reduction allows to aggregate the whole task graph under a given number n of processing elements to a single virtual job processing time with average value hv and coefficient of variation cv. By this way, the whole multi-processor system can be modeled by a queuing system of type GI/G/1 from which the response time TR and the speedup factor S(n) is derived. Finally, the influence of the stochastic properties of the workload on the performance and on energy efficiency of parallel computing will be studied and compared with serial computing on a multi-processor system modeled by a queuing system of the type M/G/n.
Paul J. Kuehn

Cyclic Blackout Mitigation Through HVAC Shifted Queue Optimization

Abstract
The increasing global demand for power has resulted in frequent blackouts in many geographies. The cost of domestic standby generation is prohibitive and novel strategies to provision measures that manage blackouts are becoming much sought after. In some scenarios certain amounts of surplus power can be identified, with the mix of available generation not being fully utilized. The paper presents a strategy that harnesses the aggregated superfluous power to fulfil essential demand in residential areas during cyclic blackouts. The solution has at its foundation, a multi-agent distributed demand management system with a supply-demand matching capability. Power is not distributed fairly to each user, and appliances which consume the most significant levels of power such as air conditioners are serviced according to the available superfluous power. The approach is evaluated through an extensive emulation framework and results show that the proposed system is capable of providing an acceptable Quality-of-Service (QoS) level during cyclic blackout periods and at the same time succeeds in smoothing demand profiles.
Kasim Al-Salim, Ivan Andonovic, Craig Michie

Stochastic Petri Net Models for the Analysis of Trade-Offs in Data Centres with Power Management

Abstract
Due to the growth in energy consumption of data centres, the demand for optimal usage of servers has become a relevant topic. This paper contributes to the early design phases of data centres by providing insight into the power-performance trade-off that arises from power management. This paper proposes a flexible set of stochastic Petri net models which can be used easily to study the trade-off between performance and power consumption.
Björn F. Postema, Boudewijn R. Haverkort

The Future Role of Data Centres in Europe

Frontmatter

GEYSER: Enabling Green Data Centres in Smart Cities

Abstract
Information Technology is a dominant player of our modern societies; Data Centres, lying at the heart of the IT landscape, have attracted attention, with their increasing energy consumption being a constant topic of concern, especially when it comes to the negative impact on the quality of their surrounding environment. Nevertheless, recent technological and societal advances are paving the way for DCs to change their role from passive energy consumers into prosumers, thus, transforming themselves into leading players within their smart district surroundings. This paper describes the innovative GEYSER approach to enabling green networked DCs to monitor, control, reuse, and optimize both their energy consumption and production, and in particular from renewable resources, towards becoming active participants within Smart Grids and Smart Cities.
Ionut Anghel, Massimo Bertoncini, Tudor Cioara, Marco Cupelli, Vasiliki Georgiadou, Pooyan Jahangiri, Antonello Monti, Seán Murphy, Anthony Schoofs, Terpsi Velivassaki

Analysis of the Influence of Application Deployment on Energy Consumption

Abstract
Energy efficiency for data centers has been recently an active research field. Several efforts have been made at the infrastructure and application levels to achieve energy efficiency and reduction of CO2 emissions. In this paper we approach the problem of application deployment to evaluate its impact on the energy consumption of applications at runtime. We use queuing networks to model different deployment configurations and to perform quantitative analysis to predict application performance and energy consumption. The results are validated against experimental data to confirm the correctness of the models when used for predictions. Comparisons between different configurations in terms of performance and energy consumption are made to suggest the optimal configuration to deploy applications on cloud environments.
Marco Gribaudo, Thi Thao Nguyen Ho, Barbara Pernici, Giuseppe Serazzi

Minimization of Costs and Energy Consumption in a Data Center by a Workload-Based Capacity Management

Abstract
In this paper we present an approach to improve power and cooling capacity management in a data center by taking into account knowledge about applications and workloads. We apply power capping techniques and proper cooling infrastructure configuration to achieve savings in energy and costs. To estimate values of a total energy consumption and costs we simulate both IT software/hardware and cooling infrastructure at once using the CoolEmAll SVD Toolkit. We also investigated the use of power capping to adjust data center operation to variable power supply and pricing. By better adjusting cooling infrastructure to specific types of workloads, we were able to find a configuration which resulted in energy, OPEX and CAPEX savings in the range of 4–25 %.
Georges Da Costa, Ariel Oleksiak, Wojciech Piatek, Jaume Salom, Laura Sisó

Building Application Profiles to Allow a Better Usage of the Renewable Energies in Data Centres

Abstract
Data centres are powerful and power-hungry facilities which aim at hosting ICT services. The current trend is to, on the one hand, try to reduce the overall consumption of a data centre, and on the other hand to prioritize the utilization of renewable energies over brown energies. Renewable energies tend to be very variable in time (e.g. solar energy), and thus renewable energy aware algorithms tries to schedule the applications running in the data centres accordingly. However, one of the main problems is that most of the time very little information is known about the applications running in data centres. More specifically, we need to have more information about the current and planned workload of an application, and the tolerance of that application to have its workload rescheduled. In this paper, we present a work in progress on Plug4Green, a flexible VM manager able to reduce energy consumption in data centres. We extend Plug4Green with the second goal of increasing the usage of renewable energy in data centres. This includes the development of specific application profiles, and a new optimization technique.
Corentin Dupont

Energy Efficiency Metrics for Data Centres

Frontmatter

Review on Performance Metrics for Energy Efficiency in Data Center: The Role of Thermal Management

Abstract
Energy consumption and thermal performance are the two most important tasks in data centers (DCs) facility management. In recent years, to monitor and control their variation several performance metrics were introduced. In this paper an overview on the main important energy and thermal metrics is provided. A critical analysis to investigate mutual relations among metrics was performed, with the aim to clarify some physical aspects regarding the assessment of DC global energy performance.
Indeed, although these metrics are commonly used to assess the energy efficiency of DCs, their usefulness for encouraging lower energy consumption was poorly investigated. Moreover, an analysis on the effect of the DC thermal performance on metrics was done. The thermal management assume a key role for achieving energy saving during the operation of a DC and for the improvement of the IT equipment reliability.
Alfonso Capozzoli, Marta Chinnici, Marco Perino, Gianluca Serale

Gain More from PUE: Assessing Data Center Infrastructure Power Adaptability

Abstract
The power usage effectiveness (PUE) for data centers is used by operators as KPI to measure the absolute infrastructure power overhead. However, this only draws conclusions on static or average operation conditions during an usual annual time period. For analyzing the aspect of dynamics in the IT to infrastructure power relation, we propose two new metrics. First, the power variability (PVar). It simply indicates the relative rates and heights of power variations. Second, the infrastructure power adaptability (IPA). It relates the power variabilities and relative average deviations of IT and infrastructure power in order to represent the scalability and adaptability of the infrastructure to the IT demands. Both metrics use the same input data also needed for a continuous PUE calculation. Thus, the applicability in a data center running a PUE-metering can be ensured. In an evaluation, we applied the IPA on power traces of a container data center (in the following denoted as CDC) and compared the results with PUE scalability, a metric with the same scope. The comparison showed, that IPA covers more operating states and is therefore more robust and reliable than its counterpart.
Daniel Schlitt, Gunnar Schomaker, Wolfgang Nebel

Backmatter

Weitere Informationen

Premium Partner

    Bildnachweise