Download pdf Advances And New Trends In Environmental And Energy Informatics Selected And Extended Contributions From The 28Th International Conference On Informatics For Environmental Protection 1St Edition Jorge ebook full chapter

Advances and New Trends in
Environmental and Energy Informatics

Selected and Extended Contributions
from the 28th International Conference
on Informatics for Environmental
Protection 1st Edition Jorge Marx
Gomez
Visit to download the full and correct content document:
https://textbookfull.com/product/advances-and-new-trends-in-environmental-and-ener
gy-informatics-selected-and-extended-contributions-from-the-28th-international-confer
ence-on-informatics-for-environmental-protection-1st-edition-jorge/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Advances and New Trends in Environmental Informatics

Stability Continuity Innovation 1st Edition Volker
Wohlgemuth
https://textbookfull.com/product/advances-and-new-trends-in-
environmental-informatics-stability-continuity-innovation-1st-
edition-volker-wohlgemuth/
Information Technology in Environmental Engineering

Proceedings of the 7th International Conference on
Information Technologies in Environmental Engineering
ITEE 2015 1st Edition Jorge Marx Gómez
https://textbookfull.com/product/information-technology-in-
environmental-engineering-proceedings-of-the-7th-international-
conference-on-information-technologies-in-environmental-
engineering-itee-2015-1st-edition-jorge-marx-gomez/
Multimedia Tools and Applications for Environmental

Biodiversity Informatics Alexis Joly
https://textbookfull.com/product/multimedia-tools-and-
applications-for-environmental-biodiversity-informatics-alexis-
joly/
Advances in Neurotechnology Electronics and Informatics

Revised Selected Papers from the 2nd International
Congress on Neurotechnology Electronics and Informatics
NEUROTECHNIX 2014 October 25 26 Rome Italy 1st Edition
Ana Rita Londral
https://textbookfull.com/product/advances-in-neurotechnology-
electronics-and-informatics-revised-selected-papers-from-the-2nd-
international-congress-on-neurotechnology-electronics-and-
Advances in Energy and Environmental Materials Yafang
Han
https://textbookfull.com/product/advances-in-energy-and-
environmental-materials-yafang-han/
Informatics in Schools Improvement of Informatics

Knowledge and Perception 9th International Conference
on Informatics in Schools Situation Evolution and
Perspectives ISSEP 2016 Münster Germany October 13 15
2016 Proceedings 1st Edition Andrej Brodnik
https://textbookfull.com/product/informatics-in-schools-
improvement-of-informatics-knowledge-and-perception-9th-
international-conference-on-informatics-in-schools-situation-
evolution-and-perspectives-issep-2016-munster-germany-octobe/
Advances in Visual Informatics 6th International Visual

Informatics Conference IVIC 2019 Bangi Malaysia
November 19 21 2019 Proceedings Halimah Badioze Zaman
https://textbookfull.com/product/advances-in-visual-
informatics-6th-international-visual-informatics-conference-
ivic-2019-bangi-malaysia-november-19-21-2019-proceedings-halimah-
badioze-zaman/
International Conference on Mobile Computing and

Sustainable Informatics : ICMCSI 2020 Jennifer S. Raj
https://textbookfull.com/product/international-conference-on-
mobile-computing-and-sustainable-informatics-
icmcsi-2020-jennifer-s-raj/
Advances in Materials Science for Environmental and

Energy Technologies VI 1st Edition Tatsuki Ohji
https://textbookfull.com/product/advances-in-materials-science-
for-environmental-and-energy-technologies-vi-1st-edition-tatsuki-
ohji/
Progress in IS
Jorge Marx Gómez

Michael Sonnenschein
Ute Vogel
Andreas Winter
Barbara Rapp
Nils Giesen Editors
Advances and
New Trends in
Environmental and
Energy Informatics
Selected and Extended Contributions
from the 28th International Conference
on Informatics for Environmental
Protection
Progress in IS
More information about this series at http://www.springer.com/series/10440
omez • Michael Sonnenschein •
Jorge Marx G
Ute Vogel • Andreas Winter • Barbara Rapp •
Nils Giesen
Editors
Advances and New Trends

in Environmental and Energy
Informatics
Selected and Extended Contributions from
the 28th International Conference on
Informatics for Environmental Protection
Editors
Jorge Marx Gomez Michael Sonnenschein
University of Oldenburg University of Oldenburg
Oldenburg, Germany Oldenburg, Germany
Ute Vogel Andreas Winter

Barbara Rapp Nils Giesen

ISSN 2196-8705 ISSN 2196-8713 (electronic)

Progress in IS
ISBN 978-3-319-23454-0 ISBN 978-3-319-23455-7 (eBook)
DOI 10.1007/978-3-319-23455-7
Library of Congress Control Number: 2015957811
Springer Cham Heidelberg New York Dordrecht London

© Springer International Publishing Switzerland 2016
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or
dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained
herein or for any errors or omissions that may have been made.
Printed on acid-free paper
Springer International Publishing AG Switzerland is part of Springer Science+Business Media

(www.springer.com)
We dedicate this book to all people whose
lives were taken and affected by terror,
repression and violence inside and outside
of universities around the world.
ThiS is a FM Blank Page
Foreword
This book is a collection of selected outstanding scientific papers based on contri-

butions to the 28th International Conference on Informatics for Environmental
Protection – EnviroInfo 2014 – which took place in September 2014 at the
Oldenburg University (Germany). The EnviroInfo conference series aims at
presenting and discussing the latest state-of-the-art development on information
and communications technology (ICT) in environmental-related fields. Selected
papers out of the EnviroInfo 2014 conference have been extended to full book
chapters by the authors and were reviewed each by three members of a 37-person
program committee to improve the contributions. This book outlines some of the
major topics in the field of ICT for energy efficiency and further relevant environ-
mental topics addressed by the EnviroInfo 2014. It includes concepts, methods,
approaches, and applications of emerging fields of research and practice in ICT
dedicated to environmental and sustainability topics and problems.
The book provides and covers a broad range of topics. It is structured in six
application-oriented main clusters comprising all chapters:
• Green IT, aiming to design, implement, and use ICT in an environment-friendly
and resource-preserving (efficient) way considering their whole life cycle
• From Smart Grids to Smart Homes, providing technical approaches and
systems with the aim of an efficient management of power generation from
renewable energy sources as well as an efficient power consumption
• Smart Transportation, aiming to provide innovative services to various modes
of transport and traffic management
• Sustainable Enterprises and Management, aiming to provide various ele-
ments involved in managing and organization’s operations, with a strong empha-
sis on maintaining socio-environmental integrity
• Environmental Decision Support, aiming to inform environmental and natural
resource management based on techniques for objective assessments of options
and their outcomes
vii
viii Foreword
• Social Media for Sustainability, becoming an indispensable tool for value

creation in education and sustainable business with community engagement
for sustainability
Crucial elements in the book clusters are the criteria environmental protection,
sustainability, and energy as well as resource efficiency which are spread all over
the book chapters.
Chapters allocated in the “Green IT” cluster deal mainly with the topics energy
awareness, resource optimization, energy efficiency, and carbon footprint improve-
ment as well as simulation approaches. Three chapters are dedicated to an opti-
mized operation of data centers with respect to energy consumption. Modeling and
simulation of data center’s operations are essential tools for this purpose. Two
chapters are oriented towards the energy efficiency of mobile and context-aware
applications.
The cluster “From Smart Grids to Smart Homes” comprises chapters concerning
the management of renewables in power grids, IT architecture analysis for smart
grids, control algorithms in smart grids, as well as energy efficiency of buildings
and eco-support features at home. First, an approach for increasing distribution grid
capacity for connecting renewable generators is presented. The next two chapters
address architectures and control algorithms for smart grids. After dealing with an
approach to urban energy planning demonstrated on the example of a district in
Vienna, a new method for reducing energy consumption at home is presented.
In the “Smart Transportation” cluster, chapters present and discuss themes on
sustainable mobility supported by mobile technologies to inform on air quality,
safer bicycling through spatial information on road traffic, and system dynamics
modeling to understand rebound effects in air traffic.
Authors of chapters allocated to “Sustainable Enterprises and Management”
argue on web-based software tools for sustainability management for small enter-
prises, while the others focus on green business process management and IT for
green improvements for bridging the gap from strategic planning to everyday work.
Contributions of chapters in the cluster “Environmental Decision Support” are
assigned to the thematic fields of supporting environmental policy making by
specific software tools, enrichment of environmental data streams by use of seman-
tic web technologies, and quality improvement of river flood prediction by enhanc-
ing flood simulation.
The remaining “Social Media for Sustainability” cluster includes a contribution
to sustainable practices in educational environments by means of social media and a
contribution on sustainable development in rural areas by neighborhood effects
using ICT. Web technologies are the key element to improve environmental
awareness and communication between different stakeholders.
We strongly believe that this book will contribute to the increasing awareness of
researchers and practitioners in the field of ICT and sustainability, and we hope that
it can reach a wide international audience and readership.
Reviewing Committee
• Hans-Knud Arndt, University of Magdeburg, Germany

• J€
org Bremer, University of Oldenburg, Germany
• Thomas Brinkhoff, Jade-Hochschule, Wilhelmshaven/Oldenburg/Elsfleth,
Germany
• Christian Bunse, Fachhochschule Stralsund, Germany
• Luis Rafael Canali, National University of Technology Cordoba, Argentina
• Lester Cowley, Nelson Mandela Metropolitan University, Port Elizabeth &
George, South Africa
• Clemens Düpmeier, Karlsruhe Institute of Technology, Germany
• Amr Eltaher, University of Duisburg, Germany
• Luis Ferreira, Polytechnic Institute of Cávado and Ave, Portugal
• Werner Geiger, Karlsruhe Institute of Technology, Germany
• Nils Giesen, University of Oldenburg, Germany
• Albrecht Gnauck, HTW Berlin, Germany
• Johannes G€ obel, University of Hamburg, Germany
• Paulina Golinska, University of Poznan, Poland
• Marion Gottschalk, OFFIS Institute for Information Technology,
Oldenburg, Germany
• Klaus Greve, University of Bonn, Germany
• Axel Hahn, University of Oldenburg, Germany
• Oliver Kramer, University of Oldenburg, Germany
• Nuno Lopes, Polytechnic Institute of Cávado and Ave, Portugal
• Somayeh Malakuti, University of Technology, Dresden, Germany
• Jorge Marx G omez, University of Oldenburg, Germany
• Ammar Memari, University of Oldenburg, Germany
• Andreas M€ oller, Leuphana University, Lüneburg, Germany
• Stefan Naumann, Hochschule Trier, Germany
• Alexandra Pehlken, University of Oldenburg, Germany
• Joachim Peinke, University of Oldenburg, Germany
ix
x Reviewing Committee
• Werner Pillmann, International Society for Environmental Protection, Vienna,

Austria
• Barbara Rapp, University of Oldenburg, Germany
• Wolf-Fritz Riekert, HDM Stuttgart, Germany
• Brenda Scholtz, Nelson Mandela Metropolitan University, Port Elizabeth &
George, South Africa
• Karl-Heinz Simon, University of Kassel, Germany
• Michael Sonnenschein, University of Oldenburg, Germany
• Frank Teuteberg, University of Osnabrück, Germany
• Ute Vogel, University of Oldenburg, Germany
• Benjamin Wagner vom Berg, University of Oldenburg, Germany
• Andreas Winter, University of Oldenburg, Germany
• Volker Wohlgemuth, HTW Berlin, Germany
Acknowledgments
First of all, we would like to thank the authors for elaborating and providing their
book chapters in time and in high quality. The editors would like to express their
sincere gratitude to all reviewers who devoted their valuable time and expertise in
order to evaluate each individual chapter. Furthermore, we would like to extend our
gratitude to all sponsors of the EnviroInfo 2014 conference. Without their financial
support, this book would not have been possible. Finally, our thanks go to Springer
Publishing House for their confidence and trust in our work.
Oldenburg, May 2015.
Jorge Marx Gomez
Michael Sonnenschein
Ute Vogel
Andreas Winter
Barbara Rapp
Nils Giesen
xi
ThiS is a FM Blank Page
Contents
Part I Green IT
1 Extending Energetic Potentials of Data Centers by Resource
Optimization to Improve Carbon Footprint . . . . . . . . . . . . . . . . . . 3
Alexander Borgerding and Gunnar Schomaker
2 Expansion of Data Centers’ Energetic Degrees of Freedom to
Employ Green Energy Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Stefan Janacek and Wolfgang Nebel
3 A Data Center Simulation Framework Based on an Ontological
Foundation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Ammar Memari, Jan Vornberger, Jorge Marx Gomez,
and Wolfgang Nebel
4 The Contexto Framework: Leveraging Energy Awareness in the
Development of Context-Aware Applications . . . . . . . . . . . . . . . . . 59
Maximilian Schirmer, Sven Bertel, and Jonas Pencke
5 Refactorings for Energy-Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . 77
Marion Gottschalk, Jan Jelschen, and Andreas Winter
Part II From Smart Grids to Smart Homes

6 The 5 % Approach as Building Block of an Energy System
Dominated by Renewables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
Enno Wieben, Thomas Kumm, Riccardo Treydel, Xin Guo, Elke Hohn,
Till Luhmann, Matthias Rohr, and Michael Stadler
7 Aligning IT Architecture Analysis and Security Standards for
Smart Grids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Mathias Uslar, Christine Rosinger, Stefanie Schlegel,
and Rafael Santodomingo-Berry
xiii
xiv Contents
8 Design, Analysis and Evaluation of Control Algorithms for

Applications in Smart Grids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
Christian Hinrichs and Michael Sonnenschein
9 Multi-actor Urban Energy Planning Support: Building
Refurbishment & Building-Integrated Solar PV . . . . . . . . . . . . . . . 157
Najd Ouhajjou, Wolfgang Loibl, Stefan Fenz, and A. Min Tjoa
10 Beyond Eco-feedback: Using Room as a Context to Design
New Eco-support Features at Home . . . . . . . . . . . . . . . . . . . . . . . . 177
Nico Castelli, Gunnar Stevens, Timo Jakobi, and Niko Sch€onau
Part III Smart Transportation

11 Supporting Sustainable Mobility Using Mobile Technologies and
Personalized Environmental Information: The Citi-Sense-MOB
Approach in Oslo, Norway . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
Núria Castell, Hai-Ying Liu, Franck R. Dauge, Mike Kobernus,
Arne J. Berre, Josef Noll, Erol Cagatay, and Reidun Gangdal
12 Spatial Information for Safer Bicycling . . . . . . . . . . . . . . . . . . . . . 219
Martin Loidl
13 Using Systems Thinking and System Dynamics Modeling to
Understand Rebound Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
Mohammad Ahmadi Achachlouei and Lorenz M. Hilty
Part IV Sustainable Enterprises and Management

14 Software and Web-Based Tools for Sustainability Management
in Micro-, Small- and Medium-Sized Enterprises . . . . . . . . . . . . . . 259
Matthew Johnson, Jantje Halberstadt, Stefan Schaltegger,
and Tobias Viere
15 Towards Collaborative Green Business Process Management
as a Conceptual Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
Timo Jakobi, Nico Castelli, Alexander Nolte, Niko Sch€onau,
and Gunnar Stevens
Part V Environmental Decision Support

16 A Generic Decision Support System for Environmental Policy
Making: Attributes, Initial Findings and Challenges . . . . . . . . . . . 297
Asmaa Mourhir, Tajjeeddine Rachidi, and Mohammed Karim
17 Towards an Environmental Decision-Making System:
A Vocabulary to Enrich Stream Data . . . . . . . . . . . . . . . . . . . . . . . 317
Peter Wetz, Tuan-Dat Trinh, Ba-Lam Do, Amin Anjomshoaa,
Elmar Kiesling, and A Min Tjoa
Contents xv
18 How a Computational Method Can Help to Improve the Quality

of River Flood Prediction by Simulation . . . . . . . . . . . . . . . . . . . . . 337
Adriana Gaudiani, Emilio Luque, Pablo Garcı́a, Mariano Re,
Marcelo Naiouf, and Armando De Giusti
Part VI Social Media for Sustainability

19 A Social Media Environmental Awareness Campaign to Promote
Sustainable Practices in Educational Environments . . . . . . . . . . . . 355
Brenda Scholtz, Clayton Burger, and Masive Zita
20 Supporting Sustainable Development in Rural Areas by
Encouraging Local Cooperation and Neighborhood Effects
Using ICT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371
Andreas Filler, Eva Kern, and Stefan Naumann
Part I
Green IT
Chapter 1
Extending Energetic Potentials of Data
Centers by Resource Optimization
to Improve Carbon Footprint
Alexander Borgerding and Gunnar Schomaker
Abstract The electric power is one of the major operating expenses in data centers.
Rising and varying energy costs induce the need of further solutions to use energy
efficiently. The first steps to improve efficiency have already been accomplished by
applying virtualization technologies. However, a practical approach for data center
power control mechanisms is still missing.
In this paper, we address the problem of energy efficiency in data centers.
Efficient and scalable power usage for data centers is needed. We present different
approaches to improve efficiency and carbon footprint as background information.
We propose an in-progress idea to extend the possibilities of power control in data
centers and to improve efficiency. Our approach is based on virtualization technol-
ogies and live-migration to improve resource utilization by comparing different
effects on virtual machine permutation on physical servers. It delivers an efficiency-
aware VM placement by assessing different virtual machine permutation. In our
approach, the applications are untouched and the technology is non-invasive
regarding the applications. This is a crucial requirement in the context of Infra-
structure-as-a-Service (IaaS) environments.
Keywords Data center • VM placement • Energy efficiency • Power-aware •

Resource management • Server virtualization
1 Introduction
The IP traffic increases year by year worldwide. New Information and Communi-
cation Technology (ICT) services are coming up and existing services are migrating
to IP technology, for example, VoIP, TV, radio and video streaming. Following
A. Borgerding (*)
University of Oldenburg, 26111 Oldenburg, Germany
e-mail: [email protected]
G. Schomaker
Software Innovation Campus Paderborn, Zukunftsmeile 1, 33102 Paderborn, Germany
e-mail: [email protected]
© Springer International Publishing Switzerland 2016 3

J. Marx Gomez et al. (eds.), Advances and New Trends in Environmental and
Energy Informatics, Progress in IS, DOI 10.1007/978-3-319-23455-7_1
4 A. Borgerding and G. Schomaker
these trends, the power consumption of ICT obtains a more and more significant
value. In the same way, data centers are growing in number and size in order to
comply with the increasing demand. As a result, their share of electric power
consumption increases too, e.g. it has doubled in the period 2000–2006 [16]. In
addition, energy costs rise continuously and the data center operators are faced with
customer questions about sustainability and carbon footprint while economical
operation is an all-over goal. The electric power consumption has become one of
the major expenses in data centers.
A high performance server in idle-state consumes up to 60 % of its peak power
[11]. To reduce the quantity of servers in idle-state, virtualization technologies are
used. Virtualization technologies allow several virtual machines (VMs) to be
operated on one physical server or machine (PM). In this way the number of servers
in idle-state can be reduced to save energy [6]. However, the rising energy costs
lead to a rising cost pressure and further solutions are needed as they will be
proposed in the following.
This paper extends our contribution to EnviroInfo 2014 – 28th International
Conference on Informatics for Environmental Protection [3] and is organized as
follows: Sect. 2 motivates and defines the problem of energy efficiency and
integrating renewable energy in data centers. Section 3 gives background on
approaches relevant to energy efficiency, virtualization technology and improving
the carbon footprint. In Sect. 4, we present the resource-efficient and energy-
adaptive approach. The paper is concluded by comments on our progressing work
in Sect. 5.
2 Problem Definition
The share of volatile renewable power sources is increasing. This leads to volatile
energy availability and lastly to varying energy price models. To deal with the
variable availability, we need an approach that ensures controllable power con-
sumption beyond general energy efficiency. Thus, we need to improve the effi-
ciency of the data center using an intelligent and efficient VM placement in order to
adapt to volatile energy availability and improve carbon footprint while keeping the
overall goal to use the invested energy as efficient as possible.
The increasing amount of IT services combined with steadily raising energy
costs place great demands on data centers. These conditions induce the need to
operate a maximum number of IT services with minimal employment of resources,
since the aim is an economical service operation. Therefore, the effectiveness of the
invested power should be at a maximum level. In this paper, we focus on the
server’s power consumption and define the efficiency of a server as the work done
per energy unit [5].
In the related work part of this paper, we analyze different kinds of approaches in
the context of energy consumption, energy efficiency and integrating renewable
power. In this research approach, we want to explore which further options exist to
1 Extending Energetic Potentials of Data Centers by Resource Optimization to. . . 5
use energy efficient and how we can take effect on the data center’s power
consumption and, finally, to adapt it to available volatile renewable energy.
To take advantage of current developments, power consumption should be
increasable in times of low energy prices and reducible otherwise while we stick
to a high efficiency level in both cases. In Service Level Agreements (SLAs) for
instance, a specific application throughput within a time frame is defined. Due to
these agreements, we can use periods of low energy prices to produce the through-
put far before the time frame exceeds. In periods of high energy prices, a scheduled
decrease of the previously built buffer can be used to save energy costs.
Some approaches [4, 10, 15] use geographically-distributed data centers to
schedule the workload across data centers with high renewable energy availability.
The methodology is only suitable in big, geographically-spread scenarios and the
overall power consumption is not affected. Hence, we do not pursue these
approaches. In general, many approaches are based on strategies with focus on
CPU utilization because CPU utilization correlates with the server’s power con-
sumption directly [5]. The utilization of other server components does not have
such an effect on the server’s power consumption. However, the application’s
performance depends not only on CPU usage, but all required resources are needed
for optimal application performance. Hence, the performance relies on other com-
ponents too and we also want to focus on these other components such as Network
Interface Card (NIC), Random Access Memory (RAM) and Hard Disk Drive
(HDD) to improve the efficiency, especially if their utilization does not have an
adverse effect on the server’s power consumption. Our assumption is that the
optimized utilization of these resources is not increasing the power consumption,
but it can be used to improve the efficiency and application performance.
There are different types of applications; some applications work stand-alone
while others rely on several components running on different VMs. Components of
the latter communicate via network and the network utilization takes effect on such
distributed applications. In our approach, we want to include these communication
topology topics. However, the applications’ requirements are changing during
operation, sometimes in large scale and in short intervals. Therefore, we need an
online algorithm that acts at runtime to respond to changing values. We need to
keep obstacles at a low level by acting agnostic to the applications. The capable
approach should be applicable without the need to change the operating applica-
tions. This is a crucial requirement in the context of Infrastructure-as-a-Service
(IaaS) environments.
Being agnostic to applications means to influence their performance without
they become aware of our methodology. For example, if an application intends to
write a file on the hard disk, it has to wait until it gets access to the hard disk. This is
a usual situation an application can handle. In the wait state, the application cannot
distinguish whether the wait was caused by another application writing on the hard
disk or by our methodology.
The problem of determining an efficient VM placement can be formulated as an
extended bin-packing problem, where VMs (objects) must be allocated to the PMs
(bins). In the bin-packing problem, objects of different volumes must be fitted into a
finite number of bins, each of the same volume, in a way that minimizes the number
of bins used. The bin-packing problem has an NP-hard complexity. Compared to
the VM allocating problem, we have a multidimensional bin packing problem.
Instead of the object size, we have to deal with several resource requirements
of VMs.
In a data center with k PMs and n VMs operated on the PMs, the number of
configuration possibilities is described by partioning a set of n elements into k
partitions while the k sets are disjoint and nonempty. This is described by the
Stirling numbers of the second kind:

1 X k
k j k
Sn , k ¼ ð1Þ jn
k! j¼0 j
In case of a data center with 10 VMs and 3 PMs, we have S10, 3 ¼ 9330 different and
possible VM allocations to the PMs that are named as configurations in this paper.
Hence, a global bin-packing solver will not be able to deliver a VM placement
for a fast acting online approach.
The formal description of the VM placing problem relating to the bin-packing
problem is as follows: A set of virtual machines V ¼ fVM1 , . . . , VMn g and a set of
physical machines P ¼ fPM1 , . . . , PMk g is given. The VMs are represented by
their resource demand vectors di . The PMs are represented by their resource
capacity vectors cs . The resource capacity vector of a PM describes the available
resources that can be requested by VMs. The goal is to find a configuration so that
for all PMs in P:
X
j
d i cs
i¼1
while j is the total number of VMs on the PM.

To measure the quality of an allocated configuration C, the efficiency E defined
by:
work done
EðCÞ ¼
unit energy
is a suitable metric [5]. The aggregated idle times of the PMs may also indicate the
quality of the configuration.
To the best of our knowledge, this is the first approach that researches on
agnostic methodologies, without scheduling components, to control the data centers
power consumption with the aim of efficiency and the possibility to increase and
decrease the power consumption as well.
3 Related Work
Power consumption and energy efficiency in data centers is a topic, on which a lot
of work has already been done. In this section, we give an overview of different
approaches.
The usage of low-power components seems to offer solutions for lower energy
consumption. Meisner et al. [12] handled the question whether low power con-
sumption correlates with energy efficiency in the data center context. They discov-
ered that the usage of low power components is not the solution. They compared
low power servers with high power servers and defined the energy efficiency of a
system as the work done per energy unit. They achieved better efficiency with the
high power servers and found that modern servers are only maximally efficient at
100 % utilization.
Another potential for improvement is to let IT requirements follow energy
availability. There are some approaches [4, 10, 14] that use local energy conditions.
They migrate the server workload to data center destinations with available renew-
able power. These ideas are finally only suitable for distributed and widespread data
centers. Data center locations at close quarters typically have the same or not
significantly different energy conditions. In the latter scenario, the consumption
of renewable energy can be increased, but the efficient power usage is not taken into
consideration.
A different idea is mentioned by Krioukov et al. [9]. In this work, a scheduler has
access to a task list, where the task with the earliest deadline is at the top. This is an
earliest deadline first (EDF) schedule. If renewable energy is available, the EDF
scheduler starts tasks from the top of the task list to use the renewable energy. If less
energy is available, tasks will be terminated. In such approaches, we have to deal
with application-specific topics. To build a graded list of tasks to schedule, we
determine the duration a task needs to be processed and we need a deadline for each
task to be processed. Terminated tasks lead to application-specific issues that need
to be resolved afterwards.
The approach of Hoyer [8] bases on prediction models to calculate the needed
server capacity in advance to reduce unused server capacity. Optimistic, pessimistic
and dynamic resource strategies were presented. This approach offers methodolo-
gies to improve efficiency, but controlling the data centers power consumption is
not focused.
Tang et al. [17] propose a thermal-aware task scheduling. The ambition is to
minimize cooling requirements and to improve the data center efficiency in this
way. They set up a central database with server information, especially server heat
information. An EDF scheduler is placing tasks with the earliest deadline on the
coldest server. Thus, they avoid hot spots and cooling requirement can be decreased
to improve efficiency. The usage of a graded task list comes with the same
disadvantages as described before. To avoid dealing with application-specific
topics, the virtual machine is a useful container to place IT loads instead of explicit
application tasks. In many approaches, for example Corradi et al. [6], power
consumption is reduced by concentrating VMs on a fewer number of servers and

powering down unused ones to save energy. Chen et al. [5] describe the power
consumption of a server as the sum of its static power consumption and its dynamic
power consumption. The static power consumption is the consumption of the server
in power-on state without workload. This amount of power can be saved with this
approach. The dynamic part of server’s power consumption correlates with its CPU
utilization, as described by Pelley et al. [13]. Thus, most methodologies are only
focused on CPU utilization.
Dalvanadi et al. [7] and Vu et al. [18] pointed out that network communication
can also influence the overall performance of an IT service and network-aware VM
placement is also an important and challenging issue. Hence, they embrace network
traffic to minimize power consumption.
As described, many approaches [1, 15, 19] use virtualization technologies to
concentrate VMs on a small number of PMs. While migrating VMs onto a PM, the
size of the RAM is a limiting factor. If the RAM-size of the PM is exhausted, further
VMs cannot be migrated onto this PM. This can be an adverse effect, especially if
resources such as CPUs are still underutilized or completely idling. The memory
sharing technology offers the possibility to condense redundant memory pages on a
PM to one page. Unneeded physical memory can be freed to improve the VMs
memory footprint. The VMs run on top of a hypervisor, which is responsible for
allocating the physical resources to individual VMs. The hypervisor identifies
identical memory pages on the different VMs on a PM and shares them among
the VMs with pointers. This frees up memory for new pages. If a VM’s information
on that shared page changes, the hypervisor writes the memory to a new page and
re-addresses a pointer. The capacity of the PM can be increased to concentrate
further VMs on the PM and to achieve higher server utilization. Wood et al. [19]
present a memory sharing-aware placement approach for virtual machines that
includes a memory fingerprinting system to determine the sharing potential
among a set of VMs. In addition, it makes use of live migration to optimize the
VM placement.
In summary, the state of the art approaches deliver several solutions in the
context of energy efficiency, but an efficiency-aware approach with combined
data center power control mechanisms is still missing.
4 Resource-Efficient and Energy-Adaptive VM Placement

Approach
In this section the in-progress idea for resource-efficient and energy-adaptive VM

placement in data centers is proposed. To optimize the server utilization, many data
center operators already use server virtualization technologies and operate several
virtual machines on one physical server. This technology is the base for our further
optimizations. In our approach, we are at the point that the first steps of
optimizations have already been done. Hence, we are running a set of VMs
concentrated on a small number of potential servers. Unused servers are already
switched off. As further input, we get a target power consumption value.
It is generally accepted that applications operate ideally if they have access to all
required server resources. With the aim of improving the data center’s efficiency,
resource-competing VMs should not be operated on the same physical server
together. Our approach is to create a VM allocation that concentrates VMs with
suitable resource requirements on the same physical server for ideal application
performance and efficiency. In this constellation, each application has access to the
required server resources and operates ideally. Finally, the overall server resources
are more utilized than before and the efficiency rises. Beside the increased effi-
ciency, this situation also leads to a higher power consumption and application
performance. This scenario is suitable for times of high energy availability. Fol-
lowing the idea of green energy usage, this technology is also capable of reducing
the data center’s power consumption in situations of less green power availability.
Therefore, the methodology can be used to explicitly reduce resource utilization by
combining resource-competing applications, leading to lower power consumption
but also to a potentially reduced application performance.
In data centers, applications induce specific power consumptions by their evoked
server load. This required amount of power is so far understood as a fixed and
restricted value. Our concept is to let this amount of power become a controllable
value by applying a corresponding VM allocation.
The power consumption PCdc of a data center breaks down as follows:
PCdc ¼ PCSupport þ PCServers
The total power consumption is the sum of the power consumption of all data center
components. Beside the power consumption of all PMs PCServers , we have the power
consumption of the support infrastructure PCSupport i.e. network components,
cooling components, UPS, lights, etc.
Chen et al. [5] describe the power consumption of a server as the sum of its static
(idle, without workload) power consumption PCServers idle and its dynamic power
consumption PCservers dyn: :
PCServers ¼ PCServers idle þ PCservers dyn:
PCservers dyn: is the amount of power we directly take influence on. It reflects the
amount of power consumption deviance between 100 % server utilization and idle
mode. Idle servers still consume 60 % of their peak power draw [11].
Hence, a sustainable part (up to 40 %) of the server’s power consumption is
controllable; it can be increased in times of high energy availability and decreased
otherwise. Our approach is based on virtualization technology and the possibility to
live-migrate VMs. The methodology is agnostic to the operating applications. This
is an advantage compared to other task scheduling-based algorithms, since these
have to deal with task execution times and other application-specific topics. In our
approach, the applications are untouched and the technology is non-invasive

regarding the applications; it only takes effect on the availability of server
resources. The variable availability of server resources is a usual setting that
applications are confronted with.
As described in the related work part of this paper, the PM’s RAM can be a
limiting factor while migrating further VMs to the PM. In addition, we make use of
the technology to share RAM across the VMs to increase the number of VMs
operated on a PM.
The following diagrams illustrate the practice, how the methodology’s strategy
migrates VMs between physical servers.
In Fig. 1.1, the initial, non-optimized situation is displayed showing a set of VMs
operated on three physical servers. The resource utilization is highlighted (lighter
colors meaning low, darker colors high utilizations). On PM2, for example the
performance is affected by high network utilization.
Our methodology achieves an equilibrium allocation regarding the resource
utilization, as shown in Fig. 1.2. VMs to migrate are chosen depending on their
RAM size and their fraction of scarce resource utilization. The subsequent VM
permutation leads to an average utilization of all involved resources. Hence, the
approach increases efficiency and power consumption by resource usage
optimization.
The configuration is suitable for times of high energy availability and low energy
prices. In periods of less available renewable energy or high energy prices, we need
to reduce the power consumption while keeping a high efficiency level.
The situation, as shown in Fig. 1.3, is the result with reduced power consumption
objectives. The CPU utilization is reduced to likewise reduce the power consump-
tion as well while the utilization of other resources is balanced. The result is the
most effective constellation at reduced power conditions. The Dynamic Voltage
PM 1 PM 2 PM 3
VM1 VM7
VM2 VM1
VM3 VM1
VM4 VM7
VM9 VM1 VM1VM11 VM7
VM12 VM1
VM13 VM1
VM5 VM4
VM6 VM1
VM7 VM1 VM8 VM4 VM1
VM10 VM1 VM4 VM1 VM1
HDD NIC RAM CPU HDD NIC RAM CPU HDD NIC RAM CPU
Fig. 1.1 Schematic VM on physical server diagram: initial situation
PM 1 PM 2 PM 3
VM1 VM7
VM2 VM1 VM1 VM3 VM7
VM9 VM1
VM4 VM1
VM11 VM7
VM12 VM1
VM13 VM1
VM5 VM4
VM6 VM1
VM7 VM1 VM8 VM4 VM1 VM1 VM4
VM10 VM1 VM1
Fig. 1.2 Schematic VM on physical server diagram: optimized situation

PM 1 PM 2 PM 3
VM1 VM7
VM2 VM1
VM3 VM1
VM4 VM7
VM9 VM1 VM1VM11 VM7
VM12 VM1
VM13 VM1
Fig. 1.3 Schematic VM on physical server diagram: aim of reduced power consumption
and Frequency Scaling (DVFS) technique is used to adapt the power consumption
to the actual CPU utilization. DVFS allows a dynamical adaption of CPU voltage
and CPU frequency according to the current resource demand.
4.1 System Model
As described, we have an NP-hard complexity if we stick to methodologies that

involve all possible configurations to find the best suitable configuration for our
actual requirements. To reduce the complexity and the long computation time, we
change from an all-embracing global solution to local solving strategies. On the one
hand, this is required for online acting approaches and on the other hand, we assume
that we will not get significantly better overall solutions if we include all VMs to
find a suitable configuration.
In Fig. 1.4, a component model of the entire system is shown. We have an
application-monitoring component that delivers information about the applications
and servers to the service level management (SLM). The SLM component contains
all service level agreements (SLAs) and calculates new power target values for the
data center to observe the SLAs. These values are propagated to all optimizers,
working on every physical server. The optimizer compares the new incoming target
values with its own actual value. If the difference is in range of a predefined
hysteresis, the optimizer does not take any action. Otherwise it starts optimization.
If the target is not in the predefined range and the actual value is lower than the
target, the optimizer resolves resource competing constellations and hosts addi-
tional VMs from the offer pool. In the offer pool, all distributed optimizers can
announce VMs, for example, if they do not fit to their actual placement strategy.
The VMs in the offer pool are represented with their resource requirements that are
the base for later VM placement swaps. If the actual value is higher than the target,
the optimizer arranges a resource competing allocation to reduce the power
consumption.
The energy availability, energy prices and service level values are independent
and global values to aggregate to a target power consumption value. This is a task
for the central service level management (SLM) component of our system. Here we
do not have any local issues to attend, so we can calculate these values globally. As
an additional effect of the globally defined target power consumption value, we
Service Level Oﬀer Pool

Applicaon Monitoring Management
VM CPU RAM NIC I/O
target
VM1 0,3 0,3 0,8 0,1
VM2 0,3 0,1 0,1 0,1
VM3 0,2 0,2 0,7 0,1
VM4 0,3 0,3 0,2 0,6
VM5 0,8 0,2 0,8 0,1
target Opmizer actual target Opmizer actual target Opmizer actual
VM1 VM2 VM3 VM4 VM5 VM6 VM7 VM8 VM9 VM10 VM11 VM12
PM1 PM2 PM3
Fig. 1.4 Schematic system model
have evenly distributed server utilization. This reduces the occurrence of hot spots,
similar to the approach mentioned by Tang et al. [17].
We use a local optimizer component working on a single PM that focuses on a
solution for its own PM. This component has to find a solution for just one PM and
the set of possible VMs is reduced to the actual operated ones and to a subset of
those in the offer pool. As input, the optimizer receives a defined target power
consumption, which has to be reached with best possible efficiency.
4.2 Algorithm
CPU utilization is the most effective value regarding power consumption as

mentioned before. In other words, the overall CPU utilization is the value to
increase or decrease to take effect on the data center’s power consumption. Our
approach uses competing resource allocations to slow down applications and in
series the CPU utilization. Consolidating VMs on a PM that utilize the same
resources except the CPU can accomplish this. Consequently, the CPU utilization
and power consumption decreases. This practice affects the application’s perfor-
mance and we need a feedback that is sent from the application-monitoring
component to the SLM component to ensure the SLAs. With the information
about the SLAs and actual application performance, the SLM component is able
to calculate power consumption target values that achieve the economic data center
objectives.
The target power consumption is broadcasted to all PMs. The PM has got an
optimizer component that receives the target and compares it with its actual value.
If the target is similar to the actual value, the optimizer does not interfere. Other-
wise it starts optimizing. While doing this, the focus is kept on balanced resource
utilization and efficiency. Hence, the overall CPU utilization is reduced or
increased but all other resources are used as efficiently as possible. Balanced
resource utilization is always the goal except for CPU utilization and resources
that are used to build the competing resource situation. Merely the attainable CPU
utilization is a variable and implicit value that corresponds to the power consump-
tion target value.
Every PM’s optimizer strives to reach the target value by optimizing its own
situation. We have an offer pool of VMs, which can be accessed by every PM’s
optimizer. The optimizer is able to read the offered VMs from other PMs or even to
offer VMs. If the target value is greater than the actual value, the optimizer removes
suitable VMs from the pool to host until the target value is reached. If the target is
lower than the actual value, the optimizer offers VMs to the pool to reduce the own
value. Furthermore, additional VMs can be hosted from the pool to create compet-
ing resource situations to reduce the CPU utilization and to reach the target value.
Developing a reduced power consumption VM allocation can be done in three
ways:
(i) Migrate VMs to other PMs. This reduces the CPU utilization and the power
consumption by DVFS technology.
(ii) The optimizer arranges a resource competing allocation, which reduces the
CPU utilization and-as a result-decreases the power consumption by DVFS
technology.
(iii) The optimizer arranges CPU overprovisioning. CPU utilization is already at
100 % and further VMs will be hosted. The additional VMs do not increase the
PM’s power consumption but reduce the power consumption of the PM they
came from. Hence, the overall power consumption is decreasing.
The strategy to reduce the power consumption starts with (i) and is cascading
down to the methodology of (iii). At first, the target is strived with (i), if this is not
leading to the required results, we go on with (ii) and lastly with (iii). Using the
methodology of (i) means, we have no further risks of SLA-violation because the
application’s performance is not influenced. In (ii) und (iii) we potentially slow
down the applications, probably increasing the risk of SLA violations. Hence, the
methodology always starts in step (i).
The formal description of the efficiency and power consumption problem is a
follows: A set of virtual machines V ¼ fVM1 , . . . , VMn g and a set of physical
machines P ¼ fPM1 , . . . , PMk g is given. The VMs are represented by their
resource demand vectors di . The PMs are represented by their resource capacity
vectors ci . The goal is to find a configuration C so that for all PMs in P:
X
j
d i cs þ xs
i¼1
where the vector xs is an offset to control under- and overprovisioning of the server
resources on PMs while j is the total number of VMs on the PMs . We use xs to control
the resource utilization on the PMs to induce the intended server utilization and
thereby their power consumption.
To measure the quality of an allocated configuration C, we have now two

different metrics. On the one hand, we have the efficiency E:
work done
EðCÞ ¼
unit energy
On the other hand, we have the difference Δ between the PMs power consumption
PCserver and the target power consumption PCtarget :

Δðtarget, CÞ ¼ PCtarget PCserver ðCÞ
The Δ represents the deviance (positive) from the target power consumption. In case
of lower target power consumptions, a lasting deviance is the indicator to go on
with the next step (ii) or (iii).
The process of reaching a suitable VM placement and the behaviour of the
locally executed optimizer is demonstrated by the following pseudo code:
Inputs: t target power consumption for local PM, p actual PM’s power
consumption, resource utilization
Output: VM placement for local PM that evokes target power consumption
1. receive new target t given by SLM component

2. if t > p and the PM’s CPU utilization is 100 %, offer VMs to other PMs via
offer pool
3. if t > p and the PM’s CPU utilization is lower than 100 % and all other
resources are underutilized, the PM invites VMs to shelter from other PMs
with high CPU utilization
4. if t > p and the PM’s CPU utilization is lower than 100 % and other
resources are strong utilized, offer VMs to other PMs to solve the com-
peting resource situation
5. if t < p and the PM’s CPU utilization is lower than 100 % and other
resources are strong utilized, invite VMs to shelter from other PMs with
high CPU utilization
6. if t < p and the PM’s CPU utilization is 100 %, invite VMs to shelter from
other PMs to create resource competing situation
7. if t ¼ p do nothing
In addition to the event of changing power consumption targets, we have further

events to deal with. During the operation a host can become over- or underloaded. A
PM’s overload might lead to SLA violations and an underload means that the
efficiency is not at optimum level. Depending on the actual power consumption
Another random document with
no related content on Scribd:
time cold. With a man, his love is deep and deeply intense for a little
while; with a woman, it is not so deep or intense at the time, but
spreads over her whole life.”
With reference to the above analysis, which certainly is true in
parts as regards the world in general, and yet which seems far too
sweeping when applied to individual cases, Elissa was one of those
whom absence did not render indifferent; she was also one of those
women whose love had spread over her life. But it could by no
means be said of her that she found that the “have beens” were
enough for ever, nor that her love had not been so deep and intense
as Maharbal’s at the time. On the contrary, it was not only equally
deep and intense, but far more violent and incapable of being kept
under control. Elissa had not therefore been satisfied with merely
living on the past, but had been ardently looking forward to a future
when her five senses might be again gratified by the presence of her
lover. Her disappointment and depression were all the greater, and
her state of “accablement” became more utter, as the loving words
and expressions conveyed to her in her lover’s letter only made her
desire his personal presence the more intensely.
As with the pieced-together letter in her hand, and the faded roses
by her side, she lay silently weeping upon her luxurious couch, she
felt as if she had been struck with blows, so limp, so crushed was
she. But after a while, proud woman that she was, she called all her
pride, all her courage, to her aid, and rising from her couch rejoined
her uncle.
Her beautiful face was very pale, and there were deep violet rings
under her eyes, when, laying her bejewelled fingers upon Mago’s
arm, she addressed him as follows:
“Mine uncle Mago, it is not good for a girl to be so much alone as I
have been for years past. Neither father nor mother have I ever had
with me, nor even thou, mine uncle Mago, nor yet have I mine uncle
Hasdrubal. Until I took my friend Sophonisba to live with me, what
society have I had, save that of the empty-headed Princess Cœcilia,
a woman utterly devoid of intellect, whose only ideas are vapid
flirtations with anything or anybody—which foolish promiscuity
maketh her somewhat a danger in the city, by the way—and how
best to take care of her complexion:
“No wonder, then, oh mine uncle! that—neglected thus, and
thrown so utterly upon mine own resources—I have dwelt far too
much in my mind upon my lover Maharbal; for lover only he is to me
henceforth; I will continue no longer the farce of calling him my
husband. Had he been my husband, or desired to be my husband,
he would have come to me now. Therefore is he but my lover and
nought else, and, my lover having failed me, I will stay here to brood
in New Carthage no longer, but will accompany thee for a while to
the war against the Romans, with this thine army that thou hast
brought. I shall presently take thee all over the defences of the town,
and thou wilt see that I have not hitherto betrayed my trust, for all is
in order. And thou, mine uncle, shalt this day present unto me one of
the superior officers of thy force, a capable man, to accompany us
round the walls, and be also present at a review of my troops, which
I intend to hold in thine honour. To such a one will I delegate mine
authority here during mine absence, and thou shalt ratify such
appointment. Were it possible for me to know whither in Italy to seek
my father Hannibal, it is to him I would now proceed, and it would
perhaps be more fitting that I should do so, but for one reason. That
reason thou canst easily fathom; it exists in the presence of
Maharbal with my father’s army. For ’twould seem to all that I were
pursuing him, or that since he would not come to me I had gone to
him, and that shall never be said of Elissa, daughter of Hannibal.
Now, mine uncle, I have said: I accompany thee if thou wilt but have
me?” and she threw an arm around his neck caressingly.
“Ay, my dear niece, right gladly will I have thee with me, and do
even as thou hast said. For ’tis true that thou hast been neglected
hitherto, and life is short, especially in times of war, and blood is
thicker than water. I would right fain have thee with me, save for the
danger that thou mayst run of thy life. Say, if I take thee, wilt thou
promise me to be very careful of thine own safety, my pretty one, my
gallant soldier’s daughter?” And gently the uncle stroked the dark
tresses of the young woman, whose pale but determined face so
near his own shone with nobility, courage, and determination.
She embraced Mago, and smiled softly but somewhat ironically.
“Thou good uncle! I knew well that thou couldst not say nay. But
take care of myself!—nay, I will make no such promise. For am I not
Hannibal’s daughter? Ay, and his representative—yea, even a
general like unto thyself, although I never yet have led my troops in
the field. Moreover, thou hast never seen me in my war harness; but
thou shalt, and that right soon too.”
And now, laughing outright, she clapped her hands loudly, when
two female slaves came running in.
“Order my charger, and prepare me mine armour instantly, and be
in readiness to attire me.”
The slave girls retired instantly to do her bidding.
“Now, mine uncle,” quoth Elissa, blowing him back a kiss as she
stood in the doorway before following them, “say farewell for a space
to Elissa the woman, for in a moment thou shalt see only Elissa the
soldier, one who will, when required, bring with her to the battle,
under old Gisco, a body of well-disciplined troops, whom she hath
trained herself and can thoroughly rely upon. Some of thy large force
can remain here to replace them in the garrison of New Carthage.”
When, a few minutes later, Elissa reappeared, fully attired in her
light but glorious armour, carrying on her left arm a shining and
beautiful shield, inlaid with the horse of Carthage in gold, and having
two or three light throwing javelines in her right hand, Mago could not
resist a cry of admiration.
“By the great gods Melcareth and Moloch, thou art beautiful! I
would to the gods, indeed, that Hannibal could but see thee thus,
Elissa; verily, he would be proud of his daughter.”
“Who is, as thou shalt learn, mine uncle, by no means a maiden
travestied in warlike panoply merely for stage effect. Wilt thou
accompany me to the verandah? Now, what object shall I strike with
this javeline?”
Mago pointed out a distant and slender tree trunk.
Poising the javeline for a second, Elissa sent the weapon whizzing
through the air, and lo! it was quivering, buried to its head in the bark
of the sapling.
“Another object?” she asked.
“The silver figure of the god of love on the fountain; but methinks
’tis over far.”
“Not too far for me,” quoth Elissa; “this is a game that I play well,
mine uncle, for I have practised greatly.”
Again a javeline flew through the air with the most marvellous
precision, striking the neck of the little silver god with such force that
it was transfixed from side to side by the gleaming steel.
“By the great goddess of love herself!” cried Mago, in admiration,
“never saw I such dexterity. ’Tis evident that her son’s arrows are but
a toy compared to Elissa’s javelines.”
Elissa smiled.
“Now, wouldst see me on my war-horse, mine uncle Mago? ’Twas
Maharbal himself who taught me to ride when but a child, and I am
on horseback, as thou shalt see, a very Numidian. I have neither
saddle nor stirrups; but, merely for show’s sake, a bridle have I, with
silver chains for reins; likewise, I have a golden saddle-cloth, to the
surcingle of which the reins are, as thou seest, attached to prevent
them falling.”
An orderly was leading a splendid bay charger, thus caparisoned,
up and down before the verandah of the palace. Taking a short run,
Elissa sprang lightly into her seat across the horse.
“Some darts,” she cried; “give me some darts.”
Some half-dozen short, but heavy-headed darts were given to her,
which she grasped with her left hand below the shield.
Then pricking the horse with the point of one of the darts that she
took in her right hand, she started off at full gallop. Away she sped
across the lawn, and in and out among the trees, at such a pace that
Mago feared to see her brains dashed out against the tree trunks.
But nay, emerging safely from the trees she swept across an open
space beyond the fish pond, all the time performing warlike
evolutions with her shield; raising it, and protecting her head, or
throwing herself flat upon the horse’s back, and covering head and
shoulders with it completely.
A third evolution she performed, and that, likewise, while still at full
gallop. Suddenly, Mago could see nothing but the glittering shield
held alongside the horse’s neck, thus protecting it. All that was
visible of Elissa herself was one small foot barely showing above the
horse’s croup, her whole body being concealed behind the horse.
Then, as the horse came round again in a circle, thundering along
the path which led before the palace verandah, Elissa, springing up
to her seat again, discharged, with the rapidity of lightning, all her
darts in rapid succession. With each she struck the object aimed at.
With the last of the whizzing weapons she transfixed and slew a
glittering peacock which, frightened by the galloping horse, flew, from
its perch upon a marble portico, screaming overhead. Then whirling
short round again, she dashed back at the same speed, stopped
suddenly by using the reins for the first time, and pulling her horse
upon his haunches, sprung to the ground in a second as lightly as
she had mounted. She ran swiftly up the steps to her uncle,
somewhat out of breath, and with a heightened colour.
“What dost thou think of my horsemanship? The princess saith
that ’tis indelicate! But what dost thou think of thy warrior niece
thyself? Is she fit to accompany thee to the war against the
Romans?”
“Fit to accompany me to the war! Thou art fit to command the
army. Why, by Moloch himself! never, save in my beloved brother-in-
arms Maharbal, who did himself instruct thee, saw I such
horsemanship, combined with such precision in throwing the
weapons. In very truth will I take thee with me unto the battle, ay,
and willingly, for woe! I say, be to the enemy who should find himself
within reach of thy darts. But one thing thou must promise me. Keep
thou ever to this Numidian style of warfare, advancing and retiring on
horseback, and casting of darts and javelines. But the use of the
sword, for which thy bodily strength would not be sufficient, ever
avoid; likewise avoid, if possible, dismounting and fighting on foot.”
“Nevertheless, the use of the sword I know too, mine uncle, for
good old Gisco hath taught it me for years past.”
“Maybe! Maybe that he hath; but, for all that, promise me to keep,
if possible, to the horse and the dart-throwing, in which thou art more
than the match for any Roman, and thou shalt come with me into the
bloodiest battle. Give me thy word, Elissa.”
“I promise thee, mine uncle Mago, to do thy bidding in this matter,
and, further, in all else appertaining unto warfare, to be entirely
subservient unto thee.”
Thus it came to pass that, after a year or two’s campaigning,
Elissa was present at the fateful battle in which Mago defeated and
overthrew Cnœus Scipio. Further, while charging alongside Mago in
the hottest of the battle, it was even the hand of Hannibal’s daughter
which discharged the missile which struck the Roman General in the
joints of his armour, and cost him his life. As at about the same time,
Hasdrubal defeated Publius Scipio, and slew him also, for a time the
Carthaginians completely regained the upper hand in Spain. For the
brothers Scipio, being both dead, there was no one left to lead the
Roman forces.
Mago and Hasdrubal now joined hands, and drove the shattered
Roman troops into various camps and cities well to the north of the
Ebro, after which, Elissa, accompanied by her uncle Hasdrubal and
all his army, returned to New Carthage for the winter. But her uncle
Mago still kept the field.
CHAPTER XII.
SOPHONISBA AND SCIPIO.
Hasdrubal, remaining in New Carthage for a space longer, when

spring set in gave to his niece one day a delightful surprise.
“Elissa,” quoth he, one morn, “wouldst thou like to travel? wouldst
thou perchance like to see the African soil whence thy fathers
sprung? ’Tis charming, I warrant thee, at this season of the year, and
well worthy of a visit.”
Elissa sprung from her seat and dropped her embroidery work, for
she had, since the battle in which she had slain Cnœus Scipio,
resumed, on her return to New Carthage, her ordinary woman’s
attire and feminine avocations.
“Visit Carthage! mine uncle?” she cried excitedly, clasping her
hands in glee; “ ’tis the dream of my life to visit that glorious home of
mine ancestors.”
“Not so fast! not so fast, my niece; I said not visit Carthage, for
there I may not send thee at present, but visit African soil. For I have
it in my head to despatch thither an embassy to Syphax, King of the
Massæsyllians, a near relative of thy lover Maharbal, of which
embassy I propose to make General Hasdrubal, son of Gisco, the
chief. I further propose to send with him his daughter, thy friend and
companion, that beautiful young girl Sophonisba, and ’tis not meet
that she should travel without a responsible female companion.
Therefore, shouldst thou fancy a short sea journey, thou art welcome
to take advantage of this opportunity. ’Twill be, methinks, a change to
thee to visit the court of a prince upon African soil, after having all thy
life met with no princes save those of the Iberian race. His kingdom
is most fertile and lovely, much resembling in all things the climate of
this southern part of Spain; the language thou knowest, for it is thine
own Phœnician tongue. Moreover, thou speakest Greek fluently,
wherewith thou canst discourse with strangers should they be
present. Now, what dost say? Wilt thou go or nay? ’tis a chance that
may not hap again in thy lifetime.”
“Go! mine uncle, of course I will go. I long, indeed, to visit African
soil; and though it be not Carthage, yet are these Numidians the
vassals and friends of Carthage. ’Tis almost the same thing.”
“Vassals of Carthage they are, and friends sometimes. Syphax
was the friend of Carthage until lately, and likewise his nephew
Massinissa, ruler of Massyllia, the adjoining country to Carthage
itself. But latterly the Romans have been tampering with both, and I
have news that they, being sorely pressed by Hannibal in Italy, are
sending, or have already sent, a new embassy with rich presents
and many promises to these princes, with a view to securing their
alliance. Therefore, it behoves me to be upon my guard, and to bribe
them also. Fortunately we have all the wealth of the silver mines of
Southern Iberia at our command, and can therefore send, without
impoverishing ourselves, such riches to these barbarian kings as the
beggarly Romans can never even dream of. And that, therefore, is
what I shall do. Would but to the gods, I could send the treasure to
my brother Hannibal himself but no man knoweth where to find him.
He hath, ’tis said, recently utterly crushed the Romans in some
tremendous battle, but no man knows, as I said but now, where he
actually is. At all events, that is the reason that the Carthaginian
Government allege for giving him no succour, and as, despite his
repeated demands for reinforcements, the Government send him
none, and they will not give me a fleet to send to his support, I
cannot myself, unfortunately, assist him in that way. But by
preventing the Numidians from joining the Roman standard I can in
one way aid Hannibal. And ’tis possible thy going into Africa might
further the matter. For thou’rt young, handsome, and clever, and thy
wits might win what the sterling qualities of General Hasdrubal, the
son of Gisco, might not, with his rough and ready tongue, be able to
accomplish. Further, Sophonisba may attract the fancy of the King
Syphax. As for thyself, thou art affianced, to say the least of it, to his
kinsman, Maharbal, so thou art not only safe from any proposals of
marriage, but wilt come into his family group with particular rights to
be treated with the greatest consideration. Moreover, thou hast tact
in the highest degree, and should, as I tell thee in confidence I
desire, the African prince become enamoured of Sophonisba, whose
charms are really remarkable, thou canst guide the maiden herself,
and impress upon her the advantages of union with a king. For
although this Numidian hath many concubines, he is yet unmarried.
And his friendship and real alliance would be of the greatest
advantage to Hannibal at the present crisis. Therefore, my niece,
thou canst by going to the court of this barbarian greatly aid my
designs. He is, it seems, a really warlike man, and well worth the
winning over to our country’s cause, no matter what the bait
employed. Moreover, he is, if not quite young, yet well-favoured, and
such as any maiden might fancy. So also I hear is Massinissa, his
nephew, but Syphax is the more powerful. But I have said enough,
and if thou wilt accompany the mission I am convinced that thou wilt
succeed.”
“So poor Sophonisba is to be the bait! is she, mine uncle? Well,
’tis in our country’s cause, and after all, ’tis something to become the
wife of a king, that is to become herself a queen. Thou canst
therefore rely upon me. Should the man not prove an absolute ogre,
and thou sayst that he is far from that and well liking, I will persuade
Sophonisba, although sorely shall I grieve to part with her, to marry
him.”
“Then that is settled, Elissa; keep thou thine own counsel entirely,
and I will arrange about the details of the mission during the next few
days. Breathe not to Sophonisba herself one single word of what I
have said to thee.”
Elissa laughed aloud and patted her uncle on the cheek.
“What dost thou take me for, oh, Hasdrubal the son of Hamilcar?
Am I like a babbling brook, or like the Princess Cœcilia, widow of thy
late namesake and brother-in-law?”
“Whom I detest most cordially. Nay, nay!” replied Hasdrubal, “may
the gods forbid that thou shouldst resemble her, for she is odious! I
have it in my mind to crucify her one of these days to encourage
virtue in the other women in the palace. For she is most unvirtuous,
and worse than that, most unwise. What can I do with her if I slay her
not, thou knowest her well Elissa?”
“Watch her carefully, or marry her to someone, that is my advice.
To crucify her would be most unjust, for she hath hitherto harmed no
one. Her sole vice is folly, but that is, it must be owned, extreme.”
“Well, well, we can see about the fool later on. I shall perhaps
know how to deal with her. Methinks I will marry her to one of my
lieutenants. There is a certain prefect of horse that would suit her
admirably. He is of gigantic stature, almost as tall as thine own
Maharbal.”
“And she adores large men,” replied Elissa. “Well, I counsel thee,
mine uncle Hasdrubal, marry thou Cœcilia unto him without delay,
then shalt thou be relieved of a constant danger in the palace. For
there is no greater danger than in the constant presence of a foolish
woman!”
“ ’Tis true, my niece—’tis most true. I must consider it. But now let
us to the harbour and see about the ships.”
So the pair left the palace together and strolled down to the
harbour, where all fitting arrangements were made for the voyage to
Africa.
A fortnight later Elissa found herself with Sophonisba, now a girl of
seventeen, and her father Hasdrubal, the son of Gisco, arriving at
the Court of Syphax, and there they were most royally entertained.
Syphax himself was a splendid Numidian. Some forty years of
age, he was handsome, affable, well-instructed, and warlike. His
bearing was indeed that of a prince. Frank and good-natured,
generous to a fault, he was a man who never suspected evil in
others, because there was absolutely no guile in his own disposition.
His leanings were all towards Carthage, for until latterly the
Carthaginians had ever treated him well, and if latterly they had not
done so, he, with his generous nature, put the neglect simply down
to the expenses incurred by the long continued war.
There were present at his court, which was most magnificent and
luxurious, his nephew Massinissa, a small but muscular and wiry
man of an entirely different type to Syphax himself, and also the
members of a Roman embassy. And the head of this embassy was
Scipio Junior, who wore his left arm in a sling, and looked pale and
an invalid. For he had been sorely wounded in two places at a
comparatively recent battle, in which fight the man who had struck
him down had been his old antagonist Maharbal. Now, by some
strange dispensation of the gods, it was his lot to meet as friends in
a foreign court not only an embassy of his country’s enemies the
Carthaginians, but also the beautiful daughter of Hannibal himself,
Elissa, the betrothed of the very man whom upon three separate
occasions he had met hand to hand, and upon every occasion to his
own discomfiture. And now that he had met Elissa, he fell deeply in
love with her at first sight.
Although it was the fashion of the Romans and Carthaginians to
call all races but their own and the Greeks “barbarians,” the word did
not carry with it the significance that it has in these days; merely
meaning at first the inhabitants of Berber, the country to the North of
Africa. In fact, these very Numidians at whose court Elissa now
found herself, were Berbers or Barbarians. The remnants of this old
race, who are still numerous in the countries of Tunisia and Algeria,
are called Berbers unto this day. And from applying first to the
Berbers, the word barbarian came to have the signification of any
foreigner of no matter what race.
If the Numidian princes were called barbarians, it must not be
imagined that they were either barbaric in their ways, or that their
residences were by any means barbaric. On the contrary, not only
had they got the long-established civilisation and culture of the
neighbouring country of Carthage to guide them in their architecture,
but they had, in the beautiful horse-shoe arch, a grace and art of
their own which, introduced into Spain many hundreds of years later
by the descendants of the Berbers, the conquering Moors, has left its
traces unto this day in the Saracenic or Moorish arches of the
Alhambra at Granada and other magnificent buildings.
The Carthaginian embassy was received in state, and when
Elissa, on first landing, was borne in an open golden litter, with
Sophonisba by her side, up to the palace steps, the two girls thought
they had never seen anything so beautiful as the view of the land
and sea from the hill upon which the palace stood, and the first
appearance of the home of Syphax.
As they ascended the hillside to the sounds of sweet and
somewhat melancholy music from the musicians of the escort, they
traversed gardens blazing with geraniums and roses; the frequent
orange blossom, shining with its waxen, heavy-scented petals on
one side, being relieved everywhere by the crimson flowers of the
pomegranate. Overhead the date palms rustled in the fresh sea-
breeze, while below shone the blue sea, with a busy harbour full of
shipping. All along the coast, into the far distance, could be seen a
succession of green headlands, forming a charming variety to the
blue of the foam-flecked sea below, and the blue of the cloud-flecked
sky above.
But if the works of nature were beautiful all around, the works of
man were beautiful also. For the front façade of the palace, beneath
the shade of which Syphax and his suite was awaiting them,
consisted of a long unbroken line of horse-shoe arches of purest
marble, these arches being supported at each heel of the horse-
shoe by double pillars of pure white marble also. Above the façade
towered the palace, a marble building studded with horse-shoe-
shaped windows everywhere. Around each of these windows, to
afford relief to the eye, was a band, a foot wide, of polished black
stone, the effect of the contrast being delightful. Leading up to the
façade, which was raised some ten feet above the level of the
ground, there was a magnificent and very wide double flight of
curved marble steps, the curve of the steps again forming a perfect
horse-shoe. The double balustrades of this horse-shoe approach
were scarlet and green, with climbing geraniums, while gracefully
festooned up the marble pillars of the façade, and above the top of
the arches, were seen the most magnificent clusters of the purple
bourgainvillier flowers, and the graceful twining convolvulus, whose
bells of deepest blue hung in brilliant contrast to the pale green
leaves.
“Oh! what a lovely place,” exclaimed Sophonisba. “Surely it must
be the home of the gods themselves. Elissa, sawst thou ever a place
so lovely as this? There is nothing in Carthage itself that can
compare with it. Oh! I would that palace were mine.”
“Who knows but it may become so some day?” replied Elissa, with
a laugh. “Syphax is unmarried, thou knowest, and thy lovely fair skin
and auburn tresses will assuredly attract him greatly if he be not of
adamant, which, my Sophonisba, I have heard he is not by any
means.”
“Hush! Elissa,” replied the young girl, blushing. “Yonder
handsome, swarthy man, in the silver inlaid armour, standing before
the rest, is doubtless Syphax himself. Ah! he descends the stair-case
to meet us. I feel nervous; my heart is all in a flutter.”
“Ay! right noble is his mien, enough to make the heart of any girl
flutter; but now to salute him. Greeting! oh King Syphax. I, thy
humble slave, whom thou seest before thee, am Elissa, daughter of
Hannibal, and this maiden by my side is Sophonisba, daughter of
General Hasdrubal, who hath preceded us.”
“Greeting! a hearty greeting, Elissa, daughter of Hannibal! Greatly
honoured am I that so beauteous a princess should deign to shed
the light of her beauty upon our poor dwelling. Welcome art thou, ay,
doubly art thou welcome, seeing that thou art the betrothed of our
kinsman Maharbal.” And he smiled pleasantly as he kissed her hand.
“Welcome to thee also, oh Sophonisba; truly so fair a flower as thou
hath never yet blossomed in the gardens of Syphax. Would to the
gods that it might take root upon our Numidian soil, then would the
palace doors be bright, and the hearts of the people happy.”
With this gallant speech, and with open admiration in his eyes, the
Numidian king bent over and kissed in turn the hand of the charming
and deeply blushing Sophonisba. Then he ordered the litter to be
lowered from the shoulders of the gorgeously-attired bearers, and
personally assisted the maidens to alight.
Elissa was attired with great splendour, much after the fashion in
which she had been clothed upon the occasion when she descended
to the harbour some years before to meet the false Adherbal, and
was looking radiant. Sophonisba was also charmingly attired, but
somewhat more simply, as became her years.
All present upon the verandah were struck by the regal splendour
of Elissa, and the sweet charm of the fair maiden, Sophonisba,
whose supple, willowy form was set off to the greatest advantage by
the simple style of her raiment.
When they had been conducted up the marble steps by Syphax,
he in turn presented to the ladies first his nephew Massinissa, then
all the nobles of his court. After them he presented to Elissa the
young Roman noble Scipio, with the nobility of whose features
Hannibal’s daughter was greatly impressed.
Publius Cornelius Scipio the younger, afterwards distinguished, on
account of his feats on Libyan soil, by the surname of Africanus, was
by no means the stripling that he had been on the occasion of the
rescue of his noble father from the hands of Maharbal and the
butcher Monomachus. He was a stately and muscularly developed
man, and, save for his temporary pallor, strong and athletic-looking.
His features were extremely regular, his eyes blue, his hair light-
brown and curling. He wore a short, fair beard, which was
exceedingly becoming to him. There was an immense charm in his
manner, as, indeed, his face seemed to proclaim.
As Elissa gazed for the first time upon this young man, whose
advent in the world was to be so fateful for Hannibal and for
Carthage, she was struck by what she considered the goodness,
although by no means weakness, of his expression.
Scipio, on his side, was for a second struck dumb by the
magnificent and voluptuous beauty before him. Thus, for a second,
the representatives of the two hostile nations remained speechless
face to face in a sort of embarrassment.
The Numidian king, standing by, laughed heartily.
“Why, my young and noble guests, what is it? Are ye afraid of one
another, or would ye continue the war upon Libyan soil? Nay, nay;
here are ye on neutral ground, and let me assure thee, oh Scipio,
that no war is allowed at the Court of Syphax, save the havoc which
can be wrought by a beauteous maiden’s eyes.”
He spoke in Greek, in which language the young man, smiling in
turn, replied, as stooping gracefully over Elissa’s hand, he raised it to
his lips:
“And that is a war in which the lady Elissa hath already won the
victory. Alas! I fear she will be ever fatal to the race of Scipio,
whether in the field or in the court. In the former, if the voice of
rumour be true, she hath already slain the uncle in well-contested
battle by her feats of arms; and now, in the latter, she hath already
half-slain the nephew with those far more potent weapons with which
the gods on Olympus have endowed her. In sooth is she a most
redoubtable foe.”
“I render thee thanks, my lord Scipio. Here, beneath the hospitable
roof-trees of King Syphax, thou hast not much to fear, at all events.
For ’tis but a simple maiden, all unarmed and defenceless, that thou
seest before thee, and no thought hath she in her heart of warfare of
any description. Therefore, see thou to it, my lord, that thou provoke
not the battle.”
Elissa smiled, displaying her pearly teeth as she uttered these
words, and as she stood thus, her crimson lips slightly parted, and a
faint flush upon her peach-like cheek, she appeared to the already
enamoured young Roman as the terrestial personification of Venus
the queen of love herself.
“Most excellent!” quoth King Syphax smilingly; “ ’tis good advice
which thou hast given unto our noble Roman guest, for were I in his
place I would fly at once rather than risk an encounter in which
defeat were assured in advance.”
Then he turned and left them, hurrying off to the side of
Sophonisba, who was shyly responding to the advances of the
Prince Massinissa, to whom she had taken an instinctive dislike,
mingled with a feeling of dread.
For the man had snaky black eyes and a cruel look, as different
from the honest and open countenance of his uncle as are the
sulphurous fumes of Erebus from the heavenly lights of Elysium.
Sophonisba turned to her host gladly as he came to her side, and
listened willingly to his kindly and, it must be owned, somewhat
amorous conversation, Massinissa effacing himself as his uncle
appeared upon the scene. For of him he stood in dread, and likewise
hated him cordially, simply for the reason that he had once been
detected in plotting against his life, and had most magnanimously
been forgiven. For ’tis ever the way with traitors to hate those whom
they have wronged, but by whose benefits they are nevertheless not
ashamed to profit.
During the ensuing weeks all was merriment at the court of
Syphax. The affairs of the two separate embassies were entirely
neglected, for the king knew full well that no sooner should he come
to a determination with either one power or the other to make an
alliance, than the representatives of both would leave. Therefore,
being a man of most jovial disposition, and likewise of a most
hospitable turn of mind, he, on one pretext or another, constantly put
off all business discussions until the morrow. Thus, both the
embassies were forced to remain, awaiting the time when the
Numidian king should find leisure to discuss the important affairs of
State, which, however decided, would inevitably plunge his
dominions and forces, now at peace, into all the horrors of war.
In the meantime, the days were passed in the pleasures of the
chase, and the nights in feasting and carousal. The country
abounded in game of all descriptions, from the lordly lion to the fierce
wild boar and the timorous deer.
The former was hunted from the backs of elephants and slain with
arrows and darts; the wild boar was pursued on horseback and slain
by the riders with their darts and javelines; the deer were killed by
the use of the bow alone, the hunters being concealed in passes in
the hills, or glades in the forests, towards which the quarry was
driven. It was in the pursuit of the wild boar that Elissa shone, for
here her magnificent horsemanship and skill in discharging her
weapons while at full gallop came chiefly into play. There were other
ladies present at the court, many of them of great beauty, but after
the arrival of the Carthaginian mission, Syphax was during these
hunting parties never to be found by the side of any save
Sophonisba; while, where Elissa was, Scipio, whose arm was soon
healed, was ever in close attendance.
One day while pursuing the boar, Elissa, being magnificently
mounted, had far outstripped all the other riders. She overtook the
boar, and casting a javeline, struck the animal behind the shoulder.
The infuriated brute turned instantly and charged her horse, which
fell headlong, casting its rider heavily to the ground. The boar, after
venting its rage upon the prostrate horse, whom with its savage
tusks it utterly destroyed, attempted to trot off, but fell down dead
close by.
Scipio, was the only one of the hunters who, save Elissa, had in
the bushy country managed to follow the chase. He came upon the
body of the horse, the now lifeless boar, and the seemingly lifeless
form of Elissa, all three close together. When she came partly to her
senses again, she found herself closely clasped in the arms of the
young Roman warrior. His lips were upon her lips, his breath mingled
with her breath; her senses had not yet completely come back to her,
she was in a dream. Passionately he clasped her to his bosom;
wildly too, in a paroxysm of grief, he cried:
“Die not, beloved, for oh, I love thee—I love thee, Elissa! Say, dost
thou love me?”
“Ay,” she replied, with swimming eyes; “ay, I love thee, and that
right truly—Maharbal!”
Then she closed her eyes once more, and became again
insensible in the arms of Scipio.
Young Scipio, gnashing his teeth with rage, laid her inert body on
the ground. Then he rode off, and finding some of the beaters, told
them to seek her and bring her back to the palace. He himself,
cursing the very name of Maharbal, rode moodily home, avoiding the
remainder of the hunting party, whom he observed in the distance.
CHAPTER XIII.
ON THE BRINK.
When we last left Maharbal upon the shores of the Adriatic he was a
prey to great sorrow at the loss of his dear friend Mago. But soon he
had no time for any personal feelings, for the army was once more in
motion. Hannibal, ever mindful of his dream, proceeded to follow out
the plan that the dream had suggested, namely, the devastation of
Italy. Accordingly, ever leaving a destroyed territory in his wake, he
marched onward and southward. Every village that he came across
he pillaged and burned, every town or walled city that he met he laid
siege to, captured, and destroyed. It was not a part of his plan of
campaign to allow his followers to hamper themselves with the
quantities of female slaves that they took prisoners, as there could
be no means of exportation for them. Therefore, merely delaying for
a few days’ repose after the capture of each place, he caused the
army to relinquish all the women they had taken, and so to march
on, ever forward, unhampered save by the enormous booty they had
acquired.
The power of Rome having been apparently paralysed, he, for a
considerable space, wandered whither he would, utterly unopposed.
Having traversed, from end to end, the countries of Picenum,
Campania, Samnium, and Apulia; having for months and months
devastated all the richest country in Italy, under the very eyes of the
following force of Romans, under the Dictator Fabius, surnamed
Cunctator or the Lingerer, he seized upon and carried by assault the
citadel and town of Cannæ, where there was an immense store of
provisions and materials of war belonging to the Romans. There he
rested for a time, and armed all his Libyan infantry with Roman
armour and Roman weapons. What a delight must not the
Carthaginian chief have felt, as he dealt out by the thousand to his
followers the suits of armour that he had taken from the Roman
warriors even in their own country. He now had, however, not only
the most absolute confidence in himself and his mission, but a
sarcastic delight in thus arming his forces with Roman harness to
fight against the Romans themselves. And this feeling was shared by
the men of mixed nationalities in his army, who, with feelings of
triumph, arrayed themselves in the trappings of the enemy whom
they were commencing to despise.
Meanwhile, the members of the Senate at Rome were tearing their
hair. They determined that an effort must be made, and this puny
invader, who, with such a ridiculously small force, had dared to
affront all the might of Rome, must be crushed forthwith. Despite,
therefore, the previous disasters, they girt their loins together most
manfully, and prepared for new and more determined efforts to wipe
Hannibal and all his crew off the face of the earth.
What the power of the Roman Senate, what the resolution of the
Roman people must have been, is exemplified by the fact that,
despite previous losses, they soon had in the field an army
amounting in number to more than four times the usual annual levy
of legions. For it consisted, counting horse and foot, of nearly ninety-
eight thousand men! And the Dictator, the lingerer Fabius, having
been proved a failure, and he and his master of the horse, and
sometimes co-dictator, Minucius, having been repeatedly defeated in
various small actions and skirmishes, this enormous force was
placed under the command of the two new consuls for the year,
Paullus Æmilius, and Terentius Varro, the former being a patrician of
great fame, the latter a popular demagogue of plebeian origin.
Æmilius had already greatly distinguished himself in the Illyrian war,
for which he had celebrated a splendid triumph; but as for Varro, he
was, although the representative of the people, nothing but a vulgar
and impudent bully, with no other knowledge of war than his own
unbounded assurance. When Hannibal, with his usual military
genius, had seized upon the citadel of Cannæ, these two consuls,
burning to retrieve the frequent recent disasters, arrived upon the
scene and took over the command. But after all that had gone
before, they were not sure of themselves, and therefore persuaded
the out-going consuls, Cnœus Servilius and Marcus Atillus, to remain
and join in the battle. Marcus Minucius likewise, who had been co-
dictator with Fabius, returned to the army to take part in the great
fight which, with all his rashness, he had not himself been able to
precipitate during his own term of office, but which he knew to be
imminent. He had already suffered a defeat at the hands of
Hannibal, and was burning to gain his revenge. And now he knew
that he had his chance against the comparatively small force of the
presumptuous invader, for never, in all her history, had Rome put
such an enormous army in the field.
Hannibal and his army were encamped upon some heights to the
south of a river called the Aufidus. This stream was remarkable in
one respect, it being the sole stream in the whole of Italia which
flows through the range of the Apennine mountains, rising on their
western side, passing through the hills, and falling into the Adriatic
Sea on the eastern side of the Italian Peninsula. From the excellent
situation of the Carthaginian camp, all the military dispositions of the
Romans could be easily observed, and by means of the spies
employed by old Sosilus, Hannibal was not long in being informed of
the dissensions between the two consuls. Never was there an
instance in which the disadvantage of a dual command was shown
more than upon the present occasion, when one consul was in
command of the whole force one day and the other the next, and
what the one did to-day the other undid to-morrow. For it was the
custom in the Roman army when both consuls were present to give
to each the supreme command on alternate days. It was a wonder,
however, that after the example of the co-dictators Fabius Cunctator
and Marcus Minucius, who had found it an utter failure a short time
before, that this system of daily alternate command had not been
abandoned. For Fabius and Minucius had found it so unworkable
that they had for a time divided the army into two, each taking his
own half. And with his half only, having risked a battle, Minucius was
utterly defeated owing to an ambush of cavalry prepared by
Hannibal. The late Master of the Horse and his troops were, upon
this occasion, only saved from utter destruction by the Lingerer
setting his own half of the army in motion, and coming to his rash
colleague’s assistance in the nick of time, and checking the
Carthaginian pursuit, with much loss to the triumphant Phœnician
force. After that, Minucius had wisely resigned his right to the
command, leaving the entire power in the hands of Fabius.

Download pdf Advances And New Trends In Environmental And Energy Informatics Selected And Extended Contributions From The 28Th International Conference On Informatics For Environmental Protection 1St Edition Jorge ebook full chapter

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Download pdf Advances And New Trends In Environmental And Energy Informatics Selected And Extended Contributions From The 28Th International Conference On Informatics For Environmental Protection 1St Edition Jorge ebook full chapter

Uploaded by

Copyright:

Available Formats

Advances and New Trends in

Environmental and Energy Informatics

Advances and New Trends in Environmental Informatics

Information Technology in Environmental Engineering

Multimedia Tools and Applications for Environmental

Advances in Neurotechnology Electronics and Informatics

Informatics in Schools Improvement of Informatics

Advances in Visual Informatics 6th International Visual

International Conference on Mobile Computing and

Advances in Materials Science for Environmental and

Jorge Marx Gómez

Advances and New Trends

Ute Vogel Andreas Winter

Barbara Rapp Nils Giesen

ISSN 2196-8705 ISSN 2196-8713 (electronic)

Library of Congress Control Number: 2015957811

Springer Cham Heidelberg New York Dordrecht London

Printed on acid-free paper

Springer International Publishing AG Switzerland is part of Springer Science+Business Media

This book is a collection of selected outstanding scientific papers based on contri-

• Social Media for Sustainability, becoming an indispensable tool for value

• Hans-Knud Arndt, University of Magdeburg, Germany

• Werner Pillmann, International Society for Environmental Protection, Vienna,

Part II From Smart Grids to Smart Homes

8 Design, Analysis and Evaluation of Control Algorithms for

Part III Smart Transportation

Part IV Sustainable Enterprises and Management

Part V Environmental Decision Support

18 How a Computational Method Can Help to Improve the Quality

Part VI Social Media for Sustainability

Alexander Borgerding and Gunnar Schomaker

Keywords Data center • VM placement • Energy efficiency • Power-aware •

© Springer International Publishing Switzerland 2016 3

while j is the total number of VMs on the PM.

consumption is reduced by concentrating VMs on a fewer number of servers and

4 Resource-Efficient and Energy-Adaptive VM Placement

In this section the in-progress idea for resource-efficient and energy-adaptive VM

PCdc ¼ PCSupport þ PCServers

PCServers ¼ PCServers idle þ PCservers dyn:

approach, the applications are untouched and the technology is non-invasive

Fig. 1.1 Schematic VM on physical server diagram: initial situation

Fig. 1.2 Schematic VM on physical server diagram: optimized situation

4.1 System Model

As described, we have an NP-hard complexity if we stick to methodologies that

Service Level Oﬀer Pool

target Opmizer actual target Opmizer actual target Opmizer actual

PM1 PM2 PM3

Fig. 1.4 Schematic system model

CPU utilization is the most effective value regarding power consumption as

To measure the quality of an allocated configuration C, we have now two

1. receive new target t given by SLM component

In addition to the event of changing power consumption targets, we have further

Hasdrubal, remaining in New Carthage for a space longer, when

You might also like